For patients with refractory liver cancer—which is difficult to treat and has a low survival rate—a rapid pathological diagnosis using artificial intelligence (AI) and personalized targeted treatment strategies have been proposed.

A joint research team led by Professors Shim Joo-hyun (Department of Gastroenterology) and Sung Chang-ok (Department of Pathology) of Asan Medical Center, Professor Shim Joong-sub of the University of Macau, Professor Park Sang-hyun of POSTECH, and Professor Ahn Ji-hyun of Hanyang University Guri Hospital announced on the 6th that the complete deficiency of the tumor suppressor gene RB1 has been identified as a new biomarker for liver cancer.

The research team developed an AI-based diagnostic model to screen for this biomarker and also confirmed the anti-cancer effects of a combination therapy administering cell division inhibitors and PARP inhibitors together.

Currently, atezolizumab and bevacizumab combination therapy is used for advanced liver cancer, but it has limitations in that it responds only to some patients and drug resistance develops. Furthermore, only a small number of patients possess genetic mutations that respond to approved targeted therapies, creating a pressing need to discover new biomarkers.

In previous studies, RB1 gene mutations were identified in liver cancer patients, but there were limitations in understanding their biological and clinical significance because previous research did not clearly distinguish between cases where only one allele was deleted and cases where both alleles were deleted.

Accordingly, the research team conducted multi-omic analyses on a total of 561 patients—including 206 patients from Asan Medical Center and 355 patients from the US National Cancer Institute's public database (TCGA)—and validated the results using an independent cohort of 450 patients.

As a result, patients with "complete RB1 deficiency (RB1-Bi)," where both RB1 gene alleles are deleted or inactivated, accounted for approximately 14.6% of all liver cancer cases.

This patient group exhibited low cancer cell differentiation and rapid tumor progression, with a 3.32-fold higher risk of death and a 3.15-fold higher risk of recurrence compared to general patients. Through this, the research team confirmed that complete RB1 deficiency is an independent poor prognostic factor for liver cancer.

To enable the screening of high-risk groups without complex genomic analysis, the research team also developed a deep learning-based pathology AI model (FR-MIL).

This model predicts complete RB1 deficiency using standard stained pathology tissue slide images alone, recording an F1 score of 84.39% to 91.58% in external validation cohorts.

The potential for treatment was also confirmed. When 876 drugs were administered to RB1-deficient liver cancer cells, the research team found that the cells selectively underwent apoptosis (cell death) when exposed to specific drugs that inhibit cell division or damaged DNA repair.

In particular, in cellular and animal experiments, the combined administration of cell division inhibitors and PARP inhibitors significantly enhanced tumor suppression effects without distinct systemic side effects such as liver toxicity.

The research team explained that this study is significant because it presented a full spectrum of solutions for high-risk liver cancer patients with poor prognoses—ranging from biomarker discovery and AI pathological diagnosis to personalized combination treatment strategies.

Professor Shim Joo-hyun stated, "It is very encouraging that a new breakthrough tailored to genetic characteristics has been established for refractory liver cancer patients who currently have limited treatment options and are prone to developing resistance," adding, "If the AI diagnostic model and combination treatment strategy developed this study are introduced into actual clinical practice, they are expected to dramatically improve patient survival rates and greatly accelerate precision personalized medicine."

Meanwhile, the results of this study were published in the latest issue of the international translational medicine journal Signal Transduction and Targeted Therapy.

[Copyright (c) Global Economic Times. All Rights Reserved.]