Onko-Gene

Critical literature watch in the molecular oncology of solid tumors

FREN

Clinical · Molecular biology · Bioinfo & AI

Biallelic loss of RB1 in hepatocellular carcinoma as synthetic lethal target for artificial intelligence-guided therapy

Summary

This study characterized biallelic inactivation of the RB1 tumor suppressor gene (RB1-Bi) in hepatocellular carcinoma (HCC), finding it in 14.6% of patients and associating it with significantly reduced overall survival. A deep learning model (FR-MIL) was developed to accurately predict RB1-Bi status from routine histopathology slides. High-throughput drug screenings revealed that RB1-Bi HCC cells are particularly sensitive to inhibitors targeting mitotic regulators and PARP inhibitors. These synthetic lethal interactions were validated in vitro and in vivo, suggesting RB1-Bi is an actionable biomarker for precision medicine in HCC.

Critical appraisal

Clinical. This research identifies biallelic RB1 loss as an independent and clinically actionable prognostic biomarker in hepatocellular carcinoma (HCC), associated with significantly reduced overall survival (adjusted HR 3.32, 95% CI 1.93-5.72, p < 0.001). This allows for the identification of a high-risk HCC subtype with specific therapeutic vulnerabilities. The detection of this biomarker could justify phase II/III clinical trials evaluating targeted therapies in these patients. Clinical impact could be seen in the medium term (3-5 years) by altering treatment strategies for RB1-Bi HCC patients.

Molecular biology. The study discovered that biallelic RB1 inactivation creates a synthetic lethal vulnerability to inhibitors of mitotic regulators (such as AURKA, PLK1, KSP) and PARP inhibitors. This discovery was made through high-throughput drug screening in isogenic HCC models and validated both in vitro and in vivo. The combination of these inhibitors showed synergistic effects. These findings pave the way for the development of companion diagnostic tests to identify patients eligible for these targeted therapies, potentially via NGS panels or ddPCR tests for RB1-Bi detection.

Bioinfo & AI. A deep learning model, FR-MIL, was developed to predict RB1-Bi status from hematoxylin and eosin-stained histopathology slides. With an F1 score of 84.39% (95% CI, ±0.02), this model demonstrates high accuracy and reproducibility across independent validation cohorts. Its clinical deployment could transform the workflow by enabling rapid and non-invasive identification of RB1-Bi HCC patients directly from routine biopsies, without requiring additional costly or time-consuming molecular tests, thereby facilitating patient stratification for targeted therapies.

The summary and the appraisal were written by a language model from the title and abstract of the original article. A curator decided to publish it. Methodology