Taehee Kim, Heejun Shin, Yong Sub Song, Jong Hyuk Lee, Hyungjin Kim, Dongmyung Shin
{"title":"人工智能软件用于在无症状人群中检测胸片上未怀疑的肺癌。","authors":"Taehee Kim, Heejun Shin, Yong Sub Song, Jong Hyuk Lee, Hyungjin Kim, Dongmyung Shin","doi":"10.1093/radadv/umae032","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Detecting clinically unsuspected lung cancer on chest radiographs is challenging. Artificial intelligence (AI) software that performs comparably to radiologists may serve as a useful tool.</p><p><strong>Purpose: </strong>To evaluate the lung cancer detection performance of a commercially available AI software and to that of humans in a healthy population.</p><p><strong>Materials and methods: </strong>This retrospective study used chest radiographs from the Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer screening trial in the United States between November 1993 and July 2001 with pathological cancer diagnosis follow-up to 2009 (median 11.3 years). The software's predictions were compared to the PLCO radiologists' reads. A reader study was performed with a subset comparing the software to 3 experienced radiologists.</p><p><strong>Results: </strong>The analysis included 24 370 individuals (mean age 62.6 <math><mo>±</mo></math> 5.4; median age 62; cancer rate 2%), with 213 individuals (mean age 63.6 <math><mo>±</mo></math> 5.5; median age 63; cancer rate 46%) for the reader study. AI achieved higher specificity (0.910 for AI vs. 0.803 for radiologists, <i>P</i> < .001), positive predictive value (0.054 for AI vs. 0.032 for radiologists, <i>P</i> < .001), but lower sensitivity (0.326 for AI vs. 0.412 for radiologists, <i>P</i> = .001) than the PLCO radiologists. When we calibrated the sensitivity of AI to match it with the PLCO radiologists, AI had higher specificity (0.815 for AI vs. 0.803 for radiologists, <i>P</i> < .001). In the reader study, AI achieved higher sensitivity than readers 1 and 3 (0.608 for AI vs. 0.588 for reader 1, <i>P</i> = .789 vs. 0.588 for reader 3, <i>P</i> = .803) but lower specificity than reader 1 (0.888 for AI vs. 0.905 for reader 1, <i>P</i> = .814). Compared to reader 2, AI showed higher specificity (0.888 for AI vs. 0.819 for reader 2, <i>P</i> = .153) but lower sensitivity (0.888 for AI vs. 0.905 for reader 1, <i>P</i> = .814).</p><p><strong>Conclusion: </strong>AI detects lung cancer on chest radiographs among asymptomatic individuals with comparable performance to experienced radiologists.</p>","PeriodicalId":519940,"journal":{"name":"Radiology advances","volume":"2 1","pages":"umae032"},"PeriodicalIF":0.0000,"publicationDate":"2024-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429178/pdf/","citationCount":"0","resultStr":"{\"title\":\"Artificial intelligence software for detecting unsuspected lung cancer on chest radiographs in an asymptomatic population.\",\"authors\":\"Taehee Kim, Heejun Shin, Yong Sub Song, Jong Hyuk Lee, Hyungjin Kim, Dongmyung Shin\",\"doi\":\"10.1093/radadv/umae032\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>Detecting clinically unsuspected lung cancer on chest radiographs is challenging. Artificial intelligence (AI) software that performs comparably to radiologists may serve as a useful tool.</p><p><strong>Purpose: </strong>To evaluate the lung cancer detection performance of a commercially available AI software and to that of humans in a healthy population.</p><p><strong>Materials and methods: </strong>This retrospective study used chest radiographs from the Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer screening trial in the United States between November 1993 and July 2001 with pathological cancer diagnosis follow-up to 2009 (median 11.3 years). The software's predictions were compared to the PLCO radiologists' reads. A reader study was performed with a subset comparing the software to 3 experienced radiologists.