AI Diagnosis Changed The Operation Mode Of Pulmonary Nodules: A retrospectively Cohort Study

Abstract

Abstract Objective: The aim of this study was to investigate whether artificial intelligence assisted Low-dose computer tomography (LDCT) diagnosis can change the surgical proposal of patients with pulmonary nodules. Methods: The clinical image data of consecutive patients with multiple pulmonary nodules who received LDCT scanning of the lungs and underwent surgical resection in Fujian Medical University Union Hospital from December 2020 to December 2021 were collected retrospectively. Patients were divided into manual group (MG) and artificial intelligence group (AIG) according to whether AI is used to assist image reading. A junior doctor and a senior doctor were included both in the two groups. The two doctors in the same group allocated cases according to the 1:1 ratio. The differences were compared between the two groups. Results: A total of 300 patients were enrolled in this study. The number of nodules need to be removed in MG was significantly less than AIG (p <0.0001 ). In terms of interpretation time, 60s (60,60) in MG was significantly longer than 30s (20,30) in AIG (p<0.0001). The missed diagnosis rate of junior doctor was significantly higher than that of senior doctor (17.0% vs. 8.5%, p=0.013). Compared with the manual diagnosis, 35 patients (11.7%) finally had an increase in the number of pulmonary nodules removed after AI assisted diagnosis. The total number of nodules for extended resection was 50. Conclusions: AI assisted LDCT diagnosis can change the surgical proposal of patients with pulmonary nodules.