956P - Pyroptosis-related pattern depicts tumor microenvironment and predicts prognosis and efficacy in hepatocellular carcinoma patients receiving combined molecular targeted therapy and immune checkpoint inhibitors: A multi-omics study

Background

Patients with liver cancer who are difficult to resect initially or have a high risk of postoperative recurrence often opt for conversion or neoadjuvant therapy. Targeted drug therapy combined with immunotherapy is a recommended treatment, but predictive factors for its efficacy are limited. Cell necrosis is linked to immune cell infiltration in tumors. We aimed to use necrosis-related genes to predict the efficacy of targeted immunotherapy in liver cancer.

Methods

We analyzed 53 necrosis-related genes in TCGA_GTEx, identifying six genes related to survival, recurrence, immune cell infiltration, and the tumor microenvironment in liver cancer. We constructed a PI score and validated it internally and externally. Single-cell data from patients before and after immunotherapy were used. We matched 131 patients receiving targeted immunotherapy using the PSM method. Immunohistochemistry was used to validate the score's clinical application. T-tests evaluated inter-group differences. Postoperative pathology and PFS assessed efficacy.

Results

Four genes (TRIM21, NLRC4, IL1A, GSDME) and two genes (NLRP6 and GZMA) were associated with recurrence risk and long-term survival. The PI was calculated at the patient level, predicting OS (HR=3.43, p=4.76e−10) and PFS (HR=2.00, p=5.04e−5). External validation confirmed the PI's predictive ability for OS (HR=1.530, p=0.014) and DFS (HR=1.548, p=0.017). Single-cell sequencing showed higher PI scores correlated with suppressed CD8+T cell function and worse immunotherapy efficacy. Surgical group pPI scores correlated with pathological remission (R=0.953, p<0.01). In the nonoperative group, median PFS was 8.5 months, with pPI score AUC=0.759 (95%CI:0.657-0.832).

Conclusions

We developed a genomic score to predict individual sensitivity to targeted immunotherapy in liver cancer, allowing personalized efficacy prediction. This score, based on pre-treatment biopsy pathology, could aid in predicting treatment efficacy in liver cancer patients undergoing conversion or neoadjuvant therapy.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.