<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.1 20151215//EN" "http://jats.nlm.nih.gov/publishing/1.1/JATS-journalpublishing1.dtd">
<article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" article-type="research-article" dtd-version="1.1" specific-use="sps-1.9" xml:lang="pt">
  <front>
    <journal-meta>
      <journal-id journal-id-type="nlm-ta">Arq Bras Cardiol</journal-id>
      <journal-id journal-id-type="publisher-id">abc</journal-id>
      <journal-title-group>
        <journal-title>Arquivos Brasileiros de Cardiologia</journal-title>
        <abbrev-journal-title abbrev-type="publisher">Arq. Bras. Cardiol.</abbrev-journal-title>
      </journal-title-group>
      <issn pub-type="ppub">0066-782X</issn>
      <issn pub-type="epub">1678-4170</issn>
      <publisher>
        <publisher-name>Sociedade Brasileira de Cardiologia - SBC</publisher-name>
      </publisher>
    </journal-meta>
    <article-meta>
      <article-id specific-use="scielo-v3" pub-id-type="publisher-id">jfHKLyXZyKN937FyqvCZB5h</article-id>
      <article-id specific-use="scielo-v2" pub-id-type="publisher-id">S0066-782X2025000800304</article-id>
      <article-id pub-id-type="other">00304</article-id>
      <article-id pub-id-type="doi">10.36660/abc.20240843</article-id>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Artigo Original</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>Aplicabilidade de Algoritmos de <italic>Machine Learning</italic> no Diagn&#243;stico de Fibrila&#231;&#227;o Atrial e SQTL por Interpreta&#231;&#227;o de Eletrocardiograma: Uma Revis&#227;o Sistem&#225;tica</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0009-0000-3187-7911</contrib-id>
          <name>
            <surname>Nascimento</surname>
            <given-names>Paulo Cainan Guimar&#227;es do</given-names>
          </name>
          <role>Concep&#231;&#227;o e desenho da pesquisa</role>
          <role>Obten&#231;&#227;o de dados</role>
          <role>An&#225;lise e interpreta&#231;&#227;o dos dados</role>
          <role>An&#225;lise estat&#237;stica</role>
          <role>Obten&#231;&#227;o de financiamento e Reda&#231;&#227;o do manuscrito</role>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Martins</surname>
            <given-names>Matthews Silva</given-names>
          </name>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff2">
            <sup>2</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-8779-1188</contrib-id>
          <name>
            <surname>Improta-Caria</surname>
            <given-names>Alex Cleber</given-names>
          </name>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff3">
            <sup>3</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-2161-3696</contrib-id>
          <name>
            <surname>Aras</surname>
            <given-names>Roque</given-names>
            <suffix>Junior</suffix>
          </name>
          <role>Concep&#231;&#227;o e desenho da pesquisa</role>
          <role>Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do</role>
          <xref ref-type="aff" rid="aff1">
            <sup>1</sup>
          </xref>
        </contrib>
      </contrib-group>
      <aff id="aff1">
        <label>1</label>
        <institution content-type="orgname">Universidade Federal da Bahia</institution>
        <institution content-type="orgdiv1">Faculdade de Medicina da Bahia</institution>
        <addr-line>
          <named-content content-type="city">Salvador</named-content>
          <named-content content-type="state">BA</named-content>
        </addr-line>
        <country country="BR">Brasil</country>
        <institution content-type="original">Universidade Federal da Bahia &#8211; Faculdade de Medicina da Bahia, Salvador, BA &#8211; Brasil</institution>
      </aff>
      <aff id="aff2">
        <label>2</label>
        <institution content-type="orgname">Universidade Federal do Esp&#237;rito Santo</institution>
        <institution content-type="orgdiv2">Laborat&#243;rio de Fisiologia Molecular e Intelig&#234;ncia Artificial</institution>
        <institution content-type="orgdiv1">Departamento de Ci&#234;ncias Fisiol&#243;gicas</institution>
        <addr-line>
          <named-content content-type="city">Vit&#243;ria</named-content>
          <named-content content-type="state">ES</named-content>
        </addr-line>
        <country country="BR">Brasil</country>
        <institution content-type="original">Universidade Federal do Esp&#237;rito Santo &#8211; Laborat&#243;rio de Fisiologia Molecular e Intelig&#234;ncia Artificial, Departamento de Ci&#234;ncias Fisiol&#243;gicas, Vit&#243;ria, ES &#8211; Brasil</institution>
      </aff>
      <aff id="aff3">
        <label>3</label>
        <institution content-type="orgname">Universidade de S&#227;o Paulo</institution>
        <institution content-type="orgdiv2">Laborat&#243;rio de Bioqu&#237;mica e Biologia Molecular do Exerc&#237;cio</institution>
        <institution content-type="orgdiv1">Escola de Educa&#231;&#227;o F&#237;sica e Esporte</institution>
        <addr-line>
          <named-content content-type="city">S&#227;o Paulo</named-content>
          <named-content content-type="state">SP</named-content>
        </addr-line>
        <country country="BR">Brasil</country>
        <institution content-type="original">Universidade de S&#227;o Paulo &#8211; Laborat&#243;rio de Bioqu&#237;mica e Biologia Molecular do Exerc&#237;cio, Escola de Educa&#231;&#227;o F&#237;sica e Esporte, S&#227;o Paulo, SP &#8211; Brasil</institution>
      </aff>
      <author-notes>
        <corresp id="c01"><label>Correspond&#234;ncia:</label> Alex Cleber Improta-Caria &#8226; Universidade de S&#227;o Paulo, Laborat&#243;rio de Bioqu&#237;mica e Biologia Molecular do Exerc&#237;cio - Av. Professor Mello Moraes, 65. CEP 05508-030, Cidade Universit&#225;ria, S&#227;o Paulo, SP - Brasil E-mail: <email>alexcaria.personal@hotmail.com</email> </corresp>
        <fn fn-type="edited-by">
          <label>Editor respons&#225;vel pela revis&#227;o:</label>
          <p>Marcio Bittencourt</p>
        </fn>
        <fn fn-type="conflict">
          <label>Potencial conflito de interesse:</label>
          <p>N&#227;o h&#225; conflito com o presente artigo.</p>
        </fn>
      </author-notes>
      <pub-date date-type="pub" publication-format="electronic">
        <day>01</day>
        <month>09</month>
        <year>2025</year>
      </pub-date>
      <pub-date date-type="collection" publication-format="electronic">
        <month>07</month>
        <year>2025</year>
      </pub-date>
      <volume>122</volume>
      <issue>8</issue>
      <elocation-id>e20240843</elocation-id>
      <history>
        <date date-type="received">
          <day>12</day>
          <month>12</month>
          <year>2024</year>
        </date>
        <date date-type="rev-recd">
          <day>23</day>
          <month>04</month>
          <year>2025</year>
        </date>
        <date date-type="accepted">
          <day>07</day>
          <month>05</month>
          <year>2025</year>
        </date>
      </history>
      <permissions>
        <license license-type="open-access" xlink:href="https://creativecommons.org/licenses/by/4.0/" xml:lang="en">
          <license-p>Este &#233; um artigo de acesso aberto distribu&#237;do sob os termos da licen&#231;a de atribui&#231;&#227;o pelo Creative Commons</license-p>
        </license>
      </permissions>
      <abstract>
        <title>Resumo</title>
        <sec>
          <title>Fundamento</title>
          <p><italic>Machine learning</italic> (ML) &#233; um tipo de algoritmo que aprende de forma aut&#244;noma a reconhecer padr&#245;es complexos. No contexto diagn&#243;stico de arritmias card&#237;acas, esses algoritmos demonstraram avan&#231;os significativos devido &#224; sua capacidade de fornecer interpreta&#231;&#227;o automatizada e reconhecimento de padr&#245;es em eletrocardiogramas (ECGs).</p>
        </sec>
        <sec>
          <title>Objetivo</title>
          <p> Analisar e identificar a aplicabilidade, validade e viabilidade de modelos de algoritmos de ML no processo diagn&#243;stico de arritmias card&#237;acas por meio de interpreta&#231;&#227;o eletrocardiogr&#225;fica automatizada.</p>
        </sec>
        <sec>
          <title>M&#233;todos</title>
          <p> Esta revis&#227;o sistem&#225;tica da literatura foi relatada de acordo com as diretrizes PRISMA. As buscas foram realizadas na Biblioteca Cochrane, EMBASE, LILACS e PubMed entre fevereiro de 2022 e novembro de 2022. O per&#237;odo do estudo abrange artigos publicados entre 2017 e 2022.</p>
        </sec>
        <sec>
          <title>Resultados</title>
          <p> A busca no banco de dados gerou 119 resultados, abrangendo tr&#234;s subtemas: S&#237;ndrome do QT Longo (SQTL), intervalo QT corrigido (QTc) e fibrila&#231;&#227;o atrial (FA). A FA foi o tema mais prevalente. Os tamanhos das amostras foram bastante vari&#225;veis. Os resultados foram, em sua maioria, satisfat&#243;rios. No diagn&#243;stico de SQTL usando Intelig&#234;ncia Artificial (IA), o algoritmo superou os m&#233;todos convencionais na distin&#231;&#227;o diagn&#243;stica. Na avalia&#231;&#227;o do QTc, n&#227;o houve diferen&#231;a entre o ECG integrado &#224; IA e o ECG convencional. No diagn&#243;stico de FA, os algoritmos, modelos e dispositivos demonstraram alta sensibilidade e especificidade, al&#233;m de maior acur&#225;cia.</p>
        </sec>
        <sec>
          <title>Conclus&#227;o</title>
          <p> Os modelos de ML no processo diagn&#243;stico de arritmias card&#237;acas s&#227;o vi&#225;veis e est&#227;o em r&#225;pido desenvolvimento. Apresentam valores de acur&#225;cia entre 96,4% e 98,2%, sensibilidade entre 92,8% e 99,4% e especificidade entre 95% e 98,1%, particularmente no diagn&#243;stico de fibrila&#231;&#227;o atrial.</p>
        </sec>
      </abstract>
      <abstract abstract-type="graphical">
        <p>
          <fig id="f01">
            <label>Figura Central:</label>
            <caption>
              <title>Aplicabilidade de Algoritmos de <italic>Machine Learning</italic> no Diagn&#243;stico de Fibrila&#231;&#227;o Atrial e SQTL por Interpreta&#231;&#227;o de Eletrocardiograma: Uma Revis&#227;o Sistem&#225;tica</title>
            </caption>
            <attrib>Figura criada usando Napkin.ai</attrib>
            <alternatives>
              <graphic xlink:href="https://minio.scielo.br/documentstore/1678-4170/jfHKLyXZyKN937FyqvCZB5h/d4f408a3e4858efd5f14babfcc9103eb10417cf8.tif"/>
              <graphic xlink:href="https://minio.scielo.br/documentstore/1678-4170/jfHKLyXZyKN937FyqvCZB5h/32d30297262f856ddcc8f475d63bd0a717a10ec9.png" specific-use="scielo-web"/>
              <graphic xlink:href="https://minio.scielo.br/documentstore/1678-4170/jfHKLyXZyKN937FyqvCZB5h/812ee245864a120016b85c3173cf7d2f77b0391c.jpg" specific-use="scielo-web" content-type="scielo-267x140"/>
            </alternatives>
          </fig>
        </p>
      </abstract>
      <kwd-group xml:lang="pt">
        <title>Palavras-chave</title>
        <kwd>Arritmias Card&#237;acas</kwd>
        <kwd>Intelig&#234;ncia Artificial</kwd>
        <kwd>Aprendizado de M&#225;quina</kwd>
        <kwd>Rede Nervosa</kwd>
        <kwd>Eletrocardiografia</kwd>
      </kwd-group>
      <funding-group>
        <award-group>
          <funding-source>Funda&#231;&#227;o de Amparo &#224; Pesquisa do Estado de S&#227;o Paulo</funding-source>
          <award-id>2022/02339-4</award-id>
        </award-group>
        <award-group>
          <funding-source>Funda&#231;&#227;o de Amparo &#224; Pesquisa do Estado de S&#227;o Paulo</funding-source>
          <award-id>#2024/17783-2</award-id>
        </award-group>
        <funding-statement>Fontes de financiamento: O presente estudo n&#227;o teve fontes de financiamento externas.</funding-statement>
      </funding-group>
      <counts>
        <fig-count count="4"/>
        <table-count count="6"/>
        <equation-count count="0"/>
        <ref-count count="29"/>
      </counts>
    </article-meta>
  </front>
  <body>
    <sec sec-type="intro">
      <title>Introdu&#231;&#227;o</title>
      <p>Segundo a Organiza&#231;&#227;o Mundial da Sa&#250;de, as doen&#231;as cardiovasculares (DCV) s&#227;o as principais causas de morte em todo o mundo, representando 32%.<sup><xref ref-type="bibr" rid="B1">1</xref></sup> Entre as DCV, as arritmias s&#227;o as mais comuns e se caracterizam por anormalidades na gera&#231;&#227;o ou condu&#231;&#227;o de impulsos el&#233;tricos, ocorrendo devido &#224; fun&#231;&#227;o card&#237;aca anormal, resultando em ritmos card&#237;acos irregulares. A ocorr&#234;ncia de arritmias pode ser indolente, mas pode causar consequ&#234;ncias graves.<sup><xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref></sup></p>
      <p>A abordagem diagn&#243;stica para arritmias card&#237;acas envolve dois componentes principais: hist&#243;ria cl&#237;nica e eletrocardiograma (ECG). A hist&#243;ria cl&#237;nica lista batimentos card&#237;acos irregulares, dispneia, fadiga, tontura e, quando n&#227;o tratada, hist&#243;rico de acidente vascular cerebral e insufici&#234;ncia card&#237;aca. A suspeita pode ser corroborada por altera&#231;&#245;es no exame f&#237;sico cardiovascular por meio de ausculta com estetosc&#243;pio e palpa&#231;&#227;o de pulsos. O diagn&#243;stico &#233; confirmado com o registro do ECG Holter por 24 horas e posterior certifica&#231;&#227;o de ritmos irregulares por um m&#233;dico especialista. Entretanto, o m&#233;todo convencional de diagn&#243;stico por ECG Holter, em geral, imp&#245;e limita&#231;&#245;es devido &#224; falta de flexibilidade mec&#226;nica ao paciente e ao m&#233;todo de an&#225;lise retrospectiva, sem monitoriza&#231;&#227;o e acompanhamento em tempo real.<sup><xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref></sup></p>
      <p>No entanto, algoritmos de Intelig&#234;ncia Artificial (IA) e seus derivados surgiram como um m&#233;todo confi&#225;vel para identificar e classificar padr&#245;es eletrocardiogr&#225;ficos que podem sugerir anormalidades.<sup><xref ref-type="bibr" rid="B5">5</xref></sup> Dorado-D&#237;az et al.<sup><xref ref-type="bibr" rid="B6">6</xref></sup> referem-se &#224; IA como um &#8220;campo da ci&#234;ncia da computa&#231;&#227;o que tenta imitar o processo cognitivo humano, a capacidade de aprendizagem e o armazenamento de conhecimento&#8221;. Em Ci&#234;ncias da Sa&#250;de, algoritmos de IA s&#227;o predominantemente aplicados em predi&#231;&#227;o, recomenda&#231;&#227;o e suporte diagn&#243;stico com base em modelos de <italic>Machine learning</italic> (ML) e <italic>Deep Learning</italic> (DL). Conceitualmente, o ML se concentra no desenvolvimento de algoritmos que permitem que sistemas de computador aprendam sem programa&#231;&#227;o expl&#237;cita. Em contraste, os modelos de DL derivam suas capacidades preditivas de Redes Neurais Artificiais (RNAs) com m&#250;ltiplas camadas de processamento de informa&#231;&#245;es.<sup><xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref></sup></p>
      <p>Nesse contexto, algoritmos de ML agora s&#227;o capazes de analisar ECGs, fornecer interpreta&#231;&#245;es automatizadas e prever riscos associados.<sup><xref ref-type="bibr" rid="B8">8</xref></sup> A literatura destaca previs&#245;es derivadas da otimiza&#231;&#227;o de etapas processuais, como processamento de sinais, extra&#231;&#227;o de vari&#225;veis significativas e classifica&#231;&#227;o de algoritmos.<sup><xref ref-type="bibr" rid="B9">9</xref></sup> Apesar da complexidade computacional intr&#237;nseca, h&#225; diversos avan&#231;os em sistemas que desempenham essa fun&#231;&#227;o.</p>
      <p>Portanto, ferramentas de IA podem complementar o monitoramento m&#233;dico, desde o n&#237;vel preventivo at&#233; o de reabilita&#231;&#227;o. O desenvolvimento de ML capaz de conduzir an&#225;lises de ECG &#233; essencial na detec&#231;&#227;o precoce de condi&#231;&#245;es card&#237;acas anormais remotamente e no aux&#237;lio ao processo diagn&#243;stico de arritmias e outras condi&#231;&#245;es de sa&#250;de.<sup><xref ref-type="bibr" rid="B5">5</xref></sup> Al&#233;m disso, o monitoramento card&#237;aco de longo prazo, o reconhecimento e a classifica&#231;&#227;o de arritmias com base na an&#225;lise do tra&#231;ado de ECG podem ser um processo demorado e desnecessariamente exaustivo para o cardiologista. Assim, t&#233;cnicas de detec&#231;&#227;o de arritmia auton&#244;mica s&#227;o &#250;teis para tais execu&#231;&#245;es. Nesse cen&#225;rio, t&#233;cnicas computadorizadas de reconhecimento de padr&#245;es potencialmente ajudar&#227;o a fornecer diagn&#243;stico e interven&#231;&#227;o m&#233;dica sempre que necess&#225;rio.<sup><xref ref-type="bibr" rid="B3">3</xref></sup></p>
      <p>Considerando o exposto, o estudo da IA e suas tecnologias derivadas na &#225;rea da sa&#250;de se mostra altamente relevante. Portanto, este trabalho teve como objetivo, por meio de uma revis&#227;o sistem&#225;tica da literatura cient&#237;fica, explorar a aplicabilidade, a validade e a viabilidade de modelos de ML no processo diagn&#243;stico de arritmias card&#237;acas. O estudo buscou identificar os benef&#237;cios e as limita&#231;&#245;es metodol&#243;gicas, os potenciais impactos cl&#237;nicos e sociais dessas tecnologias e suas implica&#231;&#245;es para a estrutura da educa&#231;&#227;o e da pr&#225;tica m&#233;dica.</p>
    </sec>
    <sec sec-type="methods">
      <title>M&#233;todos</title>
      <p>O delineamento do estudo &#233; uma revis&#227;o sistem&#225;tica da literatura conduzida utilizando os crit&#233;rios estabelecidos pelo m&#233;todo PRISMA (<italic>Preferred Reporting Items for Systematic Reviews and Meta-Analyses</italic>), conforme detalhado nas diretrizes metodol&#243;gicas para elabora&#231;&#227;o de revis&#245;es sistem&#225;ticas e metan&#225;lises de ensaios cl&#237;nicos randomizados fornecidas pelo Minist&#233;rio da Sa&#250;de do Brasil.<sup><xref ref-type="bibr" rid="B10">10</xref></sup></p>
      <p>Para a busca bibliogr&#225;fica sobre o tema pretendido, foram utilizadas as seguintes bases de dados: Biblioteca Cochrane, Embase, LILACS (Literatura Latino-Americana e do Caribe em Ci&#234;ncias da Sa&#250;de) e PubMed-MEDLINE (Sistema Online de An&#225;lise e Recupera&#231;&#227;o de Literatura M&#233;dica). Os termos de busca foram derivados do vocabul&#225;rio DeCS/MeSH (Descritores em Ci&#234;ncias da Sa&#250;de/Cabe&#231;alhos de Assunto M&#233;dico), complementados por termos de entrada. Os descritores inclu&#237;ram: arritmias; arritmia card&#237;aca ou card&#237;aca; <italic>machine learning</italic>; <italic>machine learning</italic> n&#227;o supervisionado; <italic>machine learning</italic> supervisionado; <italic>deep learning</italic>; redes neurais, computador ou redes neurais de computador ou rede neural computacional ou redes neurais computacionais; intelig&#234;ncia artificial ou IA ou intelig&#234;ncia computacional; e eletrocardiografia ou ECG ou EKG. A busca foi realizada utilizando a seguinte configura&#231;&#227;o de descritores e operadores booleanos: <italic>((cardiac arrhythmia) OR (arrhythmia) AND (machine learning)) OR (unsupervised machine learning)) OR (supervised machine learning)) OR (deep learning) OR (neural network) OR (artificial intelligence) AND (electrocardiography).</italic></p>
