Computer-aided Detection Low-dose CT for Early Lung Cancer Screening: A Feasibility Study

Objective: To analyze and compare the effect of computer-aided and artificially low dose CT in early lung cancer screening, and to provide theoretical basis for the selection of screening methods for early lung cancer. Methods: Low dose CT scans were performed on 256 high-risk patients with early lung cancer in our hospital from December 2015 to December 2017. Results: There were 42 patients of pulmonary nodules performed by low-dose CT, and the detection rate was 16.4%. There were 43 patients with pulmonary nodules detected by computer-assisted technique (manual error correction), and the detection rate was 16.8%. The difference was no statistically significant (χ²=0.014, P=0.905). There was a high consistency between manual interpretation and computer-assisted detection of high-risk patients (k=0.986). The working time of computer aided technology is lower than that of manual interpretation, saving about 70% of the working time (59.95±12.93) vs. (194.98±70.61) s, t=30.474, P=0.000). Conclusion: In this study, it was found that the efficiency of low-dose CT for early lung cancer screening with computer assisted (manual error correction) was significantly better than that of manual interpretation.