Thomas G Day, Samuel Budd, Jeremy Tan, Jacqueline Matthew, Emily Skelton, Victoria Jowett, David Lloyd, Alberto Gomez, Jo V Hajnal, Reza Razavi, Bernhard Kainz, John M Simpson
{"title":"使用人工智能的超声产前诊断左心发育不全综合征:与目前的筛查计划相比,表现如何?","authors":"Thomas G Day, Samuel Budd, Jeremy Tan, Jacqueline Matthew, Emily Skelton, Victoria Jowett, David Lloyd, Alberto Gomez, Jo V Hajnal, Reza Razavi, Bernhard Kainz, John M Simpson","doi":"10.1002/pd.6445","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Artificial intelligence (AI) has the potential to improve prenatal detection of congenital heart disease. We analysed the performance of the current national screening programme in detecting hypoplastic left heart syndrome (HLHS) to compare with our own AI model.</p><p><strong>Methods: </strong>Current screening programme performance was calculated from local and national sources. AI models were trained using four-chamber ultrasound views of the fetal heart, using a ResNet classifier.</p><p><strong>Results: </strong>Estimated current fetal screening programme sensitivity and specificity for HLHS were 94.3% and 99.985%, respectively. Depending on calibration, AI models to detect HLHS were either highly sensitive (sensitivity 100%, specificity 94.0%) or highly specific (sensitivity 93.3%, specificity 100%). Our analysis suggests that our highly sensitive model would generate 45,134 screen positive results for a gain of 14 additional HLHS cases. Our highly specific model would be associated with two fewer detected HLHS cases, and 118 fewer false positives.</p><p><strong>Conclusion: </strong>If used independently, our AI model performance is slightly worse than the performance level of the current screening programme in detecting HLHS, and this performance is likely to deteriorate further when used prospectively. This demonstrates that collaboration between humans and AI will be key for effective future clinical use.</p>","PeriodicalId":20387,"journal":{"name":"Prenatal Diagnosis","volume":null,"pages":null},"PeriodicalIF":2.7000,"publicationDate":"2024-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Prenatal diagnosis of hypoplastic left heart syndrome on ultrasound using artificial intelligence: How does performance compare to a current screening programme?\",\"authors\":\"Thomas G Day, Samuel Budd, Jeremy Tan, Jacqueline Matthew, Emily Skelton, Victoria Jowett, David Lloyd, Alberto Gomez, Jo V Hajnal, Reza Razavi, Bernhard Kainz, John M Simpson\",\"doi\":\"10.1002/pd.6445\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>Artificial intelligence (AI) has the potential to improve prenatal detection of congenital heart disease. We analysed the performance of the current national screening programme in detecting hypoplastic left heart syndrome (HLHS) to compare with our own AI model.</p><p><strong>Methods: </strong>Current screening programme performance was calculated from local and national sources. AI models were trained using four-chamber ultrasound views of the fetal heart, using a ResNet classifier.</p><p><strong>Results: </strong>Estimated current fetal screening programme sensitivity and specificity for HLHS were 94.3% and 99.985%, respectively. Depending on calibration, AI models to detect HLHS were either highly sensitive (sensitivity 100%, specificity 94.0%) or highly specific (sensitivity 93.3%, specificity 100%). Our analysis suggests that our highly sensitive model would generate 45,134 screen positive results for a gain of 14 additional HLHS cases. Our highly specific model would be associated with two fewer detected HLHS cases, and 118 fewer false positives.</p><p><strong>Conclusion: </strong>If used independently, our AI model performance is slightly worse than the performance level of the current screening programme in detecting HLHS, and this performance is likely to deteriorate further when used prospectively. This demonstrates that collaboration between humans and AI will be key for effective future clinical use.</p>\",\"PeriodicalId\":20387,\"journal\":{\"name\":\"Prenatal Diagnosis\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":2.7000,\"publicationDate\":\"2024-06-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Prenatal Diagnosis\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1002/pd.6445\",\"RegionNum\":2,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2023/9/30 0:00:00\",\"PubModel\":\"Epub\",\"JCR\":\"Q2\",\"JCRName\":\"GENETICS & HEREDITY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Prenatal Diagnosis","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1002/pd.6445","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2023/9/30 0:00:00","PubModel":"Epub","JCR":"Q2","JCRName":"GENETICS & HEREDITY","Score":null,"Total":0}
Prenatal diagnosis of hypoplastic left heart syndrome on ultrasound using artificial intelligence: How does performance compare to a current screening programme?
Background: Artificial intelligence (AI) has the potential to improve prenatal detection of congenital heart disease. We analysed the performance of the current national screening programme in detecting hypoplastic left heart syndrome (HLHS) to compare with our own AI model.
Methods: Current screening programme performance was calculated from local and national sources. AI models were trained using four-chamber ultrasound views of the fetal heart, using a ResNet classifier.
Results: Estimated current fetal screening programme sensitivity and specificity for HLHS were 94.3% and 99.985%, respectively. Depending on calibration, AI models to detect HLHS were either highly sensitive (sensitivity 100%, specificity 94.0%) or highly specific (sensitivity 93.3%, specificity 100%). Our analysis suggests that our highly sensitive model would generate 45,134 screen positive results for a gain of 14 additional HLHS cases. Our highly specific model would be associated with two fewer detected HLHS cases, and 118 fewer false positives.
Conclusion: If used independently, our AI model performance is slightly worse than the performance level of the current screening programme in detecting HLHS, and this performance is likely to deteriorate further when used prospectively. This demonstrates that collaboration between humans and AI will be key for effective future clinical use.
期刊介绍:
Prenatal Diagnosis welcomes submissions in all aspects of prenatal diagnosis with a particular focus on areas in which molecular biology and genetics interface with prenatal care and therapy, encompassing: all aspects of fetal imaging, including sonography and magnetic resonance imaging; prenatal cytogenetics, including molecular studies and array CGH; prenatal screening studies; fetal cells and cell-free nucleic acids in maternal blood and other fluids; preimplantation genetic diagnosis (PGD); prenatal diagnosis of single gene disorders, including metabolic disorders; fetal therapy; fetal and placental development and pathology; development and evaluation of laboratory services for prenatal diagnosis; psychosocial, legal, ethical and economic aspects of prenatal diagnosis; prenatal genetic counseling