      <p>Durante o processo de triagem, foram considerados os estudos que atenderam aos seguintes crit&#233;rios de inclus&#227;o: 1) pesquisas envolvendo seres humanos ou dados obtidos de bancos de dados humanos; 2) publica&#231;&#245;es em sua vers&#227;o final em peri&#243;dicos cient&#237;ficos revisados por pares; 3) estudos dos seguintes tipos: estudos de caso-controle, estudos de coorte, ensaios cl&#237;nicos, relatos de caso e s&#233;ries de casos; 4) publicados entre 2017-2022; 5) dispon&#237;veis em portugu&#234;s ou ingl&#234;s. Os crit&#233;rios de exclus&#227;o inclu&#237;ram: 1) publica&#231;&#245;es em idiomas diferentes de portugu&#234;s ou ingl&#234;s; 2) estudos com descri&#231;&#245;es experimentais sem aplica&#231;&#227;o cl&#237;nica do modelo algor&#237;tmico; 3) artigos n&#227;o relacionados ao diagn&#243;stico de arritmias card&#237;acas; 4) artigos que n&#227;o aplicam modelos de IA ao diagn&#243;stico de arritmias card&#237;acas; 5) artigos que prop&#245;em m&#233;todos para diagn&#243;stico de arritmias card&#237;acas usando exames diferentes do ECG; 6) artigos que descrevem o uso de IA para diagn&#243;stico de condi&#231;&#245;es diferentes de arritmias card&#237;acas; 7) outros tipos de publica&#231;&#245;es, como editoriais, diretrizes, livros, revis&#245;es sistem&#225;ticas e meta-an&#225;lises; 8) resultados duplicados nos bancos de dados pesquisados.</p>
      <p>O processo de triagem dos resultados da busca e extra&#231;&#227;o de dados foi baseado no m&#233;todo PRISMA e realizado por um &#250;nico operador. Esse processo resultou na cria&#231;&#227;o de um banco de dados e de um modelo de tabela, permitindo a organiza&#231;&#227;o dos dados recuperados em configura&#231;&#245;es alinhadas aos crit&#233;rios do m&#233;todo empregado, servindo como precursor para an&#225;lises futuras. Utilizando essa metodologia, o objetivo foi avaliar estudos cl&#237;nicos que abordassem o uso e a aplicabilidade de algoritmos de ML no processo diagn&#243;stico de arritmias card&#237;acas por meio da interpreta&#231;&#227;o automatizada de ECGs.</p>
    </sec>
    <sec sec-type="results">
      <title>Resultados</title>
      <p>A busca nas bases de dados gerou um total de 119 resultados, incluindo 38 da Biblioteca Cochrane, 44 da Embase, nenhum da LILACS e 37 da PubMed. Ap&#243;s a aplica&#231;&#227;o dos crit&#233;rios de exclus&#227;o, 98 publica&#231;&#245;es foram exclu&#237;das pelos seguintes motivos: n&#227;o avaliaram a detec&#231;&#227;o de arritmias card&#237;acas (83), n&#227;o aplicaram algoritmos de IA para detec&#231;&#227;o de arritmias (5), abordaram estudos em andamento (3), foram classificadas como outros tipos de textos cient&#237;ficos (3), focaram em um modelo experimental (1), descreveram a avalia&#231;&#227;o diagn&#243;stica de arritmias usando m&#233;todos diferentes do ECG (1), n&#227;o possu&#237;am uma vers&#227;o completa do texto (1) e n&#227;o forneceram o texto completo em ingl&#234;s ou portugu&#234;s (1). Consequentemente, 13 publica&#231;&#245;es foram consideradas eleg&#237;veis para revis&#227;o completa e sistem&#225;tica (<xref ref-type="fig" rid="f02">Figura 1</xref>).</p>
      <p>
        <fig id="f02">
          <label>Figura 1</label>
          <caption>
            <title>&#8211; Fluxograma da Revis&#227;o Sistem&#225;tica da Literatura.</title>
          </caption>
          <graphic xlink:href="1678-4170-abc-122-08-e20240843-40843-gf01.tif"/></fig>
      </p>
      <p>Os 13 artigos considerados eleg&#237;veis foram inclu&#237;dos na s&#237;ntese qualitativa (<xref ref-type="table" rid="t1">Tabela 1</xref>). A busca final foi realizada em 30 de novembro de 2022. Em rela&#231;&#227;o aos pa&#237;ses de origem das publica&#231;&#245;es inclu&#237;das, elas representavam sete locais: Estados Unidos (4), China (3), Coreia do Sul (2), Alemanha (1), Fran&#231;a (1), Taiwan (1) e Finl&#226;ndia (1). Em termos de delineamento do estudo, os artigos compreenderam estudos de coorte prospectivos (7), estudos de coorte retrospectivos (2), estudos de caso-controle (2) e ensaios cl&#237;nicos randomizados (2). Todos os estudos receberam aprova&#231;&#227;o dos comit&#234;s de &#233;tica de suas respectivas institui&#231;&#245;es.</p>
      <p>
        <table-wrap id="t1">
          <label>Tabela 1</label>
          <caption>
            <title>&#8211; S&#237;ntese do Desenho, Resultados e Conclus&#245;es - SQTL</title>
          </caption>
          <table frame="hsides" rules="groups">
            <colgroup>
              <col/>
              <col/>
            </colgroup>
            <thead>
              <tr>
                <th align="left">Autor, ano, (ref. N&#186;)</th>
                <th align="left">Bos, 2021<sup><bold>11</bold></sup></th>
              </tr>
            </thead>
            <tbody>
              <tr>
                <td>Desenho do estudo</td>
                <td>Estudo de caso-controle. Objetivo: Determinar se a IA &#233; capaz de distinguir pacientes com SQTL de pacientes com QTc normal usando ECG de 12 deriva&#231;&#245;es. Amostra: 2.984 pacientes previamente diagnosticados com SQTL ou que receberam alta como saud&#225;veis. O diagn&#243;stico de caso-controle foi realizado usando ECGs de 12 deriva&#231;&#245;es dispon&#237;veis de 2.059 pacientes. Desfecho principal: A IA foi capaz de distinguir pacientes com SQTL, prolongamento persistente do intervalo QT e diferencia&#231;&#227;o gen&#233;tica.</td>
              </tr>
              <tr>
                <td>Modelo IA</td>
                <td>Rede Neural Profunda.</td>
              </tr>
              <tr>
                <td>Resultados</td>
                <td>A IA foi capaz de distinguir entre as duas popula&#231;&#245;es (pacientes com SQTL e pacientes sem SQTL). Precis&#227;o: 82,2%; sensibilidade: 83,7%; especificidade: 80,6%; VPP: 83,2%; e VPN: 81,3%.</td>
              </tr>
              <tr>
                <td>Conclus&#227;o</td>
                <td>O ECG-IA superou o ECG convencional na distin&#231;&#227;o entre pacientes com SQTL e aqueles que receberam alta sem o diagn&#243;stico. Al&#233;m disso, conseguiu diferenciar entre tr&#234;s subtipos gen&#233;ticos de SQTL.</td>
              </tr>
            </tbody>
          </table>
          <table-wrap-foot>
            <fn id="TFN1">
              <p>Esta an&#225;lise foi conduzida usando um n&#237;vel de signific&#226;ncia estat&#237;stica de 5% (p = 0,05). IA: intelig&#234;ncia artificial; ECG: eletrocardiograma; SQTL: S&#237;ndrome do QT Longo.</p>
            </fn>
          </table-wrap-foot>
        </table-wrap>
      </p>
      <p>Os artigos analisados se concentraram em tr&#234;s subt&#243;picos da arritmologia card&#237;aca: S&#237;ndrome do QT Longo (SQTL), intervalo QT corrigido (QTc) e fibrila&#231;&#227;o atrial (FA). No estudo sobre SQTL, o objetivo foi avaliar a capacidade dos modelos de IA de detectar essa s&#237;ndrome em indiv&#237;duos.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Em rela&#231;&#227;o &#224; arritmia induzida por medicamentos, os pesquisadores buscaram avaliar a preval&#234;ncia de arritmia por meio da an&#225;lise automatizada do QTc em indiv&#237;duos submetidos a terapias espec&#237;ficas.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> A FA, por outro lado, apresentou o maior n&#250;mero de estudos, que visavam principalmente avaliar a precis&#227;o,<sup><xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref></sup> viabilidade,<sup><xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref></sup> confiabilidade,<sup><xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref></sup> desempenho,<sup><xref ref-type="bibr" rid="B17">17</xref></sup> efici&#234;ncia,<sup><xref ref-type="bibr" rid="B18">18</xref></sup> seguran&#231;a,<sup><xref ref-type="bibr" rid="B18">18</xref></sup> poder discriminativo,<sup><xref ref-type="bibr" rid="B19">19</xref></sup> sensibilidade, especificidade e precis&#227;o<sup><xref ref-type="bibr" rid="B20">20</xref></sup>dos algoritmos e dispositivos associados utilizados para a detec&#231;&#227;o de epis&#243;dios de FA. No entanto, outros estudos buscaram explicar o processo de tomada de decis&#227;o dos modelos de ML no diagn&#243;stico de arritmias<sup><xref ref-type="bibr" rid="B21">21</xref>-<xref ref-type="bibr" rid="B23">23</xref></sup>Os tamanhos das amostras foram altamente vari&#225;veis. Todos os estudos utilizaram amostras de pacientes ou dados coletados de bancos de dados de institui&#231;&#245;es m&#233;dicas associadas &#224; pesquisa.</p>
      <p>Analisando os resultados, a maioria dos resultados foi satisfat&#243;ria e alinhada aos objetivos iniciais. Para modelos aplic&#225;veis &#224; S&#237;ndrome do QT Longo (SQTL), o algoritmo proposto distinguiu com sucesso as duas popula&#231;&#245;es investigadas: indiv&#237;duos com SQTL e aqueles sem a s&#237;ndrome.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Em rela&#231;&#227;o &#224; detec&#231;&#227;o de FA usando algoritmos baseados em IA, os artigos relataram altas taxas de sensibilidade, especificidade, valor preditivo positivo (VPP), valor preditivo negativo (VPN), precis&#227;o e outras m&#233;tricas de desempenho.<sup><xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B23">23</xref></sup> Na avalia&#231;&#227;o de arritmias induzidas por terapia farmacol&#243;gica, houve concord&#226;ncia entre os padr&#245;es de intervalo QTc derivados de ECGs convencionais e ECGs de dispositivos vest&#237;veis.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> Quanto aos modelos algor&#237;tmicos explicativos para diagn&#243;stico de arritmia, os resultados variaram: um estudo mostrou uma &#225;rea sob a curva (AUC) relativamente moderada,<sup><xref ref-type="bibr" rid="B21">21</xref></sup> enquanto outro demonstrou melhor desempenho.<sup><xref ref-type="bibr" rid="B22">22</xref></sup></p>
      <p>Em rela&#231;&#227;o aos resultados e conclus&#245;es para o diagn&#243;stico de SQTL usando IA, o ECG baseado em IA superou o m&#233;todo convencional em capacidade de distin&#231;&#227;o diagn&#243;stica e foi capaz de diferenciar subtipos gen&#233;ticos. Para o diagn&#243;stico de FA usando IA, os algoritmos, modelos e dispositivos demonstraram alta sensibilidade na detec&#231;&#227;o de FA na &#250;ltima hora,<sup><xref ref-type="bibr" rid="B13">13</xref></sup> maior sensibilidade e especificidade em compara&#231;&#227;o ao gerenciamento convencional e outros algoritmos comparativos,<sup><xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B17">17</xref></sup> e maior precis&#227;o na detec&#231;&#227;o em diferentes posturas<sup><xref ref-type="bibr" rid="B18">18</xref></sup>e tempos.<sup><xref ref-type="bibr" rid="B19">19</xref></sup> No contexto da monitoriza&#231;&#227;o de arritmias induzidas por f&#225;rmacos, n&#227;o houve variabilidade entre os modelos de IA e os ECGs convencionais.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> No entanto, modelos algor&#237;tmicos explicativos para diagn&#243;stico de arritmia produziram resultados amb&#237;guos: um apresentou limita&#231;&#227;o na detec&#231;&#227;o de FA durante o ritmo sinusal,<sup><xref ref-type="bibr" rid="B21">21</xref></sup> enquanto outro correlacionou com sucesso as caracter&#237;sticas dos tra&#231;os de ECG com o diagn&#243;stico de arritmia usando IA.<sup><xref ref-type="bibr" rid="B22">22</xref></sup> Os desfechos, resultados e conclus&#245;es est&#227;o resumidos nas <xref ref-type="table" rid="t1">Tabelas 1</xref>, <xref ref-type="table" rid="t2">2</xref> e <xref ref-type="table" rid="t3">3</xref>.</p>
      <p>
        <table-wrap id="t2">
          <label>Tabela 2</label>
          <caption>
            <title>&#8211; S&#237;ntese do Desenho, Resultados e Conclus&#245;es &#8211; Fibrila&#231;&#227;o Atrial</title>
          </caption>
          <table frame="hsides" rules="groups">
            <colgroup>
              <col/>
              <col/>
              <col/>
              <col/>
              <col/>
            </colgroup>
            <thead>
              <tr>
                <th align="left">Autor, ano, (ref. N&#186;)</th>
                <th>Projeto</th>
                <th>Modelo de IA</th>
                <th>Resultados</th>
                <th>Conclus&#227;o</th>
              </tr>
            </thead>
            <tbody>
              <tr>
                <td>Wasserlauf et al., 2019<xref ref-type="bibr" rid="B13"><sup>13</sup></xref> </td>
                <td align="center">Estudo de coorte prospectivo. Objetivo: Comparar a precis&#227;o de um dispositivo sens&#237;vel &#224; FA atrial com a grava&#231;&#227;o simult&#226;nea de um monitor card&#237;aco vest&#237;vel. Amostra: Treinamento de IA com 7.500 dados de ECG; valida&#231;&#227;o de coorte com 26 pacientes. Principais resultados: O dispositivo com sensor de ECG demonstrou alta sensibilidade para detectar FA na &#250;ltima hora em uma popula&#231;&#227;o ambulatorial.</td>
                <td align="center">Rede Neural Profunda</td>
                <td align="center">Dos 82 epis&#243;dios detectados pelo monitor card&#237;aco implant&#225;vel, o dispositivo sens&#237;vel &#224; FA identificou 80, alcan&#231;ando uma sensibilidade de 97,5% por epis&#243;dio.</td>
                <td align="center">Os resultados demonstram que o dispositivo vest&#237;vel com o sensor de ECG, o aplicativo e o algoritmo investigativo s&#227;o altamente sens&#237;veis para detectar epis&#243;dios de FA na &#250;ltima hora em uma popula&#231;&#227;o ambulatorial e para avaliar a dura&#231;&#227;o da FA quando comparado com um MCI.</td>
              </tr>
              <tr>
                <td>Huang et al., 2021<xref ref-type="bibr" rid="B15"><sup>15</sup></xref> </td>
                <td align="center">Ensaio cl&#237;nico randomizado. Objetivo: Avaliar a viabilidade e a confiabilidade de um dispositivo de ECG autoaplic&#225;vel e sistema de monitoramento para detec&#231;&#227;o de FA. Amostra: 218 pacientes previamente submetidos &#224; abla&#231;&#227;o, randomizados em dois grupos: grupo BT monitorado pelo algoritmo de IA e grupo TF monitorado por acompanhamento m&#233;dico tradicional. Principais desfechos: A sensibilidade e a especificidade do algoritmo de IA foram superiores &#224;s da detec&#231;&#227;o automatizada de FA.</td>
                <td align="center">Algoritmo de IA</td>
                <td align="center">Viabilidade do algoritmo de IA no acompanhamento: 26.133 registros de ECG, com detec&#231;&#227;o de 12,6% de FA confirmada em revis&#227;o manual por cardiologistas, 14,8% pelo algoritmo automatizado de detec&#231;&#227;o de FA e 13,2% pelo algoritmo gen&#233;rico de IA. O modelo de IA detectou mais recorr&#234;ncia de FA na FA parox&#237;stica ap&#243;s abla&#231;&#227;o (p = 0,0099), mas n&#227;o na FA persistente (p = 0,7910). A sensibilidade e a especificidade do algoritmo gen&#233;rico de IA para detec&#231;&#227;o de FA (94,4% e 98,5%) foram maiores do que as do algoritmo autom&#225;tico de detec&#231;&#227;o (90,7% e 96,2%).</td>
                <td align="center">O acompanhamento ap&#243;s abla&#231;&#227;o utilizando o algoritmo de IA leva &#224; detec&#231;&#227;o mais frequente de recorr&#234;ncia de FA. O algoritmo de IA demonstrou maior precis&#227;o no diagn&#243;stico de ECG e foi mais eficaz do que as estrat&#233;gias tradicionais.</td>
              </tr>
              <tr>
                <td>Jacobsen et al., 2020<xref ref-type="bibr" rid="B17"><sup>17</sup></xref> </td>
                <td align="center">Estudo de coorte prospectivo. Objetivo: Avaliar o desempenho de um dispositivo m&#233;dico vest&#237;vel que utiliza tecnologia de fotopletismografia para detectar FA em pacientes hospitalizados com FA. Amostra: 102 pacientes internados em um hospital alem&#227;o com FA documentada. Principais resultados: O dispositivo demonstrou melhor capacidade de detec&#231;&#227;o de FA do que o monitoramento Holter tradicional.</td>
                <td align="center">Rede Neural Profunda</td>
                <td align="center">O dispositivo detectou epis&#243;dios de FA no conjunto de dados com uma sensibilidade de 95,2% e especificidade de 92,5%.</td>
                <td align="center">A detec&#231;&#227;o de FA por um dispositivo m&#233;dico vest&#237;vel &#233; uma abordagem vi&#225;vel e confi&#225;vel.</td>
              </tr>
              <tr>
                <td>Fu et al., 2021<xref ref-type="bibr" rid="B18"><sup>18</sup></xref> </td>
                <td align="center">Estudo de coorte prospectivo. Objetivo: Avaliar a efic&#225;cia e a seguran&#231;a da detec&#231;&#227;o de FA e fornecer um m&#233;todo confi&#225;vel e n&#227;o invasivo para o rastreamento e o tratamento da FA na pr&#225;tica di&#225;ria. Tamanho da amostra: 114 pacientes. Principais resultados: ECG din&#226;mico vest&#237;vel com algoritmo de IA pode detectar FA e analisar ritmos card&#237;acos em diferentes posturas e ap&#243;s exerc&#237;cios.</td>
                <td align="center">Algoritmo de IA</td>
                <td align="center">O m&#233;todo detectou a ocorr&#234;ncia de FA na posi&#231;&#227;o supina com acur&#225;cia de 96,4%, sensibilidade de 92,4%, especificidade de 100%, VPP de 100% e VPN de 93,8%. A detec&#231;&#227;o de FA na posi&#231;&#227;o ortost&#225;tica alcan&#231;ou acur&#225;cia de 98,2%, sensibilidade de 100%, especificidade de 100%, VPP de 100% e VPN de 96,8%.</td>
                <td align="center">O ECG din&#226;mico vari&#225;vel pode detectar FA em ritmos card&#237;acos em diferentes posturas e ap&#243;s exerc&#237;cios.</td>
              </tr>
              <tr>
                <td>Noseworthye et al., 2022<xref ref-type="bibr" rid="B19"><sup>19</sup></xref> </td>
                <td align="center">Ensaio cl&#237;nico n&#227;o randomizado. Objetivo: Avaliar o poder discriminativo do modelo de IA em adultos. Avaliar a efic&#225;cia da estrat&#233;gia de rastreio de FA guiada por IA em compara&#231;&#227;o com o tratamento habitual. Amostra: 1.003 pacientes. Principais resultados: O modelo de IA detectou ritmos e dura&#231;&#227;o da FA.</td>
                <td align="center">Algoritmo de IA</td>
                <td align="center">Cinquenta e quatro pacientes foram diagnosticados recentemente com FA com dura&#231;&#227;o de 30 segundos ou mais e estratificados em grupos de baixo e alto risco. Um padr&#227;o semelhante foi observado nos grupos com FA com dura&#231;&#227;o de 6 minutos ou mais e 24 horas ou mais. Alto risco foi associado a uma maior carga de FA (m&#233;dia de 4,97% para baixo risco; m&#233;dia de 20,32% para alto risco).</td>
                <td align="center">O dispositivo &#233; capaz de detectar FA em v&#225;rios momentos e estratificar o risco. Os resultados corroboram um programa de triagem guiado por IA, de baixo custo, altamente escal&#225;vel e centrado no paciente.</td>
