A novel diagnostic model for differentiation of lung metastasis from primary lung cancer in patients with colorectal cancer

OBJECTIVE: This study aimed to evaluate the (18)F-FDG PET/CT in differentiating lung metastasis(LM) from primary lung cancer(LC) in patients with colorectal cancer (CRC). METHODS: A total of 120 CRC patients (80 male, 40 female) who underwent (18)F-FDG PET/CT were included. The diagnosis of primary lung cancer or lung metastasis was based on histopathology The patients were divided into a training cohort and a validation cohort randomized 1:1. Independent risk factors were extracted through the clinical information and (18)F-FDG PET/CT imaging characteristics of patients in the validation cohort, and then a diagnostic model was constructed and a nomograms was made. ROC curve, calibration curve, cutoff, sensitivity, specificity, and accuracy were used to evaluate the prediction performance of the diagnostic model. RESULTS: One hundred and twenty Indeterminate lung lesions (ILLs) (77 lung metastasis, 43 primary lung cancer) were analyzed. No significant difference in clinical characteristics and imaging features between the training and the validation cohorts (P > 0. 05). Using uni-/multivariate analysis, pleural tags and contour were identified as independent predictors. These independent predictors were used to establish a diagnostic model with areas under the receiver operating characteristic curves (AUCs) of 0.92 and 0.89 in the primary and validation cohorts, respectively. The accuracy rate of the diagnostic model for differentiating LM from LC were higher than that of subjective diagnosis (P < 0.05). CONCLUSIONS: Pleural tags and contour were identified as independent predictors. The diagnostic model of ILLs in patients with CRC could help differentiate between LM and LC.