Evaluation of Detection and Diagnostic Efficiency of Pulmonary Nodules by Chest CT Based on Artificial Intelligence
-
摘要: 目的：评估基于深度学习的人工智能（AI）软件在胸部CT肺结节检出及良恶性诊断的价值。方法：收集2018年6月至2020年4月本院经手术确诊的肺结节患者172例，共切除204枚结节。将172例术前高分辨胸部CT图像导入人工智能识别系统，分别采用人工智能和影像医师阅片检出肺结节及良恶性诊断，对比两种阅片方法的敏感度、阳性预测值及假阳性结节个数。以病理结果为诊断金标准，对比AI与影像医师在恶性肺结节诊断中的敏感度、特异度及受试者工作特征（ROC）曲线下面积。结果：172例胸部高分辨CT共检出796枚真结节；AI与影像医师检出结节的敏感度分别为90.5%和75.0%，阳性预测值分别为74.5%和99.7%，假阳性结节总数分别为247个和2个。204枚经手术切除的结节中，AI、影像医师及AI联合影像医师诊断恶性肺结节的敏感度分别为93.3%、78.5%和98.6%，特异度分别为34.8%、79.7%和79.7%；AI、影像医师及AI联合影像医师诊断恶性肺结节的ROC曲线下面积分别为0.641、0.791和0.819。结论：AI检测肺结节的敏感度明显高于影像医师，但AI假阳性率亦较高；AI联合影像医师诊断恶性肺结节效能高于AI或影像医师单独诊断；建议AI联合影像医师共同检出肺结节和良恶性诊断，可以降低漏诊率、提高诊断正确率。Abstract: Objective: To evaluate the value of artificial intelligence (AI) based on deep learning in the detection and diagnosis of chest CT pulmonary nodules. Methods: A total of 172 patients with pulmonary nodules diagnosed by surgery in our hospital from June 2018 to April 2020 were collected, and a total of 204 nodules were surgically resected. 172 cases of preoperative high-resolution chest CT images were imported into the artificial intelligence recognition system, then artificial intelligence (AI) and radiologists detected and diagnosed the pulmonary nodules respectively, and we compared the sensitivity, positive predictive value and false positive rate between the two diagnostic methods. Using the pathological results as the gold standard for diagnosis, the sensitivity, specificity and area under the receiver operating characteristic (ROC) curve of AI and radiologists in the diagnosis of malignant pulmonary nodules were compared. Results: A total of 796 true nodules were detected in the 172 cases of high-resolution chest CT images. The sensitivity of AI and radiologists in detecting nodules was respectively 90.5% and 75.0%, the positive predictive value was respectively 74.5% and 99.7%, and the number of false positive nodules was respectively 247 and 2.As for the surgically resected 204 nodules, the sensitivity of AI, radiologist and AI-radiologist combination in diagnosing malignant pulmonary nodules was respectively 93.3%, 78.5% and 98.6%, the specificity was respectively 34.8%, 79.7% and 79.7%, and the area under the receiver operating characteristic (ROC) curve for the diagnosis of malignant pulmonary nodules by AI, radiologist and AI-radiologist combination was respectively 0.641, 0.791 and 0.819. Conclusion: The sensitivity of AI in detecting pulmonary nodules is significantly higher than that of radiologists, yet the false positive rate of AI is also higher. The efficiency of AI-radiologist combination in diagnosing malignant pulmonary nodules is higher than that of AI or radiologists separately. Therefore, we recommend that AI and radiologists should collaborate to detect and diagnose pulmonary nodules, which can reduce the rate of missed diagnosis and improve the diagnosis accuracy.
-
Keywords:
- artificial intelligence /
- pulmonary nodules /
- computer tomography /
- deep learning
-
-
[1] TORRE L A, BRAY F, SIEGEL R L, et al. Global cancer statistics, 2012[J]. CA-A Cancer Journal for Clinicians, 2015, 65(2):87-108. DOI: 10.3322/caac.21262.
[2] HINTON G E, OSINDERO S, TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation, 2006, 18(7):1527-1554. DOI: 10.1162/neco.2006.18.7.1527.
[3] LE CUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553):436-444. DOI: 10.1038/nature14539.
[4] SIEGEL R L, MILLER K D, JEMAL A. Cancerstatistics 2018[J]. CA-A Cancer Journal for Clinicians, 2018, 68(1):7-30. DOI: 10.3322/caac.21442.
[5] WILSHIRE C L, LOUIE B E, MANNING K A, et al. Radiologic evaluation of small lepidic adenocarcinomas to guide decision making in surgical resection[J]. Annals of Thoracic Surgery, 2015, 100(3):979-988. DOI: 10.1016/j.athoracsur.2015.04.030.