              </tr>
              <tr>
                <td>Yang et al., 2022<xref ref-type="bibr" rid="B21"><sup>21</sup></xref> </td>
                <td align="center">Estudo de coorte prospectivo. Objetivo: Propor uma abordagem de ML baseada em caracter&#237;sticas que forne&#231;a resultados explic&#225;veis e possa ser treinada em um banco de dados de tamanho vi&#225;vel. Amostra: Dados digitais de registros de ECG de 10 e 12 deriva&#231;&#245;es de um hospital taiwan&#234;s. Principais resultados: O modelo foi capaz de isolar caracter&#237;sticas temporais e espaciais das ondas P e prever padr&#245;es de FA.</td>
                <td align="center">Modelo de <italic>Machine Learning</italic></td>
                <td align="center">Predi&#231;&#227;o de FA usando o modelo de extra&#231;&#227;o de onda P por machine learning: o modelo atingiu a maior AUC (0,64). O modelo apresentou especificidade moderada (0,71), mas baixa sensibilidade (0,47).</td>
                <td align="center">Os dados mostram poder limitado para as caracter&#237;sticas propostas na detec&#231;&#227;o de pacientes com FA durante a RS. Com o processamento e a extra&#231;&#227;o de caracter&#237;sticas, &#233; poss&#237;vel construir modelos para identificar FA potencial, mas o modelo ainda &#233; impreciso.</td>
              </tr>
              <tr>
                <td>Zhu et al., 2022<xref ref-type="bibr" rid="B16"><sup>16</sup></xref> </td>
                <td align="center">Estudo de coorte prospectivo. Objetivo: Desenvolver, implementar e validar um algoritmo de detec&#231;&#227;o de FA baseado em fotopletismografia para <italic>smartwatches</italic> em pacientes com diagn&#243;stico de FA. Amostra: 204 participantes. Desfecho principal: O algoritmo permitiu a detec&#231;&#227;o passiva de FA com base em fotopletismografia usando um dispositivo vest&#237;vel.</td>
                <td align="center">Algoritmo de IA</td>
                <td align="center">O algoritmo detectou FA em 148 de 204 pacientes ao longo de 4 semanas. Sensibilidade: 87,8%; especificidade: 97,4%; poder de decis&#227;o (ritmo sinusal ou indeterminado): 67,8%.</td>
                <td align="center">O algoritmo demonstrou a viabilidade de detectar FA com precis&#227;o e estimar de forma n&#227;o invasiva a carga da FA.</td>
              </tr>
              <tr>
                <td>Jo et al.,2021<xref ref-type="bibr" rid="B22"><sup>22</sup></xref> </td>
                <td align="center">Estudo retrospectivo. Objetivo: Desenvolver e validar um modelo de <italic>deep learning</italic> explic&#225;vel (XDM) baseado em uma &#225;rvore de decis&#227;o apoiada por rede neural para classifica&#231;&#227;o de arritmias. Amostra: Um conjunto de 72.740 registros de ECG de 42.880 pacientes de um hospital sul-coreano entre 2017 e 2020.</td>
                <td align="center">Modelo de <italic>Deep Learning</italic></td>
                <td align="center">O escore F1 da valida&#231;&#227;o externa do XDM foi: RSN (Ritmo Sinusal Normal): 0,990; FA ou Flutter Atrial: 0,955; Taquicardia Supraventricular: 0,777; Bloqueio AV Completo: 0,828; Ritmo do Marcapasso: 0,671; Acur&#225;cia agregada: 0,844. FA e flutter atrial foram fortemente correlacionados com caracter&#237;sticas como a presen&#231;a da onda P e irregularidade, enquanto o bloqueio AV completo mostrou forte correla&#231;&#227;o com a dissocia&#231;&#227;o AV.</td>
                <td align="center">O modelo classifica com precis&#227;o arritmias em v&#225;rios formatos de ECG usando conjuntos de dados de valida&#231;&#227;o externa.</td>
              </tr>
              <tr>
                <td>Jo et al., 2021<xref ref-type="bibr" rid="B23"><sup>23</sup></xref> </td>
                <td align="center">Estudo retrospectivo. Objetivo: Propor um m&#233;todo para construir um XIA. Desenvolver e validar um modelo de <italic>deep learning</italic> explic&#225;vel para detectar FA usando v&#225;rios formatos de ECG. Amostra: 115.485 ECGs para desenvolvimento do modelo de IA; 12.914 ECGs para treinamento; 99.965 para valida&#231;&#227;o externa. Principais resultados: O modelo superou a autointerpreta&#231;&#227;o de ECG da m&#225;quina em todos os conjuntos de dados de valida&#231;&#227;o. Tamb&#233;m verificou o motivo (onda P e irregularidade) para sua conclus&#227;o.</td>
                <td align="center">Modelo de <italic>Deep Learning</italic></td>
                <td align="center">O modelo de IA mostrou sensibilidade de 0,928, especificidade de 0,950, VPP de 0,615, VPN de 0,993 e precis&#227;o de 0,948 na detec&#231;&#227;o de FA.</td>
                <td align="center">Os resultados indicaram que a metodologia XIA poderia ser usada para descrever o racioc&#237;nio por tr&#225;s da decis&#227;o do modelo de detectar FA com alto desempenho.</td>
              </tr>
              <tr>
                <td>Chen et al., 2020<xref ref-type="bibr" rid="B20"><sup>20</sup></xref> </td>
                <td align="center">Estudo de coorte prospectivo. Objetivo: Avaliar a sensibilidade, especificidade e precis&#227;o de um <italic>smartwatch</italic> com PPG e ECG que utiliza um algoritmo de IA para detec&#231;&#227;o de FA. Amostra: 401 pacientes entre maio e junho de 2019. Principais resultados: A sensibilidade, especificidade e precis&#227;o da pulseira equipada com algoritmos de PPG, ECG e IA demonstram desempenho satisfat&#243;rio em curto prazo.</td>
                <td align="center">Algoritmo de IA</td>
                <td align="center">O dispositivo com PPG apresentou os seguintes resultados para detec&#231;&#227;o de FA: sensibilidade: 80,00%; especificidade: 96,81%; acur&#225;cia: 90,52%; VPP: 93,75%; e VPN: 89,01%.</td>
                <td align="center">O dispositivo apresentou boa sensibilidade, especificidade e precis&#227;o na determina&#231;&#227;o da presen&#231;a de FA.</td>
              </tr>
              <tr>
                <td>V&#228;liaho et al., 2019<xref ref-type="bibr" rid="B14"><sup>14</sup></xref> </td>
                <td align="center">Estudo de caso-controle. Objetivo: Avaliar a precis&#227;o de uma pulseira PPG comercialmente dispon&#237;vel na detec&#231;&#227;o de pulsos individuais na FA e avaliar a confiabilidade de dois algoritmos de detec&#231;&#227;o de FA comumente utilizados com base na PPG. Amostra: 213 pacientes no total, 106 para o grupo FA e 107 para o grupo RS (controle). Principais achados: A pulseira equipada com PPG, utilizando dois algoritmos de detec&#231;&#227;o de FA, foi capaz de diagnosticar FA com alta sensibilidade e especificidade.</td>
                <td align="center">Algoritmo de IA</td>
                <td align="center">A detec&#231;&#227;o de FA por PPG mostrou sensibilidade de 96,2% e especificidade de 98,1% com o algoritmo de evid&#234;ncia de FA e sensibilidade de 95,3% e especificidade de 98,1% com COSEn.</td>
                <td align="center">Pulseiras de fotopletismografia podem ajudar a detectar casos assintom&#225;ticos ou &#8220;silenciosos&#8221; de FA.</td>
              </tr>
            </tbody>
          </table>
          <table-wrap-foot>
            <fn id="TFN2">
              <p>Todas as an&#225;lises foram conduzidas usando um n&#237;vel de signific&#226;ncia estat&#237;stica de 5% (p = 0,05). IA: intelig&#234;ncia artificial; ECG: eletrocardiograma; FA: fibrila&#231;&#227;o atrial; MCI: monitor card&#237;aco invasivo; VPP: valor preditivo positivo; VPN: valor preditivo negativo; XIA: modelo de IA explic&#225;vel; PPG: fotopletismografia; AV: atrioventricular.</p>
            </fn>
          </table-wrap-foot>
        </table-wrap>
      </p>
      <p>
        <table-wrap id="t3">
          <label>Tabela 3</label>
          <caption>
            <title>&#8211; S&#237;ntese de Desenho, Resultados e Conclus&#245;es - Arritmia induzida por f&#225;rmacos</title>
          </caption>
          <table frame="hsides" rules="groups">
            <colgroup>
              <col/>
              <col/>
            </colgroup>
            <thead>
              <tr>
                <th align="left">Autor, ano, (ref. n&#186;)</th>
                <th align="left">Maille et al., 2021<sup><bold>12</bold></sup></th>
              </tr>
            </thead>
            <tbody>
              <tr>
                <td>Desenho do estudo</td>
                <td>Estudo de coorte prospectivo. Objetivo: Comparar o QTc calculado usando este algoritmo em ECGs de deriva&#231;&#227;o &#250;nica de um smartwatch com o QTc medido em ECGs convencionais de 12 deriva&#231;&#245;es em pacientes com COVID-19 em est&#225;gio inicial tratados com o regime HCQ-AZM. Amostra: um total de 85 pacientes; 76 em regime medicamentoso. Principais desfechos: Nas medi&#231;&#245;es, houve concord&#226;ncia entre o padr&#227;o do intervalo QTc do ECG de 12 deriva&#231;&#245;es e o QTc medido no SW-ECG correspondente.</td>
              </tr>
              <tr>
                <td>Modelo IA</td>
                <td>Rede neural convolucional.</td>
              </tr>
              <tr>
                <td>Resultados</td>
                <td>Concord&#226;ncia entre o intervalo QTc padr&#227;o de 12 deriva&#231;&#245;es medido manualmente na deriva&#231;&#227;o II ou V5 nos dias 0, 6 e 10 e o QTc IA medido no SW-ECG correspondente.</td>
              </tr>
              <tr>
                <td>Conclus&#227;o</td>
                <td>H&#225; pouca diferen&#231;a significativa entre os m&#233;todos de IA e ECG convencional para diagnosticar arritmia induzida por medicamentos.</td>
              </tr>
            </tbody>
          </table>
          <table-wrap-foot>
            <fn id="TFN3">
              <p>Esta an&#225;lise foi conduzida usando um n&#237;vel de signific&#226;ncia estat&#237;stica de 5% (p = 0,05). IA: intelig&#234;ncia artificial; ECG: eletrocardiograma</p>
            </fn>
          </table-wrap-foot>
        </table-wrap>
      </p>
    </sec>
    <sec sec-type="discussion">
      <title>Discuss&#227;o</title>
      <p>Nesta revis&#227;o sistem&#225;tica, os algoritmos automatizados para detec&#231;&#227;o de arritmia cobriram tr&#234;s apresenta&#231;&#245;es principais: S&#237;ndrome do QT longo, outras anormalidades do intervalo QT e FA.</p>
      <sec>
        <title>S&#237;ndrome do QT longo</title>
        <p>Na S&#237;ndrome do QT Longo, o estudo analisado nesta revis&#227;o conseguiu distinguir popula&#231;&#245;es com uma acur&#225;cia de 82,2%, apresentando uma sensibilidade de 83,7% (VPP de 83,2%) e especificidade de 80,6% (VPN de 81,3%). Isso destaca que a aplica&#231;&#227;o de algoritmos de IA no diagn&#243;stico dessa arritmia demonstra um grau satisfat&#243;rio de acur&#225;cia, sensibilidade e especificidade.<sup><xref ref-type="bibr" rid="B11">11</xref></sup>Em rela&#231;&#227;o &#224;s anormalidades do intervalo QT, o estudo inclu&#237;do nesta revis&#227;o demonstrou concord&#226;ncia entre os resultados obtidos no ECG convencional de 12 deriva&#231;&#245;es e no ECG com algoritmo de IA integrado a um dispositivo vest&#237;vel equipado com tecnologia de fotopletismografia (PPG). Portanto, houve diferen&#231;a m&#237;nima entre os achados do ECG convencional e os sugeridos pelo algoritmo de IA.<sup><xref ref-type="bibr" rid="B12">12</xref></sup></p>
        <p>Hermans et al.<sup><xref ref-type="bibr" rid="B24">24</xref></sup> pesquisaram previamente caracter&#237;sticas morfol&#243;gicas que poderiam apoiar o diagn&#243;stico de SQTL e outras varia&#231;&#245;es do intervalo QT. Esses autores investigaram os valores agregados de marcadores morfol&#243;gicos da onda T no in&#237;cio do estudo e em um modelo estendido. Conclu&#237;ram que a morfologia da onda T tem um valor agregado na distin&#231;&#227;o de pacientes com SQTL de familiares com gen&#243;tipo negativo. Essa caracter&#237;stica morfol&#243;gica pode explicar a precis&#227;o do trabalho de Bos et al.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Em rela&#231;&#227;o &#224;s anormalidades do intervalo QT, Prifti et al. <sup><xref ref-type="bibr" rid="B25">25</xref></sup>relatam modelos de ML capazes de prever pacientes propensos a desenvolver prolongamento significativo do intervalo QT induzido por medicamentos.</p>
        <p>Corroborando as contribui&#231;&#245;es de Simon et al.<sup><xref ref-type="bibr" rid="B26">26</xref></sup>nas altera&#231;&#245;es eletrofisiol&#243;gicas do intervalo QT, a interpretabilidade se d&#225; por meio da precis&#227;o preditiva. M&#233;todos interpret&#225;veis permitem uma inspe&#231;&#227;o completa das intera&#231;&#245;es que levam a arritmias nessa classe, seja pela express&#227;o de uma altera&#231;&#227;o gen&#233;tica ou por intera&#231;&#245;es medicamentosas espec&#237;ficas. Essa interpretabilidade &#233; &#250;til na preven&#231;&#227;o de complica&#231;&#245;es e deve ser considerada para integrar a modelagem preditiva em ferramentas de apoio &#224; decis&#227;o cl&#237;nica.</p>
      </sec>
      <sec>
        <title>Fibrila&#231;&#227;o atrial</title>
        <p>Em rela&#231;&#227;o &#224; detec&#231;&#227;o geral de FA, caracterizada pela detec&#231;&#227;o de eventos de FA independentemente da dura&#231;&#227;o, da posi&#231;&#227;o do indiv&#237;duo ou de fatores morfol&#243;gicos espec&#237;ficos do tra&#231;ado eletrocardiogr&#225;fico, os estudos demonstraram alto grau de acur&#225;cia, sensibilidade e especificidade. Wasserlauf et al.<sup><xref ref-type="bibr" rid="B13">13</xref></sup>e Chen et al.<sup><xref ref-type="bibr" rid="B20">20</xref></sup>demonstraram que seus algoritmos derivados de IA foram eficazes na detec&#231;&#227;o de eventos de FA. No primeiro caso, o design do sensor de ECG foi altamente sens&#237;vel na detec&#231;&#227;o de epis&#243;dios gen&#233;ricos, enquanto no segundo caso, apresentou desempenho satisfat&#243;rio na avalia&#231;&#227;o de curto prazo. Huang et al.<sup><xref ref-type="bibr" rid="B15">15</xref></sup>aprofundaram o desenho do estudo e descobriram que o algoritmo gen&#233;rico de IA era mais preciso na detec&#231;&#227;o de FA em compara&#231;&#227;o ao diagn&#243;stico humano e ao algoritmo espec&#237;fico para detec&#231;&#227;o de FA.</p>
        <p>Na detec&#231;&#227;o de FA em diferentes posi&#231;&#245;es corporais e ap&#243;s exerc&#237;cio, Fu et al.<sup><xref ref-type="bibr" rid="B18">18</xref></sup>destacaram alta precis&#227;o na detec&#231;&#227;o do evento tanto na posi&#231;&#227;o supina quanto na ortost&#225;tica, bem como ap&#243;s a realiza&#231;&#227;o de exerc&#237;cios aer&#243;bicos, com valores de precis&#227;o variando de 96,4% a 98,2%. Em rela&#231;&#227;o &#224; dura&#231;&#227;o, Noseworthy et al.<sup><xref ref-type="bibr" rid="B19">19</xref></sup> e Li et al.<sup><xref ref-type="bibr" rid="B16">16</xref></sup> detectaram FA em avalia&#231;&#245;es diagn&#243;sticas com dura&#231;&#227;o de 30 segundos, 6 minutos, 24 horas e at&#233; 4 semanas.</p>
        <p>Uma fronteira da aplica&#231;&#227;o da IA em diversos campos &#233; a compreens&#227;o dos resultados obtidos. No contexto da detec&#231;&#227;o de FA, &#233; extremamente importante detectar fatores morfol&#243;gicos que expliquem a sa&#237;da do algoritmo. Para este fim, Yang et al.,<sup><xref ref-type="bibr" rid="B21">21</xref></sup>e Jo et al.<sup><xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref></sup> modelaram e aplicaram algoritmos para isolar caracter&#237;sticas-chave da FA. Yang et al.<sup><xref ref-type="bibr" rid="B21">21</xref></sup>tentaram quantificar mudan&#231;as temporais e espaciais no tra&#231;ado de ECG para detec&#231;&#227;o de FA; no entanto, o processamento e a extra&#231;&#227;o de caracter&#237;sticas resultaram em um modelo impreciso. Jo et al.,<sup><xref ref-type="bibr" rid="B22">22</xref></sup> por outro lado, demonstraram um modelo capaz de classificar a arritmia com precis&#227;o usando dados de valida&#231;&#227;o externa. Em um estudo subsequente, eles conseguiram mostrar um modelo que superou a autointerpreta&#231;&#227;o de um aparelho de ECG e identificou o racioc&#237;nio por tr&#225;s da conclus&#227;o do achado.<sup><xref ref-type="bibr" rid="B23">23</xref></sup> Portanto, estes &#250;ltimos autores demonstraram sucesso tanto na detec&#231;&#227;o do evento de FA quanto na explica&#231;&#227;o das raz&#245;es do resultado, com alta sensibilidade (92,8%) e especificidade (95%).<sup><xref ref-type="bibr" rid="B23">23</xref></sup></p>
        <p>A detec&#231;&#227;o de FA por meio de <italic>smartwatch</italic> &#233; uma ferramenta presente nas publica&#231;&#245;es analisadas (4 de 13 publica&#231;&#245;es). Na compara&#231;&#227;o realizada por Wasserlauf et al.<sup><xref ref-type="bibr" rid="B13">13</xref></sup> entre o monitor card&#237;aco invasivo (MCI) e o <italic>smartwatch</italic>, foi demonstrada alta sensibilidade para detec&#231;&#227;o de FA na &#250;ltima hora e na avalia&#231;&#227;o da dura&#231;&#227;o do epis&#243;dio arr&#237;tmico pelo dispositivo vest&#237;vel. Em contraste, o MCI apresentou maior valor preditivo positivo para FA em epis&#243;dios com dura&#231;&#227;o superior a uma hora. Da mesma forma, Chen et al.<sup><xref ref-type="bibr" rid="B20">20</xref></sup>compararam o desempenho entre o <italic>smartwatch</italic> e o ECG de 12 deriva&#231;&#245;es. Como resultado, encontraram boa sensibilidade (96,6%), especificidade (98%) e precis&#227;o (97,5%) na detec&#231;&#227;o de FA de curto prazo pelo dispositivo equipado com o algoritmo de IA.</p>
        <p>V&#228;liaho et al.<sup><xref ref-type="bibr" rid="B14">14</xref></sup>propuseram um estudo comparativo entre a detec&#231;&#227;o de ritmo de FA e ritmo sinusal utilizando smartwatches integrados a dois modelos de algoritmo de ML. Na detec&#231;&#227;o de FA, eles relataram alta sensibilidade (91,7%; VPP: 97,5%), enquanto na detec&#231;&#227;o de ritmo sinusal, relataram sensibilidade significativa (99,4%) e especificidade (98,1%). Os autores identificaram os casos de FA de in&#237;cio mais recente, mas declararam consider&#225;vel confiabilidade na detec&#231;&#227;o de epis&#243;dios de FA, independentemente de sua cronologia, com altas taxas de sensibilidade e especificidade.</p>
      </sec>
      <sec>
        <title>Arritmias induzidas por drogas</title>
        <p>Por fim, com foco na detec&#231;&#227;o de arritmias induzidas por drogas, Maille et al.<sup><xref ref-type="bibr" rid="B12">12</xref></sup>compararam dados do intervalo QT obtidos de <italic>smartwatches</italic> e ECGs de 12 deriva&#231;&#245;es. Neste estudo, os autores destacaram que, apesar da variabilidade no intervalo QT corrigido (QTc), houve concord&#226;ncia razo&#225;vel entre os resultados gerados pela IA do <italic>smartwatch</italic> e os relat&#243;rios de ECGs convencionais. Com base nessa descoberta, os autores enfatizaram que o monitoramento de pacientes por meio de <italic>smartwatches</italic> oferece vantagens distintas, incluindo a capacidade de prever, detectar e prevenir arritmias potencialmente fatais.</p>
