Liver Cancer: Can AI Identify an RB1-Deficient Subtype?

What Makes This Liver Cancer Subtype Different?

An October 14 announcement from the University of Macau highlights research into hepatocellular carcinoma, the most common primary liver cancer. The focus is biallelic RB1 inactivation: both copies of a gene that helps control cell division have lost function. [University of Macau announcement](https://www.gov.mo/en/news/411123/).

The discovery analysis included 561 people with HCC; 82, or 14.6%, had this alteration. The subtype was associated with poorer overall survival. Its frequency varied across validation cohorts, so 14.6% is not a universal prevalence estimate. These group-level findings cannot determine an individual's prognosis. The paper was published August 20. [Original study](https://www.nature.com/articles/s41392-026-02815-0).

How Could AI Help Identify These Tumors?

The system analyzes hematoxylin-and-eosin-stained specimens for patterns associated with RB1 loss. This approach could help identify candidates for further assessment without sequencing every specimen initially. [University of Macau announcement](https://www.gov.mo/en/news/411123/). The researchers tested the model in independent cohorts but say prospective studies across more centers remain necessary. [Study validation and limitations](https://www.nature.com/articles/s41392-026-02815-0).

The clinical question extends beyond recognizing a tumor pattern: will that information help select an effective treatment? The National Cancer Institute explains that biomarker testing can guide treatment or clinical-trial eligibility, but a molecular finding does not guarantee a useful drug match. Some findings have uncertain significance, and even a matching therapy may fail. This distinction is essential when interpreting the proposed AI application. [NCI biomarker testing guidance](https://www.cancer.gov/about-cancer/treatment/types/biomarker-testing-cancer-treatment).

Could This Discovery Change Liver Cancer Treatment?

Researchers combined drugs targeting cell division with PARP inhibitors, which interfere with DNA repair. The combinations suppressed tumor growth in laboratory and mouse experiments. However, the mice lacked normal immune defenses, and their tumors grew outside the liver. These limitations matter when predicting human responses. More extensive safety testing is also needed before clinical trials. [Experimental findings and limitations](https://www.nature.com/articles/s41392-026-02815-0).

For patients now, established treatment options remain central. Depending on disease extent and suitability, these include surgery, liver transplantation, ablation, embolization, radiation, targeted medicines and immunotherapy. Surgery removes cancerous tissue, while ablation destroys it locally; systemic medicines address disease through treatment that circulates through the body. Patients can discuss how their current treatment fits these options and whether an appropriate clinical trial is available. The new findings provide a research direction, not evidence that changing treatment would improve an individual patient's outcome. [NCI liver cancer treatment overview](https://www.cancer.gov/types/liver/what-is-liver-cancer/treatment).