1 / 2023-04-28 16:36:38

Integrative Models of Histopathological Images and Multi-omics Data Predict Prognosis in Endometrial Carcinoma

Endometrial Carcinoma; histopathology, proteomics, transcriptomics, genomics;

Objective. We used histopathological images for predicting molecular features, and further predict the overall survival (OS) in Endometrial Carcinoma (EC) patients.

Methods. The patients from The Cancer Genome Atlas (TCGA) were separated into the training set (n=215) and test set (n=214) in proportion of 1:1. By analyzing quantitative histological image features and setting up random forest model verified by cross-validation, we constructed prognostic models for OS. The model performance is evaluated with the time-dependent receiver operating characteristics (AUC) over the test set.

Results. Prognostic models based on histopathological image features (HIF) predicted OS in test cases (5-year AUC = 0.803). The performance of combing histopathology and omics transcends the that of genomics, transcriptomics, or proteomics alone. Additionally, the multi-dimensional omics data including HIF, genomics, transcriptomics and proteomics attained the maximum 1-, 3-, and 5-year AUCs of 0.866, 0.869, and 0.856, showcasing the greatest discrepancy in survival (HR = 18.347, 95%CI: 11.09–25.65, p < 0.001).

Conclusions. Our results indicated the complementary features of HIF can improve the prognostic performance of EC patients. The integration of HIF and multi-dimensional omics data might ameliorate survival determination and risk stratification in clinical practice.