[6] ZHOU J, LI Y, ZHANG Y, et al. Solitary ground-glass opacity nodules of stage IA pulmonary adenocarcinoma:Combination of F-FDG PET/CT and high-resolution computed tomography features to predict invasive adenocarcinoma[J]. Oncotarget, 2017, 8(14):23312-23321. DOI: 10.18632/oncotarget.15577.
[7] 李欣菱, 郭芳芳, 周振, 等. 基于深度学习的人工智能胸部CT肺结节检测效能评估[J]. 中国肺癌杂志, 2019, 22(6):336-340. DOI:10.3779/j.issn.1009-3419.2019.06.02. LI X L, GUO F F, ZHOU Z, et al. Performance of deep-learning-based artificial intelligence on detection of pulmonary nodules in chest CT[J]. Chinese Journal of Lung Cancer, 2019, 22(6):336-340. DOI:10.3779/j.issn.1009-3419.2019.06.02. (in Chinese).
[8] 张超, 王浩, 孟祥峰, 等. CT成像剂量对人工智能算法性能的影响分析[J]. 中国医疗设备, 2020, 35(4):68-70. DOI:10.3969/j.issn.1674-1633.2020.04.018. ZHANG C, WANG H, MENG X F, et al. Impact of CT dose on performance of artificial intelligence algorithm[J]. China Medical Devices, 2020, 35(4):68-70. DOI:10.3969/j.issn.1674-1633. 2020.04.018. (in Chinese).
[9] STARNES S L. Commentary:Artificial intelligence for pulmonary nodules:Machines to diagnosis cancer[J]. Journal of Thoracic and Cardiovascular Surgery, 2021, S0022-5223(21):00353-00356. DOI: 10.1016/j.jtcvs.2021.02.039.
[10] BLANC D, RACINE V, KHALIL A, et al. Artificial intelligence solution to classify pulmonary nodules on CT[J]. Diagnostic and Interventional Imaging, 2020, 101(12):803-810. DOI: 10.1016/j.diii.2020.10.004.
[11] DAS M, MIIHLENBRUCH G, MAHNKEN A H, et al. Small pulmonary nodules:Effect of two computer-aided detection systems on radiologistperformance[J]. Radiology, 2006, 241:564-571. DOI: 10.1148/radiol.2412051139.
[12] GU Y, LU X, YANG L, et al. Automatic lung nodule detection using a 3D deep convolutional neural network combined with a multi-scale prediction strategy in chest CTs[J]. Computers in Biology and Medicine, 2018, 103:220-231. DOI: 10.1016/j.compbiomed.2018.10.011.
[13] TORRES E L, FIORINA E, PENNAZIO F, et al. Large scale validation of the M5L lung CAD on heterogeneous CT datasets[J]. Medical Physics, 2015, 42(4):1477-1489. DOI: 10.1118/1.4907970.
[14] SETIO A A, CIOMPI F, LITJENS G, et al. Pulmonary nodule detection in CT images:False positive reduction using multi-view convolutional networks[J]. IEEE Transactions on Medical Imaging, 2016, 35(5):1160-1169. DOI: 10.1109/TMI.2016.2536809.
[15] 李甜, 李晓东, 刘敬禹. 人工智能辅助诊断肺结节的临床价值研究[J]. 中国全科医学杂志, 2020, 23(7):828-831 , 836. DOI:10.12114/j.issn.1007-9572.2020.00.052. LI T, LI X D, LIU J Y. Clinical value of artificial intelligence in the diagnosis of pulmonary nodules[J]. Chinese General Practice, 2020, 23(7):828-831, 836. DOI:10.12114/j.issn. 1007-9572.2020.00.052. (in Chinese).
[16] TAKENAKA M, URAMOTO H, SHIMOKAWA H, et al. Discriminative features of thin-slice computed tomography for peripheral intrapulmonary lymph nodes[J]. Asian Journal of Surgery 2013, 36(2):69-73. DOI: 10.1016/j.asjsur.2012.11.005.
[17] 明佳蕾, 方向明. 基于人工智能的CT肺结节检出临床应用及研究进展[J]. 中华放射学杂志, 2019, 53(6):522-525. DOI:10.3760/cma.j.issn.1005.1201.2019.06.017. MING J L, FANG X M. Clinical research and application of pulmonary nodules detection by CT with artificial intelligence[J]. Chinese Journal of Radiology, 2019, 53(6):522-525. DOI:10.3760/cma.j.issn.1005.1201.2019.06.017. (in Chinese).
[18] 李欣菱, 王颖. 人工智能在肺结节检测与诊断中的应用及发展[J]. 新发传染病电子杂志, 2019, 4(3):185-189. DOI:10.3877/j.issn.2096-2738.2019.03.014. LI X L, WANG Y. Application and development of artificial intelligence in detection and diagnosis of pulmonary nodules[J]. Electronic Journal of Emerging Infectious Diseases, 2019, 4(3):185-189. DOI:10.3877/j.issn.2096-2738.2019.03.014. (in Chinese).
