Objective  To explore the application value of artificial intelligence (AI) auxiliary diagnosis system in chest CT diagnosis of pulmonary nodules in plateau area. Methods Non-calcified pulmonary nodules found in chest CT of the People’s Hospital of Xizang Autonomous Region were retrospectively collected between January 2022 and May 2024. All pulmonary nodules were diagnosed based on pathological results as the gold standard using an AI assisted diagnostic system, with physicians manually reviewing the images and analyzing relevant parameters. The diagnostic efficacy of three diagnostic methods for pulmonary nodules was compared. Results  A total of 154 patients were included. Among them, there were 79 males and 75 females. The average age was (54.2±13.9) years old. There were 172 non-calcified pulmonary nodules. The combined diagnostic method of physician+AI [area under the curve (AUC)=0.848, 95% confidence interval (CI) (0.697, 0.835), P=0.003] was superior to physician [AUC=0.739, 95%CI (0.663, 0.806), P=0.031] or AI [AUC=0.771, 95%CI (0.697, 0.835), P=0.012]. There was no significant difference between AI and physician’s ability to distinguish (P>0.05). There was no significant difference in the benign and malignant rates of the three density types of pulmonary nodules identified by AI (P=0.386). The distribution of malignant pulmonary nodules in different lung segments in the plateau area was different, and the distribution was mainly in the apical segment of the right upper lobe and the posterior segment of the left upper lobe. There was no systematic deviation between physicians and AI in measuring the maximum cross-sectional area of pulmonary nodules (t=?0.687, P=0.493). Conclusions  In order to improve the efficiency and avoid the risk of misjudgment, the judgment of benign and malignant pulmonary nodules in clinical practice should be completed by doctors and AI. In the plateau area, there was no significant difference in the distribution of benign and malignant pulmonary nodules in solid, partial solid and ground glass density. Different from low altitude areas, they should be regarded as equally important identification objects when reading films. AI can effectively replace doctors to measure the cross-sectional area of nodules and improve efficiency. There are significant differences in the distribution of malignant pulmonary nodules in the plateau area, mainly in the apical segment of the upper lobe of the right lung and the posterior segment of the upper lobe of the left lung.

Citation： BIANBA Ciren, CIDAN Wangjiu, YU Jianqun, LEI Yanming. Application value of artificial intelligence assisted diagnostic system in CT evaluation of pulmonary nodules in plateau area. West China Medical Journal, 2026, 41(3): 454-459. doi: 10.7507/1002-0179.202504092 Copy

Copyright ? the editorial department of West China Medical Journal of West China Medical Publisher. All rights reserved