When should homologous recombination deficiency (HRD) testing be performed in patients with newly diagnosed high‑grade serous or endometrioid ovarian carcinoma, including timing relative to surgery, pathology review, first‑line platinum‑taxane chemotherapy, and at recurrence before PARP‑inhibitor maintenance?

When to Order HRD Testing in Ovarian Cancer

Order HRD testing (both germline/somatic BRCA and genomic scar-based assays) immediately upon histologic confirmation of high-grade serous or endometrioid ovarian carcinoma, ideally before or concurrent with first-line platinum-taxane chemotherapy, to inform PARP inhibitor maintenance decisions. 1

Timing Algorithm for HRD Testing

At Initial Diagnosis (Preferred Approach)

Initiate testing as soon as histologic diagnosis is confirmed:

• Germline BRCA1/2 testing should begin immediately after pathologic confirmation of high-grade serous or endometrioid ovarian carcinoma, even before surgery if biopsy confirms diagnosis 1
• Do not delay primary chemotherapy for genetic counseling referral or test results, as delays between surgery and chemotherapy worsen outcomes 1
• Somatic BRCA and HRD genomic scar testing should be ordered once adequate tumor tissue is available from surgery or biopsy 1
• The 18-week minimum duration of first-line platinum-based chemotherapy provides sufficient time for test completion before maintenance therapy decisions 1

Specific Testing Components

Both germline and somatic testing provide complementary information:

• Germline BRCA testing identifies inherited mutations (12-15% of cases) and has implications for family members 1
• Somatic/tumor BRCA testing identifies additional 5-7% of patients with acquired mutations 1
• Genomic scar-based HRD assays (GIS or LOH scores) identify BRCA wild-type patients with HRD who benefit from PARP inhibitors 1

Testing Strategy by Clinical Setting

First-line maintenance setting (Stage II-IV disease responding to primary treatment):

• Germline and somatic BRCA mutation testing is routinely recommended to identify patients who should receive PARP inhibitor maintenance 1
• Validated scar-based HRD testing is reasonable to establish magnitude of benefit in BRCA wild-type patients and identify those least likely to benefit 1
• Testing should be completed before finishing 6+ cycles of platinum-based chemotherapy to inform maintenance decisions 1

Platinum-sensitive recurrence:

• BRCA mutation testing and validated scar-based HRD tests are reasonable to predict magnitude of PARP inhibitor benefit for risk-benefit assessment of maintenance therapy 1
• If not performed at initial diagnosis, testing should occur at first recurrence before initiating platinum-based chemotherapy 1

Critical Caveats and Pitfalls

Tissue sample considerations:

• HRD assays reflect the time point when sample was obtained and do not provide dynamic readout 1
• Archival samples from initial diagnosis/debulking are typically used, but ideally testing at recurrence should use contemporary tissue 1
• Current assays do not detect reversion mutations that restore homologous recombination proficiency and cause PARP inhibitor resistance 1

Test selection:

• Use FDA-approved tests or validated tests performed in CLIA-approved facilities 1
• Myriad myChoice was the validated assay used in pivotal first-line trials (PAOLA-1, PRIMA, VELIA) 1
• Foundation Medicine T5 NGS assay was used in ARIEL3 trial for recurrent disease 1

Limitations of current HRD testing:

• Poor negative predictive value: BRCA wild-type and HRD-negative patients still derive some benefit from PARP inhibitors, particularