Hiroshima University researchers develop genomic framework to prioritize and identify pathogenic mutations hidden in uncertain cancer data

Hiroshima University researchers develop a system to prioritize and identify hidden pathogenic variants in cancer, improving precision oncology accuracy.

By: AXL Media

Published: Mar 31, 2026, 11:29 AM EDT

Source: Information for this report was sourced from Hiroshima University

Addressing the Clinical Challenge of Genomic Uncertainty

The rapid implementation of comprehensive genomic profiling (CGP) has revolutionized oncology by allowing treatments to be tailored to the molecular profile of a tumor, yet it has also created a significant data bottleneck. While these tests can identify thousands of genetic alterations, a large majority are classified as variants of uncertain significance (VUS), meaning their impact on cancer progression or treatment response is unknown. Researchers at Hiroshima University have now proposed a systematic framework designed to prioritize these mysterious variants for deeper functional analysis, potentially moving them from the category of "uncertain" to "pathogenic" to better inform patient care.

The Case of the Exceptional Responder

The motivation for this study originated from a singular clinical anomaly involving a patient with a poor prognosis who achieved long-term remission following platinum-based chemotherapy. Despite the presence of recurrent cancer across multiple organs, the patient's recovery was surprisingly robust. Dr. Hiroaki Niitsu, an associate professor at Hiroshima University Hospital, noted that this exceptional response suggested an underlying genetic vulnerability that standard genomic interpretations had failed to categorize. This observation prompted the team to look closer at the specific genetic variants that might be driving such unique therapeutic outcomes.

Systematic Evaluation of BRCA1 and BRCA2 Variants

To test their new prioritization framework, the researchers focused on the BRCA1 and BRCA2 genes, which are well-known drivers of hereditary breast and ovarian cancer. Analyzing data from 2,172 genomic tests conducted across 13 institutions in Japan, the team identified 526 variants within these specific genes. Of those, nearly 400 were classified as VUS, far outnumbering the variants with known pathogenic effects. The researchers utilized ten different validated computational predictors to estimate how these specific genetic changes might disrupt protein function or RNA splicing, allowing them to narrow down hundreds of possibilities to ten high-priority candidates.