Whole-genome sequencing was found to outperform current commercial methods for detecting homologous recombination deficiency (HRD) and subsequently identifying patients likely to benefit from PARP inhibitor cancer treatments, according to a study published in Communications Medicine.
Investigators from Weill Cornell Medicine and NewYork-Presbyterian analyzed hundreds of tumor samples through a precision medicine initiative also including Illumina. The results were used to train and validate an algorithm detecting homologous recombination deficiency, a DNA-repair defect. Tumors with this defect are vulnerable to PARP inhibitors, which impair DNA repair and cause lethal damage accumulation.
“A comprehensive analysis of the entire genome has advantages compared with traditional, targeted detection strategies for predicting homologous recombination deficiency,” said senior author Juan Miguel Mosquera.
Clinicians have traditionally relied on BRCA1 and BRCA2 mutations—the most common causes of this repair defect—to select patients for PARP inhibitors, particularly those with breast, ovarian, pancreatic, and prostate cancers. However, research shows that many other gene mutations can also disrupt this repair process—and in recent years, WGS, which can detect these broader changes, has become affordable enough for routine use.
The team used 305 samples from Weill Cornell and NewYork-Presbyterian patients with various cancers to train an algorithm that was developed by Isabl. The algorithm looks for a genome-wide variety of DNA defects that are known to be associated with homologous recombination repair deficiency. They then validated the algorithm using a cohort of 556 cancers and tested it against commercial methods using an additional 212 tumor samples.
The algorithm identified the deficiency in 21% of breast tumors, 20% of pancreatic and bile duct tumors, and 17% of gynecological tumors. Notably, 24% of cases lacked BRCA1 or BRCA2 mutations, revealing diverse genetic drivers.
The algorithm also corrected several “false negative” and “false positive” results from commercial assays that mismatched patient outcomes.
Plans call for larger studies to establish the algorithm as a more accurate way to guide treatment decisions for cancer patients.
Publication Details: Al Assaad, M., Hadi, K., Levine, M.F. et al. Whole genome sequencing approach to assess homologous recombination deficiency in a pan-cancer cohort. Commun Med 6, 5 (2026). https://doi.org/10.1038/s43856-025-01308-5
Investigators from Weill Cornell Medicine and NewYork-Presbyterian analyzed hundreds of tumor samples through a precision medicine initiative also including Illumina. The results were used to train and validate an algorithm detecting homologous recombination deficiency, a DNA-repair defect. Tumors with this defect are vulnerable to PARP inhibitors, which impair DNA repair and cause lethal damage accumulation.
“A comprehensive analysis of the entire genome has advantages compared with traditional, targeted detection strategies for predicting homologous recombination deficiency,” said senior author Juan Miguel Mosquera.
Clinicians have traditionally relied on BRCA1 and BRCA2 mutations—the most common causes of this repair defect—to select patients for PARP inhibitors, particularly those with breast, ovarian, pancreatic, and prostate cancers. However, research shows that many other gene mutations can also disrupt this repair process—and in recent years, WGS, which can detect these broader changes, has become affordable enough for routine use.
The team used 305 samples from Weill Cornell and NewYork-Presbyterian patients with various cancers to train an algorithm that was developed by Isabl. The algorithm looks for a genome-wide variety of DNA defects that are known to be associated with homologous recombination repair deficiency. They then validated the algorithm using a cohort of 556 cancers and tested it against commercial methods using an additional 212 tumor samples.
The algorithm identified the deficiency in 21% of breast tumors, 20% of pancreatic and bile duct tumors, and 17% of gynecological tumors. Notably, 24% of cases lacked BRCA1 or BRCA2 mutations, revealing diverse genetic drivers.
The algorithm also corrected several “false negative” and “false positive” results from commercial assays that mismatched patient outcomes.
Plans call for larger studies to establish the algorithm as a more accurate way to guide treatment decisions for cancer patients.
Publication Details: Al Assaad, M., Hadi, K., Levine, M.F. et al. Whole genome sequencing approach to assess homologous recombination deficiency in a pan-cancer cohort. Commun Med 6, 5 (2026). https://doi.org/10.1038/s43856-025-01308-5