Can visceral fat parameters based on computed tomography be used to predict occult peritoneal metastasis in gastric cancer?

Figure 2 Predictive performance of computed tomography attenuations for visceral fat in two regions of interest and two models for predicting peritoneal metastasis in gastric cancer. A: Receiving operator characteristic (ROC) curves of the visceral fat (VF) computed tomography (CT) attenuations on ROIs 1 and 2. The VF CT attenuations on ROI 2 [area under the curve (AUC): 0.657, sensitivity 93.0%, specificity 30.6%] showed a better predictive performance than those on ROI 1 (AUC: 0.599, sensitivity 60.0%, specificity 60.0%); B: ROC curves of the clinical and combined models. The two models showed comparable predictive performance for peritoneal metastasis in gastric cancer (AUC: 0.730 vs 0.749, P = 0.399)