This study aimed to develop and validate a prognostic nomogram for recurrence-free survival (RFS) after surgery in the absence of adjuvant therapy to guide the selection for adjuvant imatinib therapy based on Residual Neural Network (ResNet).The ResNet model was developed based on contrast-enhanced computed tomography (CE-CT) in a training cohort consisted of 80 patients pathologically diagnosed gastrointestinal sromal tumors (GISTs) and validated in internal and external validation cohort respectively. Independent clinicopathologic factors were integrated with the ResNet model to construct the individualized nomogram. The performance of the nomogram was evaluated in regard to discrimination, calibration, and clinical usefulness.The ResNet model was significantly associated with RFS. Integrable predictors in the individualized ResNet nomogram included the tumor site, size, and mitotic count. Compared with modified NIH, AFIP, and clinicopathologic nomogram, both ResNet nomogram and ResNet model showed a better discrimination capability with AUCs of 0·947(95%CI, 0·910–0·984) for 3-year-RFS, 0·918(0·852–0·984) for 5-year-RFS, and AUCs of 0·912 (0·851–0·973) for 3-year-RFS, 0·887(0·816–0·960) for 5-year-RFS, respectively. Calibration curve shows the good calibration of the nomogram in terms of the agreement between the estimated and the observed 3- and 5- year outcomes. Decision curve analysis showed that the ResNet nomogram had a higher overall net benefit.In conclusion, we presented a deep learning-based prognostic nomogram to predict RFS after resection of localized primary GISTs with excellent performance and could be a potential tool to select patients for adjuvant imatinib therapy.

BackgroundRecent studies have shown that long non-coding RNAs (lncRNAs) play a crucial role in the induction of cancer through epigenetic regulation, transcriptional regulation, post-transcriptional regulation and other aspects, thus participating in various biological processes such as cell proliferation, differentiation and apoptosis. As a new nova of anti-tumor therapy, immunotherapy has been shown to be effective in many tumors of which PD-1/PD-L1 monoclonal antibodies has been proofed to increase overall survival rate in advanced gastric cancer (GC). Microsatellite instability (MSI) was known as a biomarker of response to PD-1/PD-L1 monoclonal antibodies therapy. The aim of this study was to identify lncRNAs signatures able to classify MSI status and create a predictive model associated with MSI for GC patients.MethodsUsing the data of Stomach adenocarcinoma from The Cancer Genome Atlas (TCGA), we developed and validated a lncRNAs model for automatic MSI classification using a machine learning technology – support vector machine (SVM). The C-index was adopted to evaluate its accuracy. The prognostic values of overall survival (OS) and disease-free survival (DFS) were also assessed in this model.ResultsUsing the SVM, a lncRNAs model was established consisting of 16 lncRNA features. In the training cohort with 94 GC patients, accuracy was confirmed with AUC 0.976 (95% CI, 0.952 to 0.999). Veracity was also confirmed in the validation cohort (40 GC patients) with AUC 0.950 (0.889 to 0.999). High predicted score was correlated with better DFS in the patients with stage I-III and lower OS with stage I-IV.ConclusionThis study identify 16 LncRNAs signatures able to classify MSI status. The correlation between lncRNAs and MSI status indicates the potential roles of lncRNAs interacting in immunotherapy for GC patients. The pathway of these lncRNAs which might be a target in PD-1/PD-L1 immunotherapy are needed to be further study.