Predictive Modeling and Mendelian Randomization for Identifying HCC Patients with High Response to TACE with...

Hepatocellular carcinoma (HCC) is a major cause of cancer deaths. This study developed a clinical prediction model to identify patients likely to respond to transcatheter arterial chemoembolization (TACE) combined with atezolizumab and bevacizumab (Atez/Bev), using Mendelian randomization (MR) to validate the therapeutic effect on HCC via PDCD1, CD274, and VEGFA.

A retrospective analysis of 258 unresectable HCC patients administered TACE, with or without Atez/Bev therapy, was conducted. Propensity score matching (PSM) was employed to compare objective response rates (ORRs). A predictive model for response to TACE+Atez/Bev was developed using logistic regression and validated. MR was used to analyze the causal relationships between the PDCD1, CD274, VEGFA genes and Atez/Bev efficacy.

After PSM, the TACE+Atez/Bev group demonstrated significantly higher ORR compared with the TACE group (73.8% vs 56.3%, p=0.031). Multivariate logistic regression identified hepatic vein invasion, albumin (ALB), and platelets (PLT) as key outcome factors, which were used to develop a nomogram with AUCs of 0.81 and 0.89 in the training and validation cohorts, respectively. Calibration curve analysis showed good agreement with actual outcomes, and decision curve analysis highlighted the nomogram's potential for patient selection. MR analyses provided genetic support for the therapeutic target, revealing a significant protective association between PDCD1 inhibition and HCC risk (inverse-variance weighted, DrugOR=0.713, 95% CI: 0.599-0.848, p<0.001). This finding substantiates the biological rationale for the PDCD1 inhibitor atezolizumab. No significant associations were found for CD274 or VEGFA.

We developed a predictive model for TACE+Atez/Bev therapy, enabling effective patient screening. MR validated PDCD1's role in HCC immunotherapy at the genetic level.