The Application Value of Artificial Intelligence- Assisted Low-Dose 64-Slice Spiral CT Scanning in Lung Cancer Screening and Nodule Follow-Up

Abstract

Objective: To explore and observe the application value of artificial intelligence (AI)-assisted 64-slice spiral CT low-dose scanning in lung cancer screening and nodule follow-up. Methods: A retrospective analysis was conducted on 390 patients who underwent 64-slice spiral CT low-dose scanning for lung cancer screening at Qingyang Traditional Chinese Medicine Hospital from 2020 to 2024. By recording the morphology and nature of pulmonary nodules (solid, partially solid, ground-glass, and calcified nodules) and measuring nodule size (including volume, longest diameter, maximum short diameter, etc.), the diagnoses of target pulmonary nodules made by AI, senior radiologists, and senior radiologists combined with AI were compared to evaluate the clinical application value of different diagnostic methods. Results: A total of 390 subjects participated in the 64-slice spiral CT low-dose scanning for lung cancer screening. Among them, the physician + AI group detected 208 pulmonary nodules, the AI group detected 198 pulmonary nodules with a detection rate of 95.19%, and the physician group detected 194 pulmonary nodules with a detection rate of 93.27%. There was no statistical difference in the detection rate between the AI group and the physician group (χ² = 0.707, P = 0.400). No statistical differences were observed among different groups in terms of nodule density, nodule location, and the detection of positive nodules (P > 0.05). Using the positive nodules identified by the physician + AI group as the screening nodules and the pathological examination results of 6 cases of lung cancer obtained during follow-up as the confirmed results, the physician + AI group demonstrated a sensitivity of 100%, a false-negative rate of 0%, a specificity of 25.10%, a false-positive rate of 74.90%, a positive likelihood ratio of 1.34, a negative likelihood ratio of 0, a concordance rate of 26.90%, a positive predictive value of 3.2%, and a negative predictive value of 100%. Conclusion: The screening method of AI-assisted 64-slice spiral CT low-dose scanning can be used to rapidly rule out lung cancer, but positive results require further confirmation through pathological examination and other means.