Deep Learning Predicts Overall Survival of Patients With Unresectable Hepatocellular Carcinoma Treated by Transarterial Chemoembolization Plus Sorafenib

Abstract

Objectives: To develop and validate a deep learning-based overall survival (OS) prediction model in patients with hepatocellular carcinoma (HCC) treated with transarterial chemoembolization (TACE) plus sorafenib. Methods: This retrospective multicenter study consisted of 201 patients with treatmentnaïve, unresectable HCC who were treated with TACE plus sorafenib. Data from 120 patients were used as the training set for model development. A deep learning signature was constructed using the deep image features from preoperative contrast-enhanced computed tomography images. An integrated nomogram was built using Cox regression by combining the deep learning signature and clinical features. The deep learning signature and nomograms were also externally validated in an independent validation set of 81 patients. C-index was used to evaluate the performance of OS prediction. Results: The median OS of the entire set was 19.2 months and no significant difference was found between the training and validation cohort (18.6 months vs. 19.5 months, P = 0.45). The deep learning signature achieved good prediction performance with a C-index of 0.717 in the training set and 0.714 in the validation set. The integrated nomogram showed significantly better prediction performance than the clinical nomogram in the training set (0.739 vs. 0.664, P = 0.002) and validation set (0.730 vs. 0.679, P = 0.023). Conclusion: The deep learning signature provided significant added value to clinical features in the development of an integrated nomogram which may act as a potential tool for individual prognosis prediction and identifying HCC patients who may benefit from the combination therapy of TACE plus sorafenib.

Neutrophil-to-Lymphocyte and Platelet-to-Lymphocyte Ratios as Predictors of Outcomes in Patients With Unresectable Hepatocellular Carcinoma Undergoing Transarterial Chemoembolization Plus Sorafenib

“…It is difficult to predict prognosis in individual patients. Therefore, development of effective biomarkers to identify subpopulations of patients who are most likely to benefit from such a combination therapy is needed (Zhang et al, 2020b). Previous studies showed that sorafenib-related adverse events were associated with favorable prognosis in HCC patients (ref.).…”

State of the Art in Artificial Intelligence and Radiomics in Hepatocellular Carcinoma

“…Some investigations focused the contribution of textural analysis and DL, respectively, in selecting patients who can benefit from simultaneous association of TACE and sorafenib administration [ 78 , 79 ]. This combination is based on the inhibition of VEGF up-regulation induced by TACE, but its efficacy is still controversial.…”

“…This combination is based on the inhibition of VEGF up-regulation induced by TACE, but its efficacy is still controversial. Zhang and colleagues [ 79 ] validated a CNN model to predict overall survival from preoperative CE CT images of advanced HCC patients treated with TACE and sorafenib. They built an integrated nomogram based on clinical features and a DL-signature.…”

Progress of MRI Radiomics in Hepatocellular Carcinoma

“…Radiomics, a new technology, can transform the potential histopathological and physiological information in images into high-dimensional quantitative image features that can be mined (6,7).The study of radiomics will contribute to the early diagnosis and treatment of HCC and ultimately improve survival (8,9). In recent years, many studies have confirmed the application values of magnetic resonance imaging (MRI) radiomics in the diagnosis and differentiation (10,11), histological grading (12,13), microvascular invasion (MVI) assessment (14,15), radiogenomics (16,17),prediction of relapse and prognosis after surgical resection (18)(19)(20), response to transarterial chemoembolization(TACE) (21,22) and systemic treatment efficacy of HCC (23).…”

“…Traditional contrast-enhanced CT and MRI, including functional imaging, are the most commonly used biomarkers for evaluating the therapeutic response in clinical practice(159)(160)(161)(162)(163)(164)(165)(166)(167)(168)(169)(170)(171)(172). Research based on contrastenhanced CT and MR images has shown the value of radiomics and DL in predicting systemic treatment efficacy for advanced HCC(23,(173)(174)(175). Muléet al(174) analyzed the CT texture features of 92 patients before receiving sorafenib and found that the entropy of portal phase-derived entropy at fine texture scales was an independent predictor of OS, which was confirmed in their validation cohort.…”