MRI-based clinical-radiologic model to predict the efficacy and prognosis of interventional with systemic therapy in
Xingrong Zheng1, Yuqing Ren2, Xionghao Pang3
1Department of Infectious Diseases, The Third Affiliated Hospital of Sun Yat-Sen University, 600# TianHe Road, Guangzhou, Guangdong Province, 510630, China.
Abstract:
AIM: Using magnetic resonance imaging (MRI) features combined with clinical information to predict treatment response in advanced hepatocellular carcinoma (aHCC) patients receiving tyrosine kinase inhibitors (TKIs), programmed cell death protein 1 (PD-1) inhibitors with interventional therapy (TPI therapy). METHODS: This retrospective, multicenter study included patients with baseline contrast-enhanced MRI images between December 2018 and December 2023. Clinical data of patients were collected and features in MRI images were used for radiomic analysis. A clinical-radiomic model was constructed by integrating clinical variables with radiomic scores. RESULTS: A total of 239 patients were enrolled of which 15 patients achieved complete response, 161 achieved partial response, 26 displayed stable disease, and 37 displayed progressive disease. To further investigate the efficacy of TPI therapy by clinical-radiomic model, patients were randomly assigned to training (N = 191) and validation group (N = 48) according to the ratio of 3:1. The area under the curve (AUC) of the clinical-radiologic model (training cohort 0.947, validation cohort 0.803) was better than that of the clinical model (training cohort 0.674, validation cohort 0.706) and radiologic model (training cohort 0.930, validation cohort 0.790). Meanwhile, participants were divided into high-risk and low-risk groups for analysis of overall survival (OS) and progression-free survival (PFS) in the clinical-radiologic model. The prognosis of OS and PFS in high risk patients in both training (OS, HR 2.074, 95%CI 1.294–3.324, p = 0.002; PFS, HR 2.097, 95%CI 1.433–3.068, p = 0.001) and validation cohort (OS, HR 4.534, 95%CI 1.054–19.494, p = 0.026; PFS, HR 2.969, 95%CI 1.383–6.377, p = 0.003) was worse. CONCLUSION: Radiographic features based on MRI images with clinical data can predict the efficacy and risk stratification of TPI therapy in aHCC.
