人工智能诊断系统在肺结节CT筛查中的应用价值评估OA

[Objective]To analyze the application value of artificial intelligence diagnostic systems in CT screening of pulmonary nodules.[Methods]A retrospective analysis was conducted on 120 patients with pulmonary nodules who underwent chest CT scans in Panyu Hospital of Traditional Cinese Medicine from February 2025 to May 2025.Chest CT scans were completed,and the films were read by physicians and analyzed by the AI system to observe the results.[Results]Based on the unanimous consensus reached by multiple senior experts after learning about the AI and the initial physician's film reading results,the final number of nodules screened from the imaging data of 120 patients with pulmonary nodules was 136.The number of pulmonary nodules screened by the physician's manual film reading was 132,accounting for 97.06%,and 136 were screened out by the artificial intelligence diagnosis system,accounting for 100.00%.Compared with the manual film reading by physicians,the detection rate of nodule screening by the artificial intelligence diagnostic system was higher,P<0.05.Among the 136 nodules,there were 99 solid,3 partial solid,20 ground-glass and 14 calcified nodules respectively.Among them,the diagnosis rates of nodules of different densities by the artificial intelligence diagnosis system were all 100%.The detection rate was higher than that of the physician's manual film reading,which was 100.00%(99/99),33.33%(1/3),90.00%(18/20),and 100.00%(14/14),totaling 97.06%(132/136).However,in the subgroup analysis,although the detection rate of the artificial intelligence diagnostic system was higher,however,the difference did not reach statistically significant P>0.05.Regarding the detection of nodule size,for nodules<1 cm,the detection rate of the artificial intelligence diagnostic system was 90.00%(18/20),which was higher than that of the physician's manual reading of films,which was 70.00%(14/20).The detection rate of the 1-2 cm nodules by the artificial intelligence diagnosis system was 91.67%(22/24),which was higher than that of the physician's manual reading at 75.00%(18/24).The detection rate of the 2-3 cm nodules by the artificial intelligence diagnosis system was 90.48%(38/42),which was higher than that of the physician's manual reading at 78.57%(33/42).For the 3-4 cm nodules,the detection rate of the artificial intelligence diagnosis system was 100.00%(50/50),which was higher than that of the physician's manual reading at 96.00%(48/50).The total detection rate of the artificial intelligence diagnostic system was 94.12%(128/136),which was higher than that of the physician's manual film reading at 83.09%(113/136),P<0.05.[Conclusion]In CT screening of pulmonary nodules,compared with manual film reading by physicians,the artificial intelligence diagnostic system has a higher detection rate of pulmonary nodules and higher accuracy in assessing the nature and size of nodules,which is recommended to implement and promote in clinical disease diagnosis.