        <p>Considerando os diversos resultados obtidos nesta revis&#227;o sistem&#225;tica, fica evidente que algoritmos derivados de IA podem detectar padr&#245;es m&#250;ltiplos, sutis e n&#227;o lineares em um ECG. Consistente com a hip&#243;tese previamente confirmada por Attia et al.,<sup><xref ref-type="bibr" rid="B27">27</xref></sup> essas redes demonstram maior sensibilidade na detec&#231;&#227;o de FA, mesmo em ECGs com ritmo sinusal normal. Os achados est&#227;o alinhados com os relatados na literatura m&#233;dica atual, como a meta-an&#225;lise de Feeny et al.<sup><xref ref-type="bibr" rid="B28">28</xref></sup>no diagn&#243;stico de FA usando um algoritmo derivado de IA, que destaca altas taxas de sensibilidade (94%) e especificidade (96%).</p>
        <p>Entretanto, apesar da diversidade de achados nesta revis&#227;o sistem&#225;tica, h&#225; limita&#231;&#245;es metodol&#243;gicas na indexa&#231;&#227;o e disponibiliza&#231;&#227;o dos resultados nas bases de dados e nos desfechos dos dados da literatura elencados. As limita&#231;&#245;es relacionadas &#224; metodologia deste artigo est&#227;o associadas &#224; operacionaliza&#231;&#227;o por um &#250;nico operador e ao espa&#231;o temporal da pesquisa. Em rela&#231;&#227;o &#224;s bases de dados, a potencial indexa&#231;&#227;o inadequada dos artigos pode suprimir resultados com crit&#233;rios de inclus&#227;o positivos para a revis&#227;o. Em rela&#231;&#227;o &#224;s limita&#231;&#245;es dos estudos elencados neste trabalho, o universo amostral e os delineamentos dos estudos s&#227;o diversos e heterog&#234;neos, contribuindo para vieses de sele&#231;&#227;o, confirma&#231;&#227;o e confus&#227;o decorrentes de predi&#231;&#245;es oferecidas pela m&#225;quina que carregam vieses de seu modelo de treinamento.</p>
        <p>Por outro lado, as tecnologias, em geral, podem levar &#224; exclus&#227;o de grupos socialmente desfavorecidos, impedindo-os de acessar as funcionalidades oferecidas. Apesar da alta taxa de precis&#227;o na detec&#231;&#227;o de arritmias por dispositivos m&#243;veis vest&#237;veis, estes requerem conex&#227;o com outro dispositivo, geralmente um <italic>smartphone</italic>, para maximizar sua usabilidade e funcionalidade. Por outro lado, essas tecnologias permitem o monitoramento de popula&#231;&#245;es vulner&#225;veis e indiv&#237;duos em regi&#245;es remotas que podem enfrentar disparidades no acesso a cuidados m&#233;dicos.<sup><xref ref-type="bibr" rid="B29">29</xref></sup></p>
        <p>Assim, por ser uma &#225;rea especializada e em r&#225;pida expans&#227;o, a ci&#234;ncia de dados aplicada &#224; sa&#250;de requer estudos multic&#234;ntricos sobre os benef&#237;cios, as limita&#231;&#245;es metodol&#243;gicas e os impactos cl&#237;nicos e sociais das tecnologias no processo diagn&#243;stico. Ao mesmo tempo, h&#225; um enorme escopo para a implementa&#231;&#227;o de dispositivos baseados em algoritmos de IA na popula&#231;&#227;o para triagem, diagn&#243;stico e subsequente tratamento precoce de arritmias potencialmente limitantes ou amea&#231;adoras &#224; vida. A otimiza&#231;&#227;o desses processos &#233; imperativa tanto do ponto de vista m&#233;dico quanto socioecon&#244;mico, pois pode levar &#224; redu&#231;&#227;o de custos com hospitaliza&#231;&#245;es e &#224; melhoria da qualidade de vida. A aplicabilidade de recursos tecnol&#243;gicos &#224; pr&#225;tica da medicina baseada em evid&#234;ncias por meio do fornecimento cont&#237;nuo de dados em larga escala para melhorar a precis&#227;o dos dispositivos eletr&#244;nicos de diagn&#243;stico, algoritmos e conduta profissional.</p>
      </sec>
    </sec>
    <sec sec-type="conclusions">
      <title>Conclus&#227;o</title>
      <p>Com base nos resultados apresentados, a aplica&#231;&#227;o, validade e viabilidade de modelos de ML no diagn&#243;stico de arritmias card&#237;acas representam uma &#225;rea promissora e em r&#225;pido avan&#231;o. Esses modelos demonstraram alta precis&#227;o, sensibilidade e especificidade na detec&#231;&#227;o de anormalidades do ritmo card&#237;aco, particularmente FA. No entanto, sua generaliza&#231;&#227;o para outras popula&#231;&#245;es permanece uma limita&#231;&#227;o, assim como o potencial para vieses introduzidos por tend&#234;ncias nos conjuntos de dados de treinamento. Al&#233;m disso, a falta de estudos comparativos avaliando o desempenho da intelig&#234;ncia humana versus IA apresenta desafios para a valida&#231;&#227;o completa de m&#233;todos algor&#237;tmicos como ferramentas confi&#225;veis na pr&#225;tica m&#233;dica de rotina. Assim, &#224; medida que esse campo especializado continua a se expandir, estudos multic&#234;ntricos s&#227;o essenciais para avaliar os benef&#237;cios, as restri&#231;&#245;es metodol&#243;gicas e os impactos cl&#237;nicos e sociais dessas tecnologias no processo diagn&#243;stico.</p>
    </sec>
  </body>
  <back>
    <ack>
      <title>Agradecimentos</title>
      <p>Dr. Alex Cleber Improta-Caria recebe bolsa de p&#243;s-doutorado, que &#233; financiada pela Funda&#231;&#227;o de Amparo &#224; Pesquisa do Estado de S&#227;o Paulo (FAPESP), Brasil. N&#250;mero do Processo n&#186; 2022/02339-4; #2024/17783-2.</p>
    </ack>
    <ref-list>
      <title>Refer&#234;ncias</title>
      <ref id="B1">
        <label>1</label>
        <mixed-citation>1. World Health Organization. Cardiovascular Diseases (CVDs). Geneva: World Health Organization; 2021.</mixed-citation>
        <element-citation publication-type="report">
          <person-group person-group-type="author">
            <collab>World Health Organization</collab>
          </person-group>
          <source>Cardiovascular Diseases (CVDs)</source>
          <publisher-loc>Geneva</publisher-loc>
          <publisher-name>World Health Organization</publisher-name>
          <year>2021</year>
        </element-citation>
      </ref>
      <ref id="B2">
        <label>2</label>
        <mixed-citation>2. Tse G. Mechanisms of Cardiac Arrhythmias. J Arrhythm. 2016;32(2):75-81. doi: 10.1016/j.joa.2015.11.003.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Tse</surname>
              <given-names>G</given-names>
            </name>
          </person-group>
          <article-title>Mechanisms of Cardiac Arrhythmias</article-title>
          <source>J Arrhythm</source>
          <year>2016</year>
          <volume>32</volume>
          <issue>2</issue>
          <fpage>75</fpage>
          <lpage>81</lpage>
          <pub-id pub-id-type="doi">10.1016/j.joa.2015.11.003</pub-id>
        </element-citation>
      </ref>
      <ref id="B3">
        <label>3</label>
        <mixed-citation>3. Jha CK. Automated Cardiac Arrhythmia Detection Techniques: A Comprehensive Review for Prospective Approach. Comput Methods Biomech Biomed Engin. 2024:1-16. doi: 10.1080/10255842.2024.2332942.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Jha</surname>
              <given-names>CK</given-names>
            </name>
          </person-group>
          <article-title>Automated Cardiac Arrhythmia Detection Techniques: A Comprehensive Review for Prospective Approach</article-title>
          <source>Comput Methods Biomech Biomed Engin</source>
          <year>2024</year>
          <fpage>1</fpage>
          <lpage>16</lpage>
          <pub-id pub-id-type="doi">10.1080/10255842.2024.2332942</pub-id>
        </element-citation>
      </ref>
      <ref id="B4">
        <label>4</label>
        <mixed-citation>4. Pereira TMC, Sebastiao R, Conceicao RC, Sencadas V. A Review on Intelligent Systems for ECG Analysis: From Flexible Sensing Technology to Machine Learning. IEEE J Biomed Health Inform. 2025;29(5):3398-413. doi: 10.1109/JBHI.2024.3508545.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Pereira</surname>
              <given-names>TMC</given-names>
            </name>
            <name>
              <surname>Sebastiao</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Conceicao</surname>
              <given-names>RC</given-names>
            </name>
            <name>
              <surname>Sencadas</surname>
              <given-names>V</given-names>
            </name>
          </person-group>
          <article-title>A Review on Intelligent Systems for ECG Analysis: From Flexible Sensing Technology to Machine Learning</article-title>
          <source>IEEE J Biomed Health Inform</source>
          <year>2025</year>
          <volume>29</volume>
          <issue>5</issue>
          <fpage>3398</fpage>
          <lpage>3413</lpage>
          <pub-id pub-id-type="doi">10.1109/JBHI.2024.3508545</pub-id>
        </element-citation>
      </ref>
      <ref id="B5">
        <label>5</label>
        <mixed-citation>5. Hoffmann J, Mahmood S, Fogou PS, George N, Raha S, Safi S. A Survey on Machine Learning Approaches to ECG Processing. Poznan: IEEE; 2020. doi: 10.23919/SPA50552.2020.9241283.</mixed-citation>
        <element-citation publication-type="book">
          <person-group person-group-type="author">
            <name>
              <surname>Hoffmann</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Mahmood</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Fogou</surname>
              <given-names>PS</given-names>
            </name>
            <name>
              <surname>George</surname>
              <given-names>N</given-names>
            </name>
            <name>
              <surname>Raha</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Safi</surname>
              <given-names>S</given-names>
            </name>
          </person-group>
          <source>A Survey on Machine Learning Approaches to ECG Processing</source>
          <publisher-loc>Poznan</publisher-loc>
          <publisher-name>IEEE</publisher-name>
          <year>2020</year>
          <pub-id pub-id-type="doi">10.23919/SPA50552.2020.9241283</pub-id>
        </element-citation>
      </ref>
      <ref id="B6">
        <label>6</label>
        <mixed-citation>6. Dorado-D&#237;az PI, Sampedro-G&#243;mez J, Vicente-Palacios V, S&#225;nchez PL. Applications of Artificial Intelligence in Cardiology. The Future is Already Here. Rev Esp Cardiol. 2019;72(12):1065-75. doi: 10.1016/j.rec.2019.05.014.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Dorado-D&#237;az</surname>
              <given-names>PI</given-names>
            </name>
            <name>
              <surname>Sampedro-G&#243;mez</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Vicente-Palacios</surname>
              <given-names>V</given-names>
            </name>
            <name>
              <surname>S&#225;nchez</surname>
              <given-names>PL</given-names>
            </name>
          </person-group>
          <article-title>Applications of Artificial Intelligence in Cardiology. The Future is Already Here</article-title>
          <source>Rev Esp Cardiol</source>
          <year>2019</year>
          <volume>72</volume>
          <issue>12</issue>
          <fpage>1065</fpage>
          <lpage>1075</lpage>
          <pub-id pub-id-type="doi">10.1016/j.rec.2019.05.014</pub-id>
        </element-citation>
      </ref>
      <ref id="B7">
        <label>7</label>
        <mixed-citation>7. Shinde PP, Shah S. A Review of Machine Learning and Deep Learning Applications. Pune: IEEE; 2018. doi: 10.1109/ICCUBEA.2018.8697857.</mixed-citation>
        <element-citation publication-type="book">
          <person-group person-group-type="author">
            <name>
              <surname>Shinde</surname>
              <given-names>PP</given-names>
            </name>
            <name>
              <surname>Shah</surname>
              <given-names>S</given-names>
            </name>
          </person-group>
          <source>A Review of Machine Learning and Deep Learning Applications</source>
          <publisher-loc>Pune</publisher-loc>
          <publisher-name>IEEE</publisher-name>
          <year>2018</year>
          <pub-id pub-id-type="doi">10.1109/ICCUBEA.2018.8697857</pub-id>
        </element-citation>
      </ref>
      <ref id="B8">
        <label>8</label>
        <mixed-citation>8. Lopez-Jimenez F, Attia Z, Arruda-Olson AM, Carter R, Chareonthaitawee P, Jouni H, et al. Artificial Intelligence in Cardiology: Present and Future. Mayo Clin Proc. 2020;95(5):1015-39. doi: 10.1016/j.mayocp.2020.01.038.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Lopez-Jimenez</surname>
              <given-names>F</given-names>
            </name>
            <name>
              <surname>Attia</surname>
              <given-names>Z</given-names>
            </name>
            <name>
              <surname>Arruda-Olson</surname>
              <given-names>AM</given-names>
            </name>
            <name>
              <surname>Carter</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Chareonthaitawee</surname>
              <given-names>P</given-names>
            </name>
            <name>
              <surname>Jouni</surname>
              <given-names>H</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Artificial Intelligence in Cardiology: Present and Future</article-title>
          <source>Mayo Clin Proc</source>
          <year>2020</year>
          <volume>95</volume>
          <issue>5</issue>
          <fpage>1015</fpage>
          <lpage>1039</lpage>
          <pub-id pub-id-type="doi">10.1016/j.mayocp.2020.01.038</pub-id>
        </element-citation>
      </ref>
      <ref id="B9">
        <label>9</label>
        <mixed-citation>9. Shameer K, Johnson KW, Glicksberg BS, Dudley JT, Sengupta PP. Machine Learning in Cardiovascular Medicine: Are We There Yet? Heart. 2018;104(14):1156-64. doi: 10.1136/heartjnl-2017-311198.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Shameer</surname>
              <given-names>K</given-names>
            </name>
            <name>
              <surname>Johnson</surname>
              <given-names>KW</given-names>
            </name>
            <name>
              <surname>Glicksberg</surname>
              <given-names>BS</given-names>
            </name>
            <name>
              <surname>Dudley</surname>
              <given-names>JT</given-names>
            </name>
            <name>
              <surname>Sengupta</surname>
              <given-names>PP</given-names>
            </name>
          </person-group>
          <article-title>Machine Learning in Cardiovascular Medicine: Are We There Yet?</article-title>
          <source>Heart</source>
          <year>2018</year>
          <volume>104</volume>
          <issue>14</issue>
          <fpage>1156</fpage>
          <lpage>1164</lpage>
          <pub-id pub-id-type="doi">10.1136/heartjnl-2017-311198</pub-id>
        </element-citation>
      </ref>
      <ref id="B10">
        <label>10</label>
        <mixed-citation>10. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. Int J Surg. 2010;8(5):336-41. doi: 10.1016/j.ijsu.2010.02.007.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Moher</surname>
              <given-names>D</given-names>
            </name>
            <name>
              <surname>Liberati</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Tetzlaff</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Altman</surname>
              <given-names>DG</given-names>
            </name>
            <collab>PRISMA Group</collab>
          </person-group>
          <article-title>Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement</article-title>
          <source>Int J Surg</source>
          <year>2010</year>
          <volume>8</volume>
          <issue>5</issue>
          <fpage>336</fpage>
          <lpage>341</lpage>
          <pub-id pub-id-type="doi">10.1016/j.ijsu.2010.02.007</pub-id>
        </element-citation>
      </ref>
      <ref id="B11">
        <label>11</label>
        <mixed-citation>11. Bos JM, Attia ZI, Albert DE, Noseworthy PA, Friedman PA, Ackerman MJ. Use of Artificial Intelligence and Deep Neural Networks in Evaluation of Patients with Electrocardiographically Concealed Long QT Syndrome From the Surface 12-Lead Electrocardiogram. JAMA Cardiol. 2021;6(5):532-8. doi: 10.1001/jamacardio.2020.7422.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Bos</surname>
              <given-names>JM</given-names>
            </name>
            <name>
              <surname>Attia</surname>
              <given-names>ZI</given-names>
            </name>
            <name>
              <surname>Albert</surname>
              <given-names>DE</given-names>
            </name>
            <name>
              <surname>Noseworthy</surname>
              <given-names>PA</given-names>
            </name>
            <name>
              <surname>Friedman</surname>
              <given-names>PA</given-names>
            </name>
            <name>
              <surname>Ackerman</surname>
              <given-names>MJ</given-names>
            </name>
          </person-group>
          <article-title>Use of Artificial Intelligence and Deep Neural Networks in Evaluation of Patients with Electrocardiographically Concealed Long QT Syndrome From the Surface 12-Lead Electrocardiogram</article-title>
          <source>JAMA Cardiol</source>
          <year>2021</year>
          <volume>6</volume>
          <issue>5</issue>
          <fpage>532</fpage>
          <lpage>538</lpage>
          <comment>doi: 10.1001/jamacardio.2020.7422</comment>
        </element-citation>
      </ref>
      <ref id="B12">
        <label>12</label>
        <mixed-citation>12. Maille B, Wilkin M, Million M, Ress&#233;guier N, Franceschi F, Koutbi-Franceschi L, et al. Smartwatch Electrocardiogram and Artificial Intelligence for Assessing Cardiac-Rhythm Safety of Drug Therapy in the COVID-19 Pandemic. The QT-logs study. Int J Cardiol. 2021;331:333-9. doi: 10.1016/j.ijcard.2021.01.002.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Maille</surname>
              <given-names>B</given-names>
            </name>
            <name>
              <surname>Wilkin</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>Million</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>Ress&#233;guier</surname>
              <given-names>N</given-names>
            </name>
            <name>
              <surname>Franceschi</surname>
              <given-names>F</given-names>
            </name>
            <name>
              <surname>Koutbi-Franceschi</surname>
              <given-names>L</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Smartwatch Electrocardiogram and Artificial Intelligence for Assessing Cardiac-Rhythm Safety of Drug Therapy in the COVID-19 Pandemic. The QT-logs study</article-title>
          <source>Int J Cardiol</source>
          <year>2021</year>
          <volume>331</volume>
          <fpage>333</fpage>
          <lpage>339</lpage>
          <pub-id pub-id-type="doi">10.1016/j.ijcard.2021.01.002</pub-id>
        </element-citation>
      </ref>
      <ref id="B13">
        <label>13</label>
        <mixed-citation>13. Wasserlauf J, You C, Patel R, Valys A, Albert D, Passman R. Smartwatch Performance for the Detection and Quantification of Atrial Fibrillation. Circ Arrhythm Electrophysiol. 2019;12(6):e006834. doi: 10.1161/CIRCEP.118.006834.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Wasserlauf</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>You</surname>
              <given-names>C</given-names>
            </name>
            <name>
              <surname>Patel</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Valys</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Albert</surname>
              <given-names>D</given-names>
            </name>
            <name>
              <surname>Passman</surname>
              <given-names>R</given-names>
            </name>
          </person-group>
          <article-title>Smartwatch Performance for the Detection and Quantification of Atrial Fibrillation</article-title>
          <source>Circ Arrhythm Electrophysiol</source>
          <year>2019</year>
          <volume>12</volume>
          <issue>6</issue>
          <elocation-id>e006834</elocation-id>
          <pub-id pub-id-type="doi">10.1161/CIRCEP.118.006834</pub-id>
        </element-citation>
      </ref>
      <ref id="B14">
        <label>14</label>
        <mixed-citation>14. V&#228;liaho ES, Kuoppa P, Lipponen JA, Martikainen TJ, J&#228;ntti H, Rissanen TT, et al. Wrist Band Photoplethysmography in Detection of Individual Pulses in Atrial Fibrillation and Algorithm-Based Detection of Atrial Fibrillation. Europace. 2019;21(7):1031-8. doi: 10.1093/europace/euz060.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>V&#228;liaho</surname>
              <given-names>ES</given-names>
            </name>
            <name>
              <surname>Kuoppa</surname>
              <given-names>P</given-names>
            </name>
            <name>
              <surname>Lipponen</surname>
              <given-names>JA</given-names>
            </name>
            <name>
              <surname>Martikainen</surname>
              <given-names>TJ</given-names>
            </name>
            <name>
              <surname>J&#228;ntti</surname>
              <given-names>H</given-names>
            </name>
            <name>
              <surname>Rissanen</surname>
              <given-names>TT</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Wrist Band Photoplethysmography in Detection of Individual Pulses in Atrial Fibrillation and Algorithm-Based Detection of Atrial Fibrillation</article-title>
          <source>Europace</source>
          <year>2019</year>
          <volume>21</volume>
          <issue>7</issue>
          <fpage>1031</fpage>
          <lpage>1038</lpage>
          <pub-id pub-id-type="doi">10.1093/europace/euz060</pub-id>
        </element-citation>
      </ref>
      <ref id="B15">
        <label>15</label>
        <mixed-citation>15. Huang S, Zhao T, Liu C, Qin A, Dong S, Yuan B, et al. Portable Device Improves the Detection of Atrial Fibrillation after Ablation. Int Heart J. 2021;62(4):786-91. doi: 10.1536/ihj.21-067.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Huang</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Zhao</surname>
              <given-names>T</given-names>
            </name>
            <name>
              <surname>Liu</surname>
              <given-names>C</given-names>
            </name>
            <name>
              <surname>Qin</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Dong</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Yuan</surname>
              <given-names>B</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Portable Device Improves the Detection of Atrial Fibrillation after Ablation</article-title>
          <source>Int Heart J</source>
          <year>2021</year>
          <volume>62</volume>
          <issue>4</issue>
          <fpage>786</fpage>
          <lpage>791</lpage>
          <conf-loc>doi</conf-loc>
          <pub-id pub-id-type="doi">10.1536/ihj.21-067</pub-id>
        </element-citation>
      </ref>
      <ref id="B16">
        <label>16</label>
        <mixed-citation>16. Zhu L, Nathan V, Kuang J, Kim J, Avram R, Olgin J, et al. Atrial Fibrillation Detection and Atrial Fibrillation Burden Estimation via Wearables. IEEE J Biomed Health Inform. 2022;26(5):2063-74. doi: 10.1109/JBHI.2021.3131984.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Zhu</surname>
              <given-names>L</given-names>
            </name>
            <name>
              <surname>Nathan</surname>
              <given-names>V</given-names>
            </name>
            <name>
              <surname>Kuang</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Kim</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Avram</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Olgin</surname>
              <given-names>J</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Atrial Fibrillation Detection and Atrial Fibrillation Burden Estimation via Wearables</article-title>
          <source>IEEE J Biomed Health Inform</source>
          <year>2022</year>
          <volume>26</volume>
          <issue>5</issue>
          <fpage>2063</fpage>
          <lpage>2074</lpage>
          <pub-id pub-id-type="doi">10.1109/JBHI.2021.3131984</pub-id>
        </element-citation>
      </ref>
      <ref id="B17">
        <label>17</label>
        <mixed-citation>17. Jacobsen M, Dembek TA, Ziakos AP, Gholamipoor R, Kobbe G, Kollmann M, et al. Reliable Detection of Atrial Fibrillation with a Medical Wearable during Inpatient Conditions. Sensors. 2020;20(19):5517. doi: 10.3390/s20195517.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Jacobsen</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>Dembek</surname>
              <given-names>TA</given-names>
            </name>
            <name>
              <surname>Ziakos</surname>
              <given-names>AP</given-names>
            </name>
            <name>
              <surname>Gholamipoor</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Kobbe</surname>
              <given-names>G</given-names>
            </name>
            <name>
              <surname>Kollmann</surname>
              <given-names>M</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Reliable Detection of Atrial Fibrillation with a Medical Wearable during Inpatient Conditions</article-title>
          <source>Sensors</source>
          <year>2020</year>
          <volume>20</volume>
          <issue>19</issue>
          <size units="pages">5517</size>
          <pub-id pub-id-type="doi">10.3390/s20195517</pub-id>
        </element-citation>
      </ref>
      <ref id="B18">
        <label>18</label>
        <mixed-citation>18. Fu W, Li R. Diagnostic Performance of a Wearing Dynamic ECG Recorder for Atrial Fibrillation Screening: The HUAMI Heart Study. BMC Cardiovasc Disord. 2021;21(1):558. doi: 10.1186/s12872-021-02363-1.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Fu</surname>
              <given-names>W</given-names>
            </name>
            <name>
              <surname>Li</surname>
              <given-names>R</given-names>
            </name>
          </person-group>
          <article-title>Diagnostic Performance of a Wearing Dynamic ECG Recorder for Atrial Fibrillation Screening: The HUAMI Heart Study</article-title>
          <source>BMC Cardiovasc Disord</source>
          <year>2021</year>
          <volume>21</volume>
          <issue>1</issue>
          <size units="pages">558</size>
          <pub-id pub-id-type="doi">10.1186/s12872-021-02363-1</pub-id>
        </element-citation>
      </ref>
      <ref id="B19">
        <label>19</label>
        <mixed-citation>19. Noseworthy PA, Attia ZI, Behnken EM, Giblon RE, Bews KA, Liu S, et al. Artificial Intelligence-Guided Screening for Atrial Fibrillation Using Electrocardiogram during Sinus Rhythm: A Prospective Non-Randomised Interventional Trial. Lancet. 2022;400(10359):1206-12. doi: 10.1016/S0140-6736(22)01637-3.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Noseworthy</surname>
              <given-names>PA</given-names>
            </name>
            <name>
              <surname>Attia</surname>
              <given-names>ZI</given-names>
            </name>
            <name>
              <surname>Behnken</surname>
              <given-names>EM</given-names>
            </name>
            <name>
              <surname>Giblon</surname>
              <given-names>RE</given-names>
            </name>
            <name>
              <surname>Bews</surname>
              <given-names>KA</given-names>
            </name>
            <name>
              <surname>Liu</surname>
              <given-names>S</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Artificial Intelligence-Guided Screening for Atrial Fibrillation Using Electrocardiogram during Sinus Rhythm: A Prospective Non-Randomised Interventional Trial</article-title>
          <source>Lancet</source>
          <year>2022</year>
          <volume>400</volume>
          <issue>10359</issue>
          <fpage>1206</fpage>
          <lpage>1212</lpage>
          <conf-loc>doi</conf-loc>
          <pub-id pub-id-type="doi">10.1016/S0140-6736(22)01637-3</pub-id>
        </element-citation>
      </ref>
      <ref id="B20">
        <label>20</label>
        <mixed-citation>20. Chen E, Jiang J, Su R, Gao M, Zhu S, Zhou J, et al. A New Smart Wristband Equipped with an Artificial Intelligence Algorithm to Detect Atrial Fibrillation. Heart Rhythm. 2020;17(5 Pt B):847-53. doi: 10.1016/j.hrthm.2020.01.034.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Chen</surname>
              <given-names>E</given-names>
            </name>
            <name>
              <surname>Jiang</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Su</surname>
              <given-names>R</given-names>
            </name>
            <name>
              <surname>Gao</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>Zhu</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Zhou</surname>
              <given-names>J</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>A New Smart Wristband Equipped with an Artificial Intelligence Algorithm to Detect Atrial Fibrillation</article-title>
          <source>Heart Rhythm</source>
          <year>2020</year>
          <volume>17</volume>
          <issue>5 Pt B</issue>
          <fpage>847</fpage>
          <lpage>853</lpage>
          <pub-id pub-id-type="doi">10.1016/j.hrthm.2020.01.034</pub-id>
        </element-citation>
      </ref>
      <ref id="B21">
        <label>21</label>
        <mixed-citation>21. Yang HW, Hsiao CY, Peng YQ, Lin TY, Tsai LW, Lin C, et al. Identification of Patients with Potential Atrial Fibrillation during Sinus Rhythm Using Isolated P Wave Characteristics from 12-Lead ECGs. J Pers Med. 2022;12(10):1608. doi: 10.3390/jpm12101608.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Yang</surname>
              <given-names>HW</given-names>
            </name>
            <name>
              <surname>Hsiao</surname>
              <given-names>CY</given-names>
            </name>
            <name>
              <surname>Peng</surname>
              <given-names>YQ</given-names>
            </name>
            <name>
              <surname>Lin</surname>
              <given-names>TY</given-names>
            </name>
            <name>
              <surname>Tsai</surname>
              <given-names>LW</given-names>
            </name>
            <name>
              <surname>Lin</surname>
              <given-names>C</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Identification of Patients with Potential Atrial Fibrillation during Sinus Rhythm Using Isolated P Wave Characteristics from 12-Lead ECGs</article-title>
          <source>J Pers Med</source>
          <year>2022</year>
          <volume>12</volume>
          <issue>10</issue>
          <size units="pages">1608</size>
          <pub-id pub-id-type="doi">10.3390/jpm12101608</pub-id>
        </element-citation>
      </ref>
      <ref id="B22">
        <label>22</label>
        <mixed-citation>22. Jo YY, Kwon JM, Jeon KH, Cho YH, Shin JH, Lee YJ, et al. Detection and Classification of Arrhythmia Using an Explainable Deep Learning Model. J Electrocardiol. 2021;67:124-32. doi: 10.1016/j.jelectrocard.2021.06.006.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Jo</surname>
              <given-names>YY</given-names>
            </name>
            <name>
              <surname>Kwon</surname>
              <given-names>JM</given-names>
            </name>
            <name>
              <surname>Jeon</surname>
              <given-names>KH</given-names>
            </name>
            <name>
              <surname>Cho</surname>
              <given-names>YH</given-names>
            </name>
            <name>
              <surname>Shin</surname>
              <given-names>JH</given-names>
            </name>
            <name>
              <surname>Lee</surname>
              <given-names>YJ</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Detection and Classification of Arrhythmia Using an Explainable Deep Learning Model</article-title>
          <source>J Electrocardiol</source>
          <year>2021</year>
          <volume>67</volume>
          <fpage>124</fpage>
          <lpage>132</lpage>
          <pub-id pub-id-type="doi">10.1016/j.jelectrocard.2021.06.006</pub-id>
        </element-citation>
      </ref>
      <ref id="B23">
        <label>23</label>
        <mixed-citation>23. Jo YY, Cho Y, Lee SY, Kwon JM, Kim KH, Jeon KH, et al. Explainable Artificial Intelligence to Detect Atrial Fibrillation Using Electrocardiogram. Int J Cardiol. 2021;328:104-10. doi: 10.1016/j.ijcard.2020.11.053.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Jo</surname>
              <given-names>YY</given-names>
            </name>
            <name>
              <surname>Cho</surname>
              <given-names>Y</given-names>
            </name>
            <name>
              <surname>Lee</surname>
              <given-names>SY</given-names>
            </name>
            <name>
              <surname>Kwon</surname>
              <given-names>JM</given-names>
            </name>
            <name>
              <surname>Kim</surname>
              <given-names>KH</given-names>
            </name>
            <name>
              <surname>Jeon</surname>
              <given-names>KH</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Explainable Artificial Intelligence to Detect Atrial Fibrillation Using Electrocardiogram</article-title>
          <source>Int J Cardiol</source>
          <year>2021</year>
          <volume>328</volume>
          <fpage>104</fpage>
          <lpage>110</lpage>
          <pub-id pub-id-type="doi">10.1016/j.ijcard.2020.11.053</pub-id>
        </element-citation>
      </ref>
      <ref id="B24">
        <label>24</label>
        <mixed-citation>24. Hermans BJM, Stoks J, Bennis FC, Vink AS, Garde A, Wilde AAM, et al. Support Vector Machine-Based Assessment of the T-Wave Morphology Improves Long QT Syndrome Diagnosis. Europace. 2018;20(Suppl 3):113-9. doi: 10.1093/europace/euy243.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Hermans</surname>
              <given-names>BJM</given-names>
            </name>
            <name>
              <surname>Stoks</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Bennis</surname>
              <given-names>FC</given-names>
            </name>
            <name>
              <surname>Vink</surname>
              <given-names>AS</given-names>
            </name>
            <name>
              <surname>Garde</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Wilde</surname>
              <given-names>AAM</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Support Vector Machine-Based Assessment of the T-Wave Morphology Improves Long QT Syndrome Diagnosis</article-title>
          <source>Europace</source>
          <year>2018</year>
          <volume>20</volume>
          <supplement>Suppl 3</supplement>
          <fpage>113</fpage>
          <lpage>119</lpage>
          <pub-id pub-id-type="doi">10.1093/europace/euy243</pub-id>
        </element-citation>
      </ref>
      <ref id="B25">
        <label>25</label>
        <mixed-citation>25. Prifti E, Fall A, Davogustto G, Pulini A, Denjoy I, Funck-Brentano C, et al. Deep Learning Analysis of Electrocardiogram for Risk Prediction of Drug-Induced Arrhythmias and Diagnosis of Long QT Syndrome. Eur Heart J. 2021;42(38):3948-61. doi: 10.1093/eurheartj/ehab588.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Prifti</surname>
              <given-names>E</given-names>
            </name>
            <name>
              <surname>Fall</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Davogustto</surname>
              <given-names>G</given-names>
            </name>
            <name>
              <surname>Pulini</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Denjoy</surname>
              <given-names>I</given-names>
            </name>
            <name>
              <surname>Funck-Brentano</surname>
              <given-names>C</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Deep Learning Analysis of Electrocardiogram for Risk Prediction of Drug-Induced Arrhythmias and Diagnosis of Long QT Syndrome</article-title>
          <source>Eur Heart J</source>
          <year>2021</year>
          <volume>42</volume>
          <issue>38</issue>
          <fpage>3948</fpage>
          <lpage>3961</lpage>
          <pub-id pub-id-type="doi">10.1093/eurheartj/ehab588</pub-id>
        </element-citation>
      </ref>
      <ref id="B26">
        <label>26</label>
        <mixed-citation>26. Simon ST, Trinkley KE, Malone DC, Rosenberg MA. Interpretable Machine Learning Prediction of Drug-Induced QT Prolongation: Electronic Health Record Analysis. J Med Internet Res. 2022;24(12):e42163. doi: 10.2196/42163.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Simon</surname>
              <given-names>ST</given-names>
            </name>
            <name>
              <surname>Trinkley</surname>
              <given-names>KE</given-names>
            </name>
            <name>
              <surname>Malone</surname>
              <given-names>DC</given-names>
            </name>
            <name>
              <surname>Rosenberg</surname>
              <given-names>MA</given-names>
            </name>
          </person-group>
          <article-title>Interpretable Machine Learning Prediction of Drug-Induced QT Prolongation: Electronic Health Record Analysis</article-title>
          <source>J Med Internet Res</source>
          <year>2022</year>
          <volume>24</volume>
          <issue>12</issue>
          <elocation-id>e42163</elocation-id>
          <pub-id pub-id-type="doi">10.2196/42163</pub-id>
        </element-citation>
      </ref>
      <ref id="B27">
        <label>27</label>
        <mixed-citation>27. Attia ZI, Harmon DM, Behr ER, Friedman PA. Application of Artificial Intelligence to the Electrocardiogram. Eur Heart J. 2021;42(46):4717-30. doi: 10.1093/eurheartj/ehab649.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Attia</surname>
              <given-names>ZI</given-names>
            </name>
            <name>
              <surname>Harmon</surname>
              <given-names>DM</given-names>
            </name>
