Ying Zhao ¹ Sen Wang ^2,3 Yue Wang ¹ Jun Li ¹ Jinghong Liu ¹ Yuhui Liu ⁴ Haitong Ji ⁴ Wenhan Su ⁴ Qinhe Zhang ¹ Qingwei Song ¹ Yu Yao ^2,3 Ailian Liu ^1,5*

• ¹ Department of Radiology, First Affiliated Hospital, Dalian Medical University, Dalian, Liaoning Province, China
• ² Chengdu Institute of Computer Application, Chinese Academy of Sciences (CAS), Chengdu, Sichuan Province, China
• ³ School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing, China
• ⁴ College of Medical Imaging, Dalian Medical University, Dalian, Liaoning Province, China
• ⁵ Dalian Engineering Research Center for Artificial Intelligence in Medical Imaging, Dalian, China

Purpose: To explore the role of deep learning (DL) and radiomics-based integrated approach based on contrast enhanced magnetic resonance imaging (CEMRI) for predicting early recurrence (ER) in hepatocellular carcinoma (HCC) patients after curative resection. Methods: Total 165 HCC patients (ER, n = 96 vs. non-early recurrence (NER), n = 69) were retrospectively collected and divided into a training cohort (n = 132) and a validation cohort (n = 33). From pretreatment CEMR images, a total of 3111 radiomics features were extracted, and radiomics models were constructed using five machine learning classifiers (logistic regression, support vector machine, k-nearest neighbor, extreme gradient Boosting, and multilayer perceptron). DL models were established via three variations of ResNet architecture. The clinical-radiological (CR), radiomics combined with clinical-radiological (RCR), and deep learning combined with RCR (DLRCR) models were constructed. Model discrimination, calibration, and clinical utilities were evaluated by receiver operating characteristic curve, calibration curve, and decision curve analysis, respectively. The best-performing model was compared with the widely used staging systems and preoperative prognostic indexes. Results: The RCR model (area under the curve (AUC): 0.841 and 0.811) and the optimal radiomics model (AUC: 0.839 and 0.804) achieved better performance than the CR model (AUC: 0.662 and 0.752) in the training and validation cohorts, respectively. The optimal DL model (AUC: 0.870 and 0.826) outperformed the radiomics model in the both cohorts. The DL, radiomics, and CR predictors (aspartate aminotransferase (AST) and tumor diameter) were combined to construct the DLRCR model. The DLRCR model presented the best performance over any model, yielding an AUC, an accuracy, a sensitivity, a specificity of 0.917, 0.886, 0.889, and 0.882 in the training cohort and of 0.844, 0.818, 0.800, and 0.846 in the validation cohort, respectively. The DLRCR model achieved better clinical utility compared to the clinical staging systems and prognostic indexes. Conclusion: Both radiomics and DL models derived from CEMRI can predict HCC recurrence, and DL and radiomics-based integrated approach can provide a more effective tool for the precise prediction of ER for HCC patients undergoing resection.