Deep Learning Prediction of Metastasis in Locally Advanced Colon Cancer Using Binary Histologic Tumor Images

SIMPLE SUMMARY: Deep learning methods are increasingly being applied for tissue classification to improve diagnosis and optimize therapy stratification. In this study, we developed the Binary ImaGe Colon Metastasis classifier (BIg-CoMet), a semi-guided approach for the stratification of colon cancer patients into two risk groups according to the occurrence of distant metastasis, using an InceptionResNetV2-based deep learning model trained on binary images. For a validation collective (n = 128), we showed that BIg-CoMet was able to stratify patients appropriately. The predicted high-risk group showed a worse clinical course for being metastasis-free, and the risk group was a prognostic factor for the occurrence of metastasis. These results were also found for both Union Internationale Contre le Cancer (UICC) subgroups. We demonstrated that Big-CoMet is useful for the stratification of colon cancer patients into risk groups based on images reflecting tumor architecture. ABSTRACT: In this study, we developed the Binary ImaGe Colon Metastasis classifier (BIg-CoMet), a semi-guided approach for the stratification of colon cancer patients into two risk groups for the occurrence of distant metastasis, using an InceptionResNetV2-based deep learning model trained on binary images. We enrolled 291 colon cancer patients with pT3 and pT4 adenocarcinomas and converted one cytokeratin-stained representative tumor section per case into a binary image. Image augmentation and dropout layers were incorporated to avoid overfitting. In a validation collective (n = 128), BIg-CoMet was able to discriminate well between patients with and without metastasis (AUC: 0.842, 95% CI: 0.774–0.911). Further, the Kaplan–Meier curves of the metastasis-free survival showed a highly significant worse clinical course for the high-risk group (log-rank test: p < 0.001), and we demonstrated superiority over other established risk factors. A multivariable Cox regression analysis adjusted for confounders supported the use of risk groups as a prognostic factor for the occurrence of metastasis (hazard ratio (HR): 5.4, 95% CI: 2.5–11.7, p < 0.001). BIg-CoMet achieved good performance for both UICC subgroups, especially for UICC III (n = 53), with a positive predictive value of 80%. Our study demonstrates the ability to stratify colon cancer patients via a semi-guided process on images that primarily reflect tumor architecture.