Researchers sought to create an SVM classifier for synchronous PC prediction utilizing artificial intelligence’s ResNet-3D technique for a study. PC from CRC detection and staging was still problematic. The main tumors in synchronous PC were identified on preoperative contrast-enhanced computed tomography (CT) imaging. To develop a ResNet3D + SVM classifier, the characteristics of the nearby peritoneum were extracted. In the test set, the performance of the ResNet3D + SVM classifier was compared with routine CT, which radiologists evaluated. The training set included 19,814 pictures from 54 PC patients and 76 non-PC individuals. A total of 7,837 pictures from 40 test patients made up the test set. The ResNet-3D analyzed the test photos in about 34 seconds. Investigators created an SVM classifier by combining ResNet-3D characteristics with 12 PC-specific features (P<0.05) to improve PC detection accuracy. In the test set, the ResNet3D + SVM classifier had a sensitivity of 93.75%, specificity of 94.44%, positive predictive value (PPV) of 93.75%, and negative predictive value (NPV) of 94.44%, with an AUC of 0.922 (0.912–0.944). The results were better than a standard contrast-enhanced CT scan (AUC: 0.791). The ResNet3D + SVM classifier, which used the ResNet-3D framework and was based on a deep learning algorithm, had shown tremendous promise in predicting synchronous PC in CRC.