[19] GINNEKEN V B. Fifty years of computer analysis in chest imaging:Rule based machine learning deep learning[J]. Radiological Physics and Technology, 2017, 10(1):23-32. DOI: 10.1007/s12194-017-0394-5.
[20] NIBALI A, HE Z, WOLLERSHEIM D. Pulmonary nodule classification with deep residual networks[J]. International Journal of Computer Assisted Radiology and Surgery, 2017, 12(10):1799-1808. DOI: 10.1007/s11548-017-1605-6.
-
期刊类型引用(19)
1. 万鑫龙，邓尚彪，高鹏. 基于深度学习的肺结节良恶性CT智能鉴别系统的构建及应用. 临床医学工程. 2025(01): 12-15 . 
百度学术
2. 袁冬健，彭磊，张玉荣，沈春冬，茅竞文. 人工智能CT扫描联合血清CEA检测对肺结节良恶性的诊断价值研究. 影像研究与医学应用. 2025(05): 109-111+114 . 百度学术
3. 苏寅晨，张晓琴. 人工智能辅助诊断系统在肺结节检测及良恶性判断中的应用价值. CT理论与应用研究. 2024(03): 325-331 . 本站查看
4. 杨嫚. 人工智能技术辅助诊断肺小结节的临床应用价值. 中国现代医药杂志. 2024(06): 75-77 . 百度学术
5. 钟丽茹，罗娜，贺露瑶. 多个人工智能辅助诊断系统对肺结节的诊断效能比较. 影像研究与医学应用. 2024(17): 26-29 . 百度学术
6. 贺呈祥，栗欣，张怀瑢. 基于人工智能与影像医生对肺结节检出效能的应用价值. 宁夏医科大学学报. 2024(09): 911-914 . 百度学术
7. 王璟琛，柴军. 人工智能体积密度法判断肺亚实性结节的浸润性研究. CT理论与应用研究. 2023(02): 241-248 . 本站查看
8. 王彬，顾文权，蔡孙美，周杰，张琦，雷撼. 基于人工智能的胸部CT肺结节检出及良恶性诊断效能评估. 影像研究与医学应用. 2023(05): 95-97 . 百度学术
9. 张添辉，刘舒珊，范伟雄. AI辅助实习医生和规培医生鉴别良恶性肺结节的价值. 影像研究与医学应用. 2023(05): 83-85 . 百度学术
10. 周伟文，谭学渊，余佐时. AI肺小结节检测系统用于低剂量CT肺筛查中的应用研究. 中国CT和MRI杂志. 2023(05): 43-45 . 百度学术
11. 陈新花，黄晓旗，李建龙，郭佑民. 基于两种AI软件对肺结节的诊断价值评价研究. CT理论与应用研究. 2023(04): 493-499 . 本站查看
12. 梁付奎，傅晓琴，李彩琴，刘芸，刘山波. 人工智能AI联合低剂量肺部CT扫描在肺结节诊断中的应用价值分析. 影像研究与医学应用. 2023(15): 28-30 . 百度学术
13. 方荣荣，陈深远，张先仙. 人工智能辅助技术在肺磨玻璃结节CT定性诊断中的准确率研究. 实用医学影像杂志. 2023(04): 266-269 . 百度学术
14. 张磊. 人工智能辅助诊断系统在低剂量肺部CT扫描筛查肺原位癌的应用价值. 影像研究与医学应用. 2023(23): 104-106 . 百度学术
15. 卢翼，陈友三，杨磊，肖诗银. 基于人工智能的CT检查对肝肿瘤患者良恶性的诊断价值. 影像研究与医学应用. 2023(24): 74-76 . 百度学术
16. 吴久纯，李甜，李晓东，卓越，张玉娇，刘敬禹. 基于人工智能随访预测肺结节增长的影响因素研究. 中国全科医学. 2022(17): 2115-2120 . 百度学术
17. 谢春晓，顾文豪. 人工智能CT定量分析肺结节的临床应用价值. 影像研究与医学应用. 2022(14): 74-76 . 百度学术
18. 耿然. 基于深度学习的人工智能影像辅助诊断系统对肺结节的诊断效能评价. 新乡医学院学报. 2022(11): 1031-1035 . 百度学术
19. 刘甜，范丽. 人工智能在肺癌筛查中的研究进展. 中国医学计算机成像杂志. 2022(06): 681-684 . 百度学术
其他类型引用(2)
计量
- 文章访问数: 584
- HTML全文浏览量: 50
- PDF下载量: 83
- 被引次数: 21
下载:
