A risk scoring system for predicting visceral pleural invasion in non-small lung cancer patients

Objective This study aimed to construct a simple scoring system for predicting visceral pleural invasion of non-small cell lung cancer (NSCLC) from computed tomography (CT) findings and clinicopathological factors in lesions directly under the pleural membrane. Methods Among 376 cases of surgically treated NSCLC, cases in which the tumor was <= 7 cm in diameter and in contact with the pleura on the CT image were retrospectively extracted and examined. The CT findings and clinicopathological factors associated with the presence of pathological pleural invasion in each case were examined by Fisher's exact test. A score was then assigned based on the odds ratio obtained for each factor, and a risk scoring system for predicting pleural invasion was constructed. Result In the 138 extracted cases, pathological visceral pleural invasion was found in 64 cases. The scoring system predicting pleural invasion could be defined as follows: pl risk score = 3 (tumor diameter in CT >= 24 mm) + 3 (tumor contact length with pleura in CT >= 16 mm) + 3 (smoking index >= 400) + 3 (clinically lymph node positive) + 2 (tumor with cavity in CT) + 2 (serum CEA level > 4.4 ng/mL). A score was calculated for each case and an ROC curve was created. The cutoff value was score 8 and the area under curve (AUC) was 0.68. Conclusion Our findings suggest that visceral pleural invasion can be predicted using a score calculated from several simple CT findings and clinicopathologic factors.