            <name>
              <surname>Behr</surname>
              <given-names>ER</given-names>
            </name>
            <name>
              <surname>Friedman</surname>
              <given-names>PA</given-names>
            </name>
          </person-group>
          <article-title>Application of Artificial Intelligence to the Electrocardiogram</article-title>
          <source>Eur Heart J</source>
          <year>2021</year>
          <volume>42</volume>
          <issue>46</issue>
          <fpage>4717</fpage>
          <lpage>4730</lpage>
          <pub-id pub-id-type="doi">10.1093/eurheartj/ehab649</pub-id>
        </element-citation>
      </ref>
      <ref id="B28">
        <label>28</label>
        <mixed-citation>28. Feeny AK, Chung MK, Madabhushi A, Attia ZI, Cikes M, Firouznia M, et al. Artificial Intelligence and Machine Learning in Arrhythmias and Cardiac Electrophysiology. Circ Arrhythm Electrophysiol. 2020;13(8):e007952. doi: 10.1161/CIRCEP.119.007952.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Feeny</surname>
              <given-names>AK</given-names>
            </name>
            <name>
              <surname>Chung</surname>
              <given-names>MK</given-names>
            </name>
            <name>
              <surname>Madabhushi</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>Attia</surname>
              <given-names>ZI</given-names>
            </name>
            <name>
              <surname>Cikes</surname>
              <given-names>M</given-names>
            </name>
            <name>
              <surname>Firouznia</surname>
              <given-names>M</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>Artificial Intelligence and Machine Learning in Arrhythmias and Cardiac Electrophysiology</article-title>
          <source>Circ Arrhythm Electrophysiol</source>
          <year>2020</year>
          <volume>13</volume>
          <issue>8</issue>
          <elocation-id>e007952</elocation-id>
          <pub-id pub-id-type="doi">10.1161/CIRCEP.119.007952</pub-id>
        </element-citation>
      </ref>
      <ref id="B29">
        <label>29</label>
        <mixed-citation>29. Garikapati K, Turnbull S, Bennett RG, Campbell TG, Kanawati J, Wong MS, et al. The Role of Contemporary Wearable and Handheld Devices in the Diagnosis and Management of Cardiac Arrhythmias. Heart Lung Circ. 2022;31(11):1432-49. doi: 10.1016/j.hlc.2022.08.001.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Garikapati</surname>
              <given-names>K</given-names>
            </name>
            <name>
              <surname>Turnbull</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Bennett</surname>
              <given-names>RG</given-names>
            </name>
            <name>
              <surname>Campbell</surname>
              <given-names>TG</given-names>
            </name>
            <name>
              <surname>Kanawati</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Wong</surname>
              <given-names>MS</given-names>
            </name>
            <etal>et al</etal>
          </person-group>
          <article-title>The Role of Contemporary Wearable and Handheld Devices in the Diagnosis and Management of Cardiac Arrhythmias</article-title>
          <source>Heart Lung Circ</source>
          <year>2022</year>
          <volume>31</volume>
          <issue>11</issue>
          <fpage>1432</fpage>
          <lpage>1449</lpage>
          <pub-id pub-id-type="doi">10.1016/j.hlc.2022.08.001</pub-id>
        </element-citation>
      </ref>
    </ref-list>
    <fn-group>
      <fn fn-type="other">
        <label>Vincula&#231;&#227;o acad&#234;mica:</label>
        <p> N&#227;o h&#225; vincula&#231;&#227;o deste estudo a programas de p&#243;s-gradua&#231;&#227;o.</p>
      </fn>
      <fn fn-type="other">
        <label>Aprova&#231;&#227;o &#233;tica e consentimento informado:</label>
        <p> Este artigo n&#227;o cont&#233;m estudos com humanos ou animais realizados por nenhum dos autores.</p>
      </fn>
      <fn fn-type="other">
        <label>Uso de Intelig&#234;ncia Artificial:</label>
        <p> Durante a prepara&#231;&#227;o deste trabalho, o(s) autor(es) usaram ChatGPT. O estudo n&#227;o foi gerado pela intelig&#234;ncia artificial, foi gerado pelo primeiro autor (Paulo Cainan), esse mesmo autor desenvolveu e escreveu o texto inteiro do artigo em portugu&#234;s. Paulo analisou e interpretou os dados e os escreveu em portugu&#234;s. Quando o artigo foi finalizado e revisado, todos escritos em portugu&#234;s, usamos o ChatGPT apenas para tradu&#231;&#227;o de portugu&#234;s para ingl&#234;s. Ap&#243;s o uso desta ferramenta/servi&#231;o, o(s) autor(es) revisaram e editaram o conte&#250;do conforme necess&#225;rio e assumem total responsabilidade pelo conte&#250;do do artigo publicado.</p>
      </fn>
      <fn fn-type="data-availability" specific-use="data-in-article">
        <label>Disponibilidade de Dados:</label>
        <p>Os conte&#250;dos subjacentes ao texto da pesquisa est&#227;o contidos no manuscrito.</p>
      </fn>
      <fn fn-type="financial-disclosure">
        <label>Fontes de financiamento:</label>
        <p> O presente estudo n&#227;o teve fontes de financiamento externas.</p>
      </fn>
    </fn-group>
  </back>
  <sub-article article-type="translation" id="TRen" xml:lang="en">
    <front-stub>
      <article-id pub-id-type="doi">10.36660/abc.20240843i</article-id>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Original Article</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>Applicability of Machine Learning Algorithms in Diagnosis of Atrial Fibrillation and LQTS by Electrocardiogram Interpretation: A Systematic Review</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0009-0000-3187-7911</contrib-id>
          <name>
            <surname>Nascimento</surname>
            <given-names>Paulo Cainan Guimar&#227;es do</given-names>
          </name>
          <role>Conception and design of the research</role>
          <role>Acquisition of data</role>
          <role>Analysis and interpretation of the data</role>
          <role>Statistical analysis</role>
          <role>Obtaining financing and Writing of the manuscript</role>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <name>
            <surname>Martins</surname>
            <given-names>Matthews Silva</given-names>
          </name>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff2001">
            <sup>2</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-8779-1188</contrib-id>
          <name>
            <surname>Improta-Caria</surname>
            <given-names>Alex Cleber</given-names>
          </name>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff3001">
            <sup>3</sup>
          </xref>
        </contrib>
        <contrib contrib-type="author">
          <contrib-id contrib-id-type="orcid">0000-0002-2161-3696</contrib-id>
          <name>
            <surname>Aras</surname>
            <given-names>Roque</given-names>
            <suffix>Junior</suffix>
          </name>
          <role>Conception and design of the research</role>
          <role>Critical revision of the manuscript for content</role>
          <xref ref-type="aff" rid="aff1001">
            <sup>1</sup>
          </xref>
        </contrib>
      </contrib-group>
      <aff id="aff1001">
        <label>1</label>
        <country country="BR">Brazil</country>
        <institution content-type="original">Universidade Federal da Bahia &#8211; Faculdade de Medicina da Bahia, Salvador, BA &#8211; Brazil</institution>
      </aff>
      <aff id="aff2001">
        <label>2</label>
        <country country="BR">Brazil</country>
        <institution content-type="original">Universidade Federal do Esp&#237;rito Santo &#8211; Laborat&#243;rio de Fisiologia Molecular e Intelig&#234;ncia Artificial, Departamento de Ci&#234;ncias Fisiol&#243;gicas, Vit&#243;ria, ES &#8211; Brazil</institution>
      </aff>
      <aff id="aff3001">
        <label>3</label>
        <country country="BR">Brazil</country>
        <institution content-type="original">Universidade de S&#227;o Paulo &#8211; Laborat&#243;rio de Bioqu&#237;mica e Biologia Molecular do Exerc&#237;cio, Escola de Educa&#231;&#227;o F&#237;sica e Esporte, S&#227;o Paulo, SP &#8211; Brazil</institution>
      </aff>
      <author-notes>
        <corresp id="c01001"><label>Mailing Address:</label> Alex Cleber Improta-Caria &#8226; Universidade de S&#227;o Paulo, Laborat&#243;rio de Bioqu&#237;mica e Biologia Molecular do Exerc&#237;cio - Av. Professor Mello Moraes, 65. Postal Code 05508-030, Cidade Universit&#225;ria, S&#227;o Paulo, SP - Brazil E-mail: alexcaria.personal@hotmail.com </corresp>
        <fn fn-type="conflict">
          <label>Potential conflict of interest:</label>
          <p>No potential conflict of interest relevant to this article was reported.</p>
        </fn>
        <fn fn-type="edited-by">
          <label>Editor responsible for the review:</label>
          <p>Marcio Bittencourt</p>
        </fn>
      </author-notes>
      <abstract>
        <title>Abstract</title>
        <sec>
          <title>Background</title>
          <p> Machine Learning (ML) is a type of algorithm that autonomously learns to recognize complex patterns. In the diagnostic context of cardiac arrhythmias, these algorithms have shown significant advancements due to their ability to provide automated interpretation and pattern recognition in electrocardiograms (ECGs).</p>
        </sec>
        <sec>
          <title>Objective</title>
          <p> To analyze and identify the applicability, validity, and feasibility of ML algorithm models in the diagnostic process of cardiac arrhythmias through automated electrocardiogram interpretation.</p>
        </sec>
        <sec>
          <title>Methods</title>
          <p> This systematic literature review was reported according to the PRISMA guidelines. The searches were conducted in the Cochrane Library, EMBASE, LILACS, and PubMed between February 2022 and November 2022. The study period encompasses articles published between 2017 and 2022.</p>
        </sec>
        <sec>
          <title>Results</title>
          <p> The database search yielded 119 results, covering three subthemes: Long QT Syndrome (LQTS), corrected QT interval (QTc), and atrial fibrillation (AF). AF was the most prevalent theme. The sample sizes were quite variable. The outcomes were mostly satisfactory. In the diagnosis of LQTS using Artificial Intelligence (AI), the algorithm outperformed conventional methods in diagnostic distinction. In the evaluation of QTc, there was no difference between the AI-integrated ECG and the conventional ECG. In the diagnosis of AF, the algorithms, models, and devices demonstrated high sensitivity and specificity, along with greater accuracy.</p>
        </sec>
        <sec>
          <title>Conclusion</title>
          <p> ML models in the diagnostic process of cardiac arrhythmias are feasible and rapidly developing. They demonstrate accuracy values between 96.4% and 98.2%, sensitivity between 92.8% and 99.4%, and specificity between 95% and 98.1%, particularly in the diagnosis of atrial fibrillation.</p>
        </sec>
      </abstract>
      <abstract abstract-type="graphical">
        <p>
          <fig id="f01001">
            <label>Central Illustration:</label>
            <caption>
              <title>Applicability of Machine Learning Algorithms in Diagnosis of Atrial Fibrillation and LQTS by Electrocardiogram Interpretation: A Systematic Review</title>
            </caption>
            <attrib>Figure created using Napkin.ai</attrib>
            <graphic xlink:href="1678-4170-abc-122-08-e20240843-40843-gf01-en.tif"/></fig>
        </p>
      </abstract>
      <kwd-group xml:lang="en">
        <title>Keywords</title>
        <kwd>Cardiac Arrhythmias</kwd>
        <kwd>Artificial Intelligence</kwd>
        <kwd>Machine Learning</kwd>
        <kwd>Neural Network</kwd>
        <kwd>Electrocardiography</kwd>
      </kwd-group>
    </front-stub>
    <body>
      <sec sec-type="intro">
        <title>Introduction</title>
        <p>According to the World Health Organization, cardiovascular diseases (CVD) are the leading causes of death worldwide, accounting for 32%.<sup><xref ref-type="bibr" rid="B1">1</xref></sup> Among CVD, arrhythmias are the most common and are characterized by abnormalities in the generation or conduction of electrical impulses and occur due to abnormal heart function, resulting in irregular heart rhythms. The occurrence of arrhythmias may be indolent, but they can cause serious outcomes.<sup><xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref></sup></p>
        <p>The diagnostic approach for cardiac arrhythmias involves two main components: clinical history and electrocardiogram (ECG). The clinical history lists irregular heartbeats, dyspnea, fatigue, dizziness, and, when left untreated, a history of stroke and heart failure. The suspicion can be corroborated by changes in the cardiovascular physical examination through auscultation with a stethoscope and palpation of pulses. The diagnosis is confirmed with the recording of the ECG Holter for 24 hours and subsequent certification of irregular rhythms by a specialist physician. However, the conventional method of diagnosis by ECG Holter, in general, imposes limitations due to the lack of mechanical flexibility to the patient and the retrospective analysis method, without real-time monitoring and follow-up.<sup><xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref></sup></p>
        <p>However, Artificial Intelligence (AI) algorithms and their derivatives have emerged as a reliable method to identify and classify electrocardiographic patterns that may suggest abnormalities.<sup><xref ref-type="bibr" rid="B5">5</xref></sup> Dorado-D&#237;az et al.<sup><xref ref-type="bibr" rid="B6">6</xref></sup> refer to AI as a &#8220;field of computer science that attempts to mimic the human cognitive process, learning capacity, and knowledge storage&#8221;. In Health Sciences, AI algorithms are predominantly applied in prediction, recommendation, and diagnostic support based on Machine Learning (ML) and Deep Learning (DL) models. Conceptually, ML focuses on developing algorithms that enable computer systems to learn without explicit programming. In contrast, DL models derive their predictive capabilities from Artificial Neural Networks (ANNs) with multiple layers of information processing.<sup><xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref></sup></p>
        <p>In this context, ML algorithms are now capable of analyzing ECGs, providing automated interpretations, and predicting associated risks.<sup><xref ref-type="bibr" rid="B8">8</xref></sup> The literature highlights predictions derived from the optimization of procedural steps, such as signal processing, extraction of significant variables, and algorithm classification.<sup><xref ref-type="bibr" rid="B9">9</xref></sup> Despite the intrinsic computational complexity, there are several advances in systems that perform this role.</p>
        <p>Therefore, AI tools can complement medical monitoring, from preventive to rehabilitation levels. The development of ML capable of conducting ECG analysis is essential in the early detection of abnormal cardiac conditions remotely and assisting the diagnostic process of arrhythmias and other health conditions.<sup><xref ref-type="bibr" rid="B5">5</xref></sup> Furthermore, long-term cardiac monitoring, recognition, and classification of arrhythmias based on the analysis of the ECG tracing can be a time-consuming and unnecessarily exhaustive process for the cardiologist. Thus, autonomic arrhythmia detection techniques are useful for such executions. In this scenario, computerized pattern recognition techniques will potentially help provide diagnosis and medical intervention whenever necessary.<sup><xref ref-type="bibr" rid="B3">3</xref></sup></p>
        <p>Considering the above, the study of AI and its derivative technologies in the healthcare field proves to be highly relevant. Therefore, this work aimed, through a systematic review of scientific literature, to explore the applicability, validity, and feasibility of ML models in the diagnostic process of cardiac arrhythmias. The study sought to identify methodological benefits and limitations, potential clinical and social impacts of these technologies, and their implications for the structure of medical education and practice.</p>
      </sec>
      <sec sec-type="methods">
        <title>Methods</title>
        <p>The study design is a systematic literature review conducted using the criteria established by the PRISMA method (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), as detailed in the methodological guidelines for the preparation of systematic reviews and meta-analyses of randomized clinical trials provided by the Brazilian Ministry of Health.<sup><xref ref-type="bibr" rid="B10">10</xref></sup></p>
        <p>For the literature search on the intended topic, the following databases were used: Cochrane Library, Embase, LILACS (Latin American and Caribbean Health Sciences Literature), and PubMed-MEDLINE (Medical Literature Analysis and Retrieval System Online). The search terms were derived from the DeCS/MeSH (Health Sciences Descriptors/Medical Subject Headings) vocabulary, supplemented by entry terms. The descriptors included: arrhythmias; cardiac or cardiac arrhythmia; machine learning; unsupervised machine learning; supervised machine learning; deep learning; neural networks, computer or computer neural networks or computational neural network or computational neural networks; artificial intelligence or AI or computational intelligence; and electrocardiography or ECG or EKG. The search was conducted using the following configuration of descriptors and Boolean operators: ((cardiac arrhythmia) OR (arrhythmia) AND (machine learning)) OR (unsupervised machine learning)) OR (supervised machine learning)) OR (deep learning) OR (neural network) OR (artificial intelligence) AND (electrocardiography).</p>
        <p>During the screening process, studies meeting the following inclusion criteria were considered: 1) research involving human subjects or data obtained from human databases; 2) publications in their final version in peer-reviewed scientific journals; 3) studies of the following types: case-control studies, cohort studies, clinical trials, case reports, and case series; 4) published between 2017-2022; 5) available in Portuguese or English. Exclusion criteria included: 1) publications in languages other than Portuguese or English; 2) studies with experimental descriptions lacking clinical application of the algorithmic model; 3) articles unrelated to the diagnosis of cardiac arrhythmias; 4) articles that do not apply AI models to the diagnosis of cardiac arrhythmias; 5) articles proposing methods for diagnosing cardiac arrhythmias using examinations other than ECG; 6) articles describing the use of AI for diagnosing conditions other than cardiac arrhythmias; 7) other types of publications, such as editorials, guidelines, books, systematic reviews, and meta-analyses; 8) duplicated results across the searched databases.</p>
