Clinical decision-making in oncology involves multi-modal data such as radiology and clinical factors. In recent years, several computer-aided multi-modal decision systems have been developed to predict the recurrence of hepatocellular carcinoma (HCC) after hepatectomy, but these models simply concatenate features naively at the feature level, which would create redundancy and hinder model performance. Particularly, we found that integrating deep multi-modal models based on tensor fusion, which can better handle relevant information between different modalities and reduce model redundancy to obtain better results. This finding suggests the presence of independent, complementary prognostic information between radiology and clinical modalities. Hence we propose a multi-modal learning network that uses a tensor-based fusion approach with clinical parameters, radiomics features and corresponding complex interrelationships between pathological data to predict postoperative early recurrence of single hepatocellular carcinoma. In order to maximize the information gathered from each modality, we introduced a multi-modal fusion loss function based on orthogonal loss to assist multi-modal training. With IRB's approval, we collected 176 cases with radiomics (MRI) and pathological features diagnosed by experienced clinicians, and established an approach based on tensor fusion that takes pathological findings and postoperative early recurrence as ground truth. Training with 140 cases and tested with 36 cases with 5-fold cross validation, the proposed network achieved the AUC of 0.883, which showed great potential in predicting postoperative early HCC recurrence. The ablation experimental results also showed that the auxiliary loss function had a statistically significant improvement in model performance (p<0.01).

Background/aims: To assess the performance of transient elastography (TE), two-dimensional shear wave elastography (2D-SWE), and magnetic resonance elastography (MRE) for staging significant fibrosis and cirrhosis in untreated chronic hepatitis B (CHB) patients. Methods: Pubmed, Embase, Web of Science and Cochrane Library were searched for terms involving CHB, TE, SWE, and MRE. Other etiologies of chronic liver disease (CLD), previous treatment in patients or articles not published in SCI journals were excluded. Hierarchical non-linear models were used to evaluate the diagnostic accuracy of TE, 2D-SWE and MRE. Heterogeneity was explored via analysis of threshold effect and meta-regression. Results: Twenty-eight articles with a total of 4540 untreated CHB patients were included. The summary AUROC using TE, 2D-SWE and MRE for predicting significant fibrosis (SF) were 0.84, 0.89, and 0.99, respectively. MRE is more accurate than both TE (P＜0.01) and 2D-SWE (P＜0.01) in staging SF. 2D-SWE is superior to TE in detecting SF (P＜0.01). The summary AUROC employing TE, 2D-SWE and MRE for detecting cirrhosis were 0.9, 0.94, and 0.99, respectively. TE displayed a similar diagnostic accuracy with 2D-SWE in staging cirrhosis (P=0.14). MRE and 2D-SWE are comparable for staging cirrhosis (P=0.08). MRE is superior than TE (P＜0.01) in staging cirrhosis.Conclusion: TE, 2D-SWE, and MRE express acceptable diagnostic accuracies in staging staging significant fibrosis and cirrhosis in untreated CHB patients. Both MRE and 2D-SWE are better choices while the TE can be regarded as a secondary option.