</p><p><strong>Results: </strong>The analysis included 24 370 individuals (mean age 62.6 <math><mo>±</mo></math> 5.4; median age 62; cancer rate 2%), with 213 individuals (mean age 63.6 <math><mo>±</mo></math> 5.5; median age 63; cancer rate 46%) for the reader study. AI achieved higher specificity (0.910 for AI vs. 0.803 for radiologists, <i>P</i> < .001), positive predictive value (0.054 for AI vs. 0.032 for radiologists, <i>P</i> < .001), but lower sensitivity (0.326 for AI vs. 0.412 for radiologists, <i>P</i> = .001) than the PLCO radiologists. When we calibrated the sensitivity of AI to match it with the PLCO radiologists, AI had higher specificity (0.815 for AI vs. 0.803 for radiologists, <i>P</i> < .001). In the reader study, AI achieved higher sensitivity than readers 1 and 3 (0.608 for AI vs. 0.588 for reader 1, <i>P</i> = .789 vs. 0.588 for reader 3, <i>P</i> = .803) but lower specificity than reader 1 (0.888 for AI vs. 0.905 for reader 1, <i>P</i> = .814). Compared to reader 2, AI showed higher specificity (0.888 for AI vs. 0.819 for reader 2, <i>P</i> = .153) but lower sensitivity (0.888 for AI vs. 0.905 for reader 1, <i>P</i> = .814).</p><p><strong>Conclusion: </strong>AI detects lung cancer on chest radiographs among asymptomatic individuals with comparable performance to experienced radiologists.</p>\",\"PeriodicalId\":519940,\"journal\":{\"name\":\"Radiology advances\",\"volume\":\"2 1\",\"pages\":\"umae032\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-12-09\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12429178/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Radiology advances\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1093/radadv/umae032\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"2025/1/1 0:00:00\",\"PubModel\":\"eCollection\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Radiology advances","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1093/radadv/umae032","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/1/1 0:00:00","PubModel":"eCollection","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
背景：在胸片上发现临床未怀疑的肺癌是具有挑战性的。表现堪比放射科医生的人工智能（AI）软件可能会成为一个有用的工具。目的：评估市售人工智能软件的肺癌检测性能，以及对健康人群的检测性能。材料和方法：本回顾性研究使用1993年11月至2001年7月在美国进行的前列腺、肺、结直肠和卵巢（PLCO）癌症筛查试验的胸部x线片，病理癌症诊断随访至2009年（中位11.3年）。该软件的预测与PLCO放射科医生的读数进行了比较。进行了一项读者研究，将该软件与3名经验丰富的放射科医生进行了比较。结果：分析纳入24370例（平均年龄62.6±5.4岁，中位年龄62岁，癌症发病率2%），读者研究纳入213例（平均年龄63.6±5.5岁，中位年龄63岁，癌症发病率46%）。AI获得了更高的特异性（AI为0.910，放射科医师为0.803）。001)比PLCO放射科医生。当我们校准AI的灵敏度以使其与PLCO放射科医生相匹配时，AI具有更高的特异性(AI为0.815，放射科医生为0.803,P P =。789 vs.读者3 0.588,P =。803)，但特异性低于阅读器1 （AI为0.888，阅读器1为0.905,P = .814）。与读写器2相比，AI具有更高的特异性(0.888 vs 0.819, P =；153)但灵敏度较低（AI为0.888，阅读器1为0.905,P = .814）。结论：人工智能在无症状个体的胸片上检测出肺癌，其表现与经验丰富的放射科医生相当。
Artificial intelligence software for detecting unsuspected lung cancer on chest radiographs in an asymptomatic population.
Background: Detecting clinically unsuspected lung cancer on chest radiographs is challenging. Artificial intelligence (AI) software that performs comparably to radiologists may serve as a useful tool.
Purpose: To evaluate the lung cancer detection performance of a commercially available AI software and to that of humans in a healthy population.
Materials and methods: This retrospective study used chest radiographs from the Prostate, Lung, Colorectal, and Ovarian (PLCO) cancer screening trial in the United States between November 1993 and July 2001 with pathological cancer diagnosis follow-up to 2009 (median 11.3 years). The software's predictions were compared to the PLCO radiologists' reads. A reader study was performed with a subset comparing the software to 3 experienced radiologists.
Results: The analysis included 24 370 individuals (mean age 62.6 5.4; median age 62; cancer rate 2%), with 213 individuals (mean age 63.6 5.5; median age 63; cancer rate 46%) for the reader study. AI achieved higher specificity (0.910 for AI vs. 0.803 for radiologists, P < .001), positive predictive value (0.054 for AI vs. 0.032 for radiologists, P < .001), but lower sensitivity (0.326 for AI vs. 0.412 for radiologists, P = .001) than the PLCO radiologists. When we calibrated the sensitivity of AI to match it with the PLCO radiologists, AI had higher specificity (0.815 for AI vs. 0.803 for radiologists, P < .001). In the reader study, AI achieved higher sensitivity than readers 1 and 3 (0.608 for AI vs. 0.588 for reader 1, P = .789 vs. 0.588 for reader 3, P = .803) but lower specificity than reader 1 (0.888 for AI vs. 0.905 for reader 1, P = .814). Compared to reader 2, AI showed higher specificity (0.888 for AI vs. 0.819 for reader 2, P = .153) but lower sensitivity (0.888 for AI vs. 0.905 for reader 1, P = .814).
Conclusion: AI detects lung cancer on chest radiographs among asymptomatic individuals with comparable performance to experienced radiologists.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