        <p>The screening process for search results and data extraction was based on the PRISMA method and carried out by a single operator. This process resulted in the creation of a database and a table model, enabling the organization of the retrieved data in configurations aligned with the employed method&#8217;s criteria serving as a precursor to future analyses. Using this methodology, the aim was to evaluate clinical studies addressing the use and applicability of ML algorithms in the diagnostic process of cardiac arrhythmias through the automated interpretation of ECGs.</p>
      </sec>
      <sec sec-type="results">
        <title>Results</title>
        <p>The database search yielded a total of 119 results, including 38 from the Cochrane Library, 44 from Embase, none from LILACS, and 37 from PubMed. After applying the exclusion criteria, 98 publications were excluded for the following reasons: did not evaluate the detection of cardiac arrhythmias (83), did not apply AI algorithms for arrhythmia detection (5), addressed ongoing studies (3), were classified as other types of scientific texts (3), focused on an experimental model (1), described the diagnostic evaluation of arrhythmias using methods other than ECG (1), lacked a full-text version (1), and did not provide the full text in English or Portuguese (1). Consequently, 13 publications were deemed eligible for full and systematic review (<xref ref-type="fig" rid="f02001">Figure 1</xref>).</p>
        <p>
          <fig id="f02001">
            <label>Figure 1</label>
            <caption>
              <title>&#8211; Flowchart of the Systematic Literature Review.</title>
            </caption>
            <graphic xlink:href="1678-4170-abc-122-08-e20240843-40843-gf02-en.tif"/></fig>
        </p>
        <p>The 13 articles deemed eligible were included in the qualitative synthesis (<xref ref-type="table" rid="t1001">Table 1</xref>). The final search was conducted on November 30, 2022. Regarding the countries of origin of the included publications, they represented seven locations: the United States (4), China (3), South Korea (2), Germany (1), France (1), Taiwan (1), and Finland (1). In terms of study design, the articles comprised prospective cohort studies (7), retrospective cohort studies (2), case-control studies (2), and randomized clinical trials (2). All studies had received approval from the ethics committees of their respective institutions.</p>
        <p>
          <table-wrap id="t1001">
            <label>Table 1</label>
            <caption>
              <title>&#8211; Synthesis of Design, Results, and Conclusions - LQTS</title>
            </caption>
            <table frame="hsides" rules="groups">
              <colgroup>
                <col/>
                <col/>
              </colgroup>
              <thead>
                <tr>
                  <th align="left">Author, year, (ref. N&#186;)</th>
                  <th align="left">BOS, 2021<sup><bold>11</bold></sup></th>
                </tr>
              </thead>
              <tbody>
                <tr>
                  <td>Study design</td>
                  <td>Case-control study. Objective: Determine whether AI is capable of distinguishing patients with Long QT Syndrome from patients with a normal QTc using 12-lead ECG. Sample: 2.984 patients previously diagnosed with LQTS or discharged as healthy. Case-control diagnosis was performed using available 12-lead ECGs from 2.059 patients. Main outcome: AI was capable of distinguishing patients with LQTS, persistent QT prolongation, and genetic differentiation.</td>
                </tr>
                <tr>
                  <td>IA model</td>
                  <td>Deep Neural Network.</td>
                </tr>
                <tr>
                  <td>Results</td>
                  <td>AI was able to distinguish between the two populations (patients with LQTS and patients without LQTS). Accuracy: 82.2%; sensitivity: 83.7%; specificity: 80.6%; PPV: 83.2%; and NPV: 81.3%.</td>
                </tr>
                <tr>
                  <td>Conclusion</td>
                  <td>The AI-ECG outperformed conventional ECG in distinguishing patients with LQTS from those discharged without the diagnosis. Additionally, it could differentiate between three genetic subtypes of LQTS.</td>
                </tr>
              </tbody>
            </table>
            <table-wrap-foot>
              <fn id="TFN1001">
                <p>This analysis was conducted using a statistical significance level of 5% (p = 0.05). AI: artificial intelligence; ECG: electrocardiogram; LQTS: Long QT Syndrome.</p>
              </fn>
            </table-wrap-foot>
          </table-wrap>
        </p>
        <p>The analyzed articles focused on three subtopics of cardiac arrhythmology: Long QT Syndrome (LQTS), corrected QT interval (QTc), and atrial fibrillation (AF). In the study on LQTS, the aim was to assess the ability of AI models to detect this syndrome in individuals.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Regarding drug-induced arrhythmia, the researchers sought to evaluate the prevalence of arrhythmia through automated analysis of QTc in individuals undergoing specific therapies.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> AF, on the other hand, provided the highest number of studies, which primarily aimed to assess the accuracy,<sup><xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref></sup> feasibility,<sup><xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref></sup> reliability,<sup><xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref></sup> performance,<sup><xref ref-type="bibr" rid="B17">17</xref></sup> efficiency,<sup><xref ref-type="bibr" rid="B18">18</xref></sup> safety,<sup><xref ref-type="bibr" rid="B18">18</xref></sup> discriminative power,<sup><xref ref-type="bibr" rid="B19">19</xref></sup> sensitivity, specificity, and precision<sup><xref ref-type="bibr" rid="B20">20</xref></sup> of the algorithms and associated devices used to detect AF episodes. However, other studies aimed to explain the decision-making process of ML models in the diagnosis of arrhythmias.<sup><xref ref-type="bibr" rid="B21">21</xref></sup> The sample sizes were highly variable. All studies used patient samples or data collected from medical institution databases associated with the research.</p>
        <p>Analyzing the outcomes, most results were satisfactory and aligned with the initial objectives. For models applicable to Long QT Syndrome (LQTS), the proposed algorithm successfully distinguished between the two populations under investigation: individuals with LQTS and those without it.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Regarding the detection of AF using AI-based algorithms, the articles reported high rates of sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy, and other performance metrics.<sup><xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B23">23</xref></sup> In the assessment of arrhythmias induced by pharmacological therapy, there was concordance between the QTc interval patterns derived from conventional ECGs and wearable device ECGs.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> As for explanatory algorithmic models for arrhythmia diagnostics, the outcomes varied: one study showed a relatively moderate area under the curve (AUC),<sup><xref ref-type="bibr" rid="B21">21</xref></sup> while another demonstrated better performance.<sup><xref ref-type="bibr" rid="B22">22</xref></sup></p>
        <p>Regarding the results and conclusions for diagnosing LQTS using AI, the AI-based ECG surpassed the conventional method in diagnostic distinction capabilities and was able to differentiate genetic subtypes. For the diagnosis of AF using AI, the algorithms, models, and devices demonstrated high sensitivity in detecting AF within the last hour,<sup><xref ref-type="bibr" rid="B13">13</xref></sup> higher sensitivity and specificity compared to conventional management and other comparative algorithms,<sup><xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B17">17</xref></sup> and greater accuracy in detection across different postures<sup><xref ref-type="bibr" rid="B18">18</xref></sup> and times.<sup><xref ref-type="bibr" rid="B19">19</xref></sup> In the context of monitoring drug-induced arrhythmias, there was no variability between AI models and conventional ECGs.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> However, explanatory algorithmic models for arrhythmia diagnostics yielded ambiguous results: one exhibited limitation in detecting AF during sinus rhythm,<sup><xref ref-type="bibr" rid="B21">21</xref></sup> whereas another successfully correlated ECG trace characteristics with arrhythmia diagnosis using AI.<sup><xref ref-type="bibr" rid="B22">22</xref></sup> The outcomes, results, and conclusions are summarized in <xref ref-type="table" rid="t1001">Tables 1</xref>, <xref ref-type="table" rid="t2001">2</xref> and <xref ref-type="table" rid="t3001">3</xref>.</p>
        <p/>
        <p/>
        <p/>
      </sec>
      <sec sec-type="discussion">
        <title>Discussion</title>
        <p>In this systematic review, the automated algorithms for arrhythmia detection covered three major presentations: Long QT Syndrome, other QT interval abnormalities, and AF.</p>
        <sec>
          <title>Long QT syndrome</title>
          <p>In Long QT Syndrome, the study analyzed in this review was able to distinguish populations with an accuracy of 82.2%, presenting a sensitivity of 83.7% (PPV 83.2%) and specificity of 80.6% (NPV 81.3%). This highlights that the application of AI algorithms in diagnosing this arrhythmia demonstrates a satisfactory degree of accuracy, sensitivity, and specificity.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Regarding QT interval abnormalities, the study included in this review showed concordance between the results obtained from the conventional 12-lead ECG and the ECG with an AI algorithm integrated into a wearable device equipped with photoplethysmography (PPG) technology. Therefore, there was minimal difference between the findings of the conventional ECG and those suggested by the AI algorithm.<sup><xref ref-type="bibr" rid="B12">12</xref></sup></p>
          <p>Hermans et al.<sup><xref ref-type="bibr" rid="B24">24</xref></sup> previously researched morphological characteristics that could support the diagnosis of LQTS and other QT interval variations. These authors investigated the aggregated values of morphological markers of the T-wave at baseline and in an extended model. They concluded that T-wave morphology has an added value in distinguishing patients with LQTS from family members with a negative genotype. This morphological characteristic can explain the accuracy in the work of Bos et al.<sup><xref ref-type="bibr" rid="B11">11</xref></sup> Regarding QT interval abnormalities, Prifti et al.<sup><xref ref-type="bibr" rid="B25">25</xref></sup> report ML models capable of predicting patients prone to developing significant drug-induced QT prolongation.</p>
          <p>Corroborating the contributions of Simon et al.<sup><xref ref-type="bibr" rid="B26">26</xref></sup> on electrophysiological changes in the QT interval, interpretability occurs through predictive accuracy. Interpretable methods allow for a thorough inspection of the interactions that lead to arrhythmias in this class, whether due to the expression of a genetic alteration or specific drug interactions. This interpretability is useful in preventing complications and should be considered for integrating predictive modeling into clinical decision-support tools.</p>
        </sec>
        <sec>
          <title>Atrial fibrillation</title>
          <p>Regarding the general detection of AF, characterized by detecting AF events regardless of duration, the individual&#8217;s position, or specific morphological factors of the electrocardiographic trace, the studies showed a high degree of accuracy, sensitivity, and specificity. Wasserlauf et al.<sup><xref ref-type="bibr" rid="B13">13</xref></sup> and Erdong Chen et al.<sup><xref ref-type="bibr" rid="B20">20</xref></sup> demonstrated that their AI-derived algorithms were effective in detecting AF events. In the first case, the ECG sensor design was highly sensitive in detecting generic episodes, while in the second case, it showed satisfactory performance in short-term evaluation. Huang et al.<sup><xref ref-type="bibr" rid="B15">15</xref></sup> deepened their study design and found that the generic AI algorithm was more accurate in detecting AF compared to human diagnosis and the specific algorithm for detecting AF.</p>
          <p>In the detection of AF in different body positions and after exercise, Fu et al.<sup><xref ref-type="bibr" rid="B18">18</xref></sup> highlighted high accuracy in detecting the event in both supine and orthostatic positions, as well as after performing aerobic exercise, with accuracy values ranging from 96.4% to 98.2%. Regarding duration, Noseworthy et al.<sup><xref ref-type="bibr" rid="B19">19</xref></sup> and Li et al.<sup><xref ref-type="bibr" rid="B16">16</xref></sup> detected AF in diagnostic assessments lasting 30 seconds, 6 minutes, 24 hours, and up to 4 weeks.</p>
          <p>One frontier of AI application in various fields is the understanding of the results obtained. In the context of AF detection, it is extremely important to detect morphological factors that explain the algorithm&#8217;s output. To this end, Yang et al.<sup><xref ref-type="bibr" rid="B21">21</xref></sup> and Jo et al.<sup><xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref></sup> modeled and applied algorithms to isolate key features of AF. Yang et al.<sup><xref ref-type="bibr" rid="B21">21</xref></sup> attempted to quantify temporal and spatial changes in the ECG tracing for AF detection; however, the processing and feature extraction resulted in an imprecise model. Jo et al.,<sup><xref ref-type="bibr" rid="B22">22</xref></sup> on the other hand, demonstrated a model capable of accurately classifying the arrhythmia using external validation data. In a subsequent study, they were able to show a model that surpassed the self-interpretation of an ECG machine and identified the reasoning behind the conclusion of the finding.<sup><xref ref-type="bibr" rid="B23">23</xref></sup> Therefore, these latter authors demonstrated success both in detecting the AF event and in explaining the reasons for the result, with high sensitivity (92.8%) and specificity (95%).<sup><xref ref-type="bibr" rid="B23">23</xref></sup></p>
          <p>The detection of AF via smartwatch device is a tool present in the analyzed publications (4 of 13 publications). In the comparison made by Wasserlauf et al.<sup><xref ref-type="bibr" rid="B13">13</xref></sup> between the invasive cardiac monitor (ICM) and the smartwatch, high sensitivity for detecting AF in the last hour and in the assessment of the arrhythmic episode duration by the wearable device was demonstrated. In contrast, the ICM showed a higher positive predictive value for AF in episodes lasting longer than one hour. Similarly, Chen et al.<sup><xref ref-type="bibr" rid="B20">20</xref></sup> compared the performance between the smartwatch and the 12-lead ECG. As a result, they found good sensitivity (96.6%), specificity (98%), and accuracy (97.5%) in the short-term AF detection by the device equipped with the AI algorithm.</p>
          <p>V&#228;liaho et al.<sup><xref ref-type="bibr" rid="B14">14</xref></sup> proposed a comparative study between the detection of AF rhythm and sinus rhythm using smartwatches integrated with two ML algorithm models. In AF detection, they reported high sensitivity (91.7%; PPV: 97.5%), while in detecting sinus rhythm, they reported significant sensitivity (99.4%) and specificity (98.1%). The authors identified the most recent-onset AF cases but stated considerable reliability in detecting AF episodes regardless of their chronology, with high sensitivity and specificity rates.</p>
        </sec>
        <sec>
          <title>Drug-induced arrhythmias</title>
          <p>Lastly, focusing on the detection of drug-induced arrhythmias, Maille et al.<sup><xref ref-type="bibr" rid="B12">12</xref></sup> compared QT interval data obtained from smartwatches and 12-lead ECGs. In this study, the authors highlighted that, despite variability in the corrected QT (QTc) interval, there was reasonable agreement between the outputs generated by the smartwatch&#8217;s AI and the reports from conventional ECGs. Based on this finding, the authors emphasized that patient monitoring through smartwatches offers distinct advantages, including the ability to predict, detect, and prevent potentially fatal arrhythmias.</p>
          <p>Given the diverse results obtained in this systematic review, it is evident that AI-derived algorithms can detect multiple, subtle, and nonlinear patterns in an ECG. Consistent with the hypothesis previously confirmed by Attia et al.,<sup><xref ref-type="bibr" rid="B27">27</xref></sup> these networks demonstrate greater sensitivity in detecting AF, even in ECGs with normal sinus rhythm. The findings align with those reported in the current medical literature, such as the meta-analysis by Feeny et al.<sup><xref ref-type="bibr" rid="B28">28</xref></sup> on the diagnosis of AF using an AI-derived algorithm, which highlights high sensitivity (94%) and specificity (96%) rates.</p>
          <p>However, despite the diversity of findings in this systematic review, there are methodological limitations in the indexing and provision of results in the databases and the outcomes of the listed literature data. The limitations related to the methodology of this article are associated with the operationalization by a single operator and the temporal space of research. Regarding the databases, the potential inadequate indexing of articles may suppress results with positive inclusion criteria for the review. Regarding the limitations of the studies listed in this work, the sample universe and study designs are diverse and heterogeneous, contributing to selection, confirmation, and confusion biases resulting from predictions offered by the machine that carry biases from its training model.</p>
          <p>On the other hand, technologies, in general, can lead to the exclusion of socially disadvantaged groups, preventing them from accessing the functionalities offered. Despite the high accuracy rate in detecting arrhythmias by wearable mobile devices, these require connection to another device, usually a smartphone, to maximize their usability and functionality. On the other hand, these technologies allow the monitoring of vulnerable populations and individuals in remote regions who may face disparities in access to medical care.<sup><xref ref-type="bibr" rid="B29">29</xref></sup></p>
          <p>Thus, as it is a rapidly expanding and specialized area, data science applied to health requires multicenter studies on the benefits, methodological limitations, and clinical and social impacts of technologies in the diagnostic process. At the same time, there is enormous scope for implementing devices based on AI algorithms in the population for screening, diagnosis, and subsequent early treatment of potentially life-limiting or life-threatening arrhythmias. Optimizing these processes is imperative from both a medical and socioeconomic point of view, as they can lead to reduced costs for hospitalizations and improved quality of life. The applicability of technological resources to the practice of evidence-based medicine through the continuous provision of large-scale data to improve the accuracy of electronic diagnostic devices, algorithms, and professional conduct.</p>
        </sec>
      </sec>
      <sec sec-type="conclusions">
        <title>Conclusion</title>
        <p>Based on the presented results, the application, validity, and feasibility of ML models in diagnosing cardiac arrhythmias represent a promising and rapidly advancing area. These models have shown high accuracy, sensitivity, and specificity in detecting cardiac rhythm abnormalities, particularly AF. However, their generalizability to other populations remains a limitation, as does the potential for biases introduced by trends in the training datasets. Additionally, the lack of comparative studies assessing the performance of human intelligence versus AI presents challenges to fully validating algorithmic methods as reliable tools in routine medical practice. Thus, as this specialized field continues to expand, multicenter studies are essential to assess the benefits, methodological constraints, and clinical and societal impacts of these technologies in the diagnostic process.</p>
        <p>
          <table-wrap id="t2001">
            <label>Table 2</label>
            <caption>
              <title>&#8211; Synthesis of Design, Results, and Conclusions &#8211; Atrial Fibrillation</title>
            </caption>
            <table frame="hsides" rules="groups">
              <colgroup>
                <col/>
                <col/>
                <col/>
                <col/>
                <col/>
              </colgroup>
              <thead>
                <tr>
                  <th align="left">Author, year, (ref. n&#186;)</th>
                  <th>Design</th>
                  <th>AI model</th>
                  <th>Results</th>
                  <th>Conclusion</th>
                </tr>
              </thead>
              <tbody>
                <tr>
                  <td>Wasserlauf et al., 2019<xref ref-type="bibr" rid="B13"><sup>13</sup></xref> </td>
                  <td align="center">Prospective cohort study. Objective: Compare the accuracy of an atrial AF-sensitive device with simultaneous recording from a wearable cardiac monitor. Sample: AI training with 7.500 ECG data; cohort validation with 26 patients. Main outcomes: The ECG sensor device demonstrated high sensitivity for detecting AF within the last hour in an ambulatory population.</td>
                  <td align="center">Deep Neural Network</td>
                  <td align="center">Of the 82 episodes detected by the implantable cardiac monitor, the AF-sensitive device identified 80, achieving a sensitivity of 97.5% per episode.</td>
                  <td align="center">The results demonstrate that the wearable device with the ECG sensor, the application, and the investigative algorithm is highly sensitive for detecting AF episodes within the last hour in an ambulatory population and for assessing AF duration when compared with an ICM.</td>
                </tr>
                <tr>
                  <td>Huang et al., 2021<xref ref-type="bibr" rid="B15"><sup>15</sup></xref> </td>
                  <td align="center">Randomized clinical trial. Objective: To evaluate the feasibility and reliability of a self-applied ECG device and monitoring system for detecting AF. Sample: 218 patients previously submitted to ablation, randomized into two groups: BT group monitored by the AI algorithm and TF group monitored by traditional medical follow-up. Main outcomes: The sensitivity and specificity of the AI algorithm were higher than that of the automated AF detection.</td>
                  <td align="center">AI Algorithm</td>
                  <td align="center">Feasibility of the AI algorithm in follow-up: 26.133 ECG records, with detection of 12.6% of AF confirmed in manual review by cardiologists, 14.8% by the automated AF detection algorithm, and 13.2% by the generic AI algorithm. The AI model detected more AF recurrence in paroxysmal AF after ablation (p = 0.0099) but not in persistent AF (p = 0.7910). The sensitivity and specificity of the generic AI algorithm for detecting AF (94.4% and 98.5%) were higher than those of the automatic detection algorithm (90.7% and 96.2%).</td>
                  <td align="center">Follow-up after ablation using the AI algorithm leads to more frequent detection of AF recurrence. The AI algorithm demonstrated improved accuracy of ECG diagnosis and was more effective than traditional strategies.</td>
                </tr>
                <tr>
                  <td>Jacobsen et al., 2020<xref ref-type="bibr" rid="B17"><sup>17</sup></xref> </td>
                  <td align="center">Prospective cohort study. Objective: To evaluate the performance of a wearable medical device employing photoplethysmography technology to detect AF in hospitalized patients with AF. Sample: 102 patients admitted to a German hospital with documented AF. Main outcomes: The device demonstrated better AF detection capability than traditional Holter monitoring.</td>
                  <td align="center">Deep Neural Network</td>
                  <td align="center">The device detected AF episodes in the dataset with a sensitivity of 95.2% and specificity of 92.5%.</td>
                  <td align="center">AF detection by a wearable medical device is a feasible and reliable approach.</td>
                </tr>
                <tr>
                  <td>Fu et al., 2021<xref ref-type="bibr" rid="B18"><sup>18</sup></xref> </td>
                  <td align="center">Prospective cohort study. Objective: To assess the efficacy and safety of AF detection and provide a reliable, non-invasive method for the screening and management of AF in daily practice. Sample size: 114 patients. Main outcomes: Wearable dynamic ECG with AI algorithm can detect AF and analyze heart rhythms in different postures and after exercise.</td>
                  <td align="center">AI algorithm</td>
                  <td align="center">The method detected the occurrence of AF in the supine position with an accuracy of 96.4%, sensitivity of 92.4%, specificity of 100%, PPV of 100%, and NPV of 93.8%. The detection of AF in the orthostatic position achieved an accuracy of 98.2%, sensitivity of 100%, specificity of 100%, PPV of 100%, and NPV of 96.8%.</td>
                  <td align="center">The dynamic variable ECG can detect AF in heart rhythms across different postures and after exercise.</td>
                </tr>
                <tr>
                  <td>Noseworthy et al., 2022<xref ref-type="bibr" rid="B19"><sup>19</sup></xref> </td>
                  <td align="center">Non-randomized clinical trial. Objective: To evaluate the discriminative power of the AI model in adults. Assess the effectiveness of the AI-guided AF screening strategy compared to usual care. Sample: 1.003 patients. Main outcomes: The AI model detected AF rhythms and duration.</td>
                  <td align="center">AI algorithm</td>
                  <td align="center">Fifty-four patients were newly diagnosed with AF lasting 30 seconds or more and stratified into low and high-risk groups. A similar pattern was observed in groups with AF lasting 6 minutes or more and 24 hours or more. High risk was associated with a higher AF burden (mean 4.97% for low risk; mean 20.32% for high risk).</td>
                  <td align="center">The device is capable of detecting AF at various time points and stratifying risk. The results support a low-cost, massively scalable, patient-centered AI-guided screening program.</td>
                </tr>
                <tr>
                  <td>Yang et al., 2022<xref ref-type="bibr" rid="B21"><sup>21</sup></xref> </td>
                  <td align="center">Prospective cohort study. Objective: Proposition of a feature-based ML approach that provides explainable results and can be trained on a feasibly sized database. Sample: Digital data from 10- and 12-lead ECG recordings from a Taiwanese hospital. Main outcomes: The model was able to isolate temporal and spatial features of P waves and predict AF patterns.</td>
                  <td align="center">Machine Learning model</td>
                  <td align="center">AF prediction using machine learning P-wave extraction model: the model achieved the highest AUC (0.64). The model showed moderate specificity (0.71) but low sensitivity (0.47).</td>
                  <td align="center">The data show limited power for the proposed features in detecting patients with AF during RS. With feature processing and extraction, it is possible to build models to identify potential AF, but the model is still imprecise.</td>
                </tr>
                <tr>
                  <td>Zhu et al., 2022<xref ref-type="bibr" rid="B16"><sup>16</sup></xref> </td>
                  <td align="center">Prospective cohort study. Objective: Develop, implement, and validate a photoplethysmography-based AF detection algorithm for smartwatches in patients diagnosed with AF. Sample: 204 participants. Main outcome: The algorithm enabled passive detection of AF based on photoplethysmography using a wearable device.</td>
                  <td align="center">AI algorithm</td>
                  <td align="center">The algorithm detected AF in 148 out of 204 patients over 4 weeks. Sensitivity: 87.8%; specificity: 97.4%; decision power (whether sinus rhythm or indeterminate): 67.8%.</td>
                  <td align="center">The algorithm demonstrated the feasibility of accurately detecting AF and non-invasively estimating the burden of AF.</td>
                </tr>
                <tr>
                  <td>Jo et al., 2021<xref ref-type="bibr" rid="B22"><sup>22</sup></xref> </td>
                  <td align="center">Retrospective study. Objective: To develop and validate an explainable deep learning model (XDM) based on a neural network-backed decision tree (NBET) for arrhythmia classification. Sample: A set of 72.740 ECG records from 42.880 patients from a South Korean hospital between 2017 and 2020.</td>
                  <td align="center">Deep Learning model</td>
                  <td align="center">The F1 score of the external validation of the XDM was: NSR (Normal Sinus Rhythm): 0.990; AF or Atrial Flutter: 0.955; Supraventricular Tachycardia: 0.777; Complete AV Block: 0.828; Pacemaker Rhythm: 0.671; Aggregate accuracy: 0.844. AF and atrial flutter were strongly correlated with characteristics such as the presence of the P wave and irregularity, while complete AV block showed a strong correlation with AV dissociation.</td>
                  <td align="center">The model accurately classifies arrhythmias in various ECG formats using external validation datasets.</td>
                </tr>
                <tr>
                  <td>Jo et al., 2021<xref ref-type="bibr" rid="B23"><sup>23</sup></xref> </td>
                  <td align="center">Retrospective study. Objective: Propose a method to construct an explainable AI (XAI) model. Develop and validate an explainable deep learning model for detecting AF using various ECG formats. Sample: 115.485 ECGs for AI model development; 12.914 ECGs for training; 99.965 for external validation. Main outcomes: The model outperformed the machine&#8217;s self-interpretation of ECG in all validation datasets. It also verified the reason (P wave and irregularity) for its conclusion.</td>
                  <td align="center">Deep Learning model</td>
                  <td align="center">The AI model showed a sensitivity of 0.928, specificity of 0.950, PPV of 0.615, NPV of 0.993, and accuracy of 0.948 in detecting AF.</td>
                  <td align="center">The results indicated that the XAI methodology could be used to describe the reasoning behind the model&#8217;s decision to detect AF with high performance.</td>
                </tr>
                <tr>
                  <td>Chen et al., 2020<xref ref-type="bibr" rid="B20"><sup>20</sup></xref> </td>
                  <td align="center">Prospective cohort study. Objective: Evaluate the sensitivity, specificity, and precision of a smartwatch with PPG and ECG that uses an AI algorithm for AF detection. Sample: 401 patients between May and June 2019. Main outcomes: The sensitivity, specificity, and accuracy of the wristband equipped with PPG, ECG, and AI algorithms demonstrate satisfactory performance in the short term.</td>
                  <td align="center">AI algorithm</td>
                  <td align="center">The device with PPG showed the following results for AF detection: sensitivity: 80.00%; specificity: 96.81%; accuracy: 90.52%; PPV: 93.75%; and NPV: 89.01%.</td>
                  <td align="center">The device provided good sensitivity, specificity, and accuracy in determining the presence of AF.</td>
                </tr>
                <tr>
                  <td>V&#228;liaho et al., 2019<xref ref-type="bibr" rid="B14"><sup>14</sup></xref> </td>
                  <td align="center">Case-control study. Objective: Evaluate the accuracy of a commercially available PPG wristband in detecting individual pulses in AF and assess the reliability of two commonly used AF detection algorithms based on PPG. Sample: 213 total patients, 106 for the FA group and 107 for the SR group (control). Main findings: The wristband equipped with PPG, using two AF detection algorithms, was able to diagnose AF with high sensitivity and specificity.</td>
                  <td align="center">AI algorithm</td>
                  <td align="center">The detection of AF by PPG showed a sensitivity of 96.2% and specificity of 98.1% with the AF evidence algorithm and a sensitivity of 95.3% and specificity of 98.1% with COSEn.</td>
                  <td align="center">Photoplethysmography wristbands can help detect asymptomatic or &#8220;silent&#8221; cases of AF.</td>
                </tr>
              </tbody>
            </table>
            <table-wrap-foot>
              <fn id="TFN2001">
                <p>All analyses were conducted using a statistical significance level of 5% (p = 0.05). AI: artificial intelligence; ECG: electrocardiogram; FA: atrial fibrillation; MCI: invasive heart monitor; VPP: positive predictive value; VPN: negative predictive value; XAI: explainable AI; PPG: photoplethysmography; AV: atrioventricular.</p>
              </fn>
            </table-wrap-foot>
          </table-wrap>
        </p>
        <p>
          <table-wrap id="t3001">
            <label>Table 3</label>
            <caption>
              <title>&#8211; Synthesis of Design, Results, and Conclusions - Drug-induced arrhythmia</title>
            </caption>
            <table frame="hsides" rules="groups">
              <colgroup>
                <col/>
                <col/>
              </colgroup>
              <thead>
                <tr>
                  <th align="left">Author, year, (ref. n&#186;)</th>
                  <th align="left">Maille et al., 2021<sup><bold>12</bold></sup></th>
                </tr>
              </thead>
              <tbody>
                <tr>
                  <td>Study design</td>
                  <td>Prospective cohort study. Objective: Compare the QTc calculated using this algorithm on single-lead ECGs from a smartwatch with the QTc measured on conventional 12-lead ECGs in early-stage COVID-19 patients treated with the HCQ-AZM regimen. Sample: a total of 85 patients; 76 under medication regimen. Main outcomes: In the measurements, there was agreement between the QTc interval standard of the 12-lead ECG and the QTc measured in the corresponding SW-ECG.</td>
                </tr>
                <tr>
                  <td>IA model</td>
                  <td>Convolutional neural network.</td>
                </tr>
                <tr>
                  <td>Results</td>
                  <td>Agreement between the standard 12-lead QTc interval measured manually in the lead II or V5 on days 0, 6, and 10 and the AI-QTc measured in the corresponding SW-ECG.</td>
                </tr>
                <tr>
                  <td>Conclusion</td>
                  <td>There is little significant difference between the AI and conventional ECG methods for diagnosing drug-induced arrhythmia.</td>
                </tr>
              </tbody>
            </table>
            <table-wrap-foot>
              <fn id="TFN3001">
                <p>This analysis was conducted using a statistical significance level of 5% (p = 0.05). AI: artificial intelligence; ECG: electrocardiogram.</p>
              </fn>
            </table-wrap-foot>
          </table-wrap>
        </p>
      </sec>
    </body>
    <back>
      <ack>
        <title>Acknowledgments</title>
        <p>Dr. Alex Cleber Improta-Caria is a recipient of a postdoctoral fellowship funded by the S&#227;o Paulo Research Foundation (FAPESP), Brazil. Process Number #2022/02339-4; #2024/17783-2.</p>
      </ack>
      <fn-group>
        <fn fn-type="other">
          <label>Study association:</label>
          <p> This study is not associated with any thesis or dissertation work.</p>
        </fn>
        <fn fn-type="other">
          <label>Ethics approval and consent to participate:</label>
          <p> This article does not contain any studies with human participants or animals performed by any of the authors.</p>
        </fn>
        <fn fn-type="other">
          <label>Use of Artificial Intelligence:</label>
          <p> During the preparation of this work, the author(s) used ChatGPT. The study was not generated by artificial intelligence, it was generated by the first author (Paulo Cainan), this same author developed and wrote the entire text of the article in Portuguese. Paulo analyzed and interpreted the data and wrote it in Portuguese. When the article was finalized and revised, all written in Portuguese, we used chatGPT only for translation from Portuguese to English. After using this tool/service, the author(s) reviewed and edited the content as needed and take full responsibility for the content of the published article.</p>
        </fn>
        <fn fn-type="data-availability" specific-use="data-in-article">
          <label>Data Availability:</label>
          <p> The underlying content of the research text is contained within the manuscript.</p>
        </fn>
        <fn fn-type="financial-disclosure">
          <label>Sources of funding:</label>
          <p> There were no external funding sources for this study.</p>
        </fn>
      </fn-group>
    </back>
  </sub-article>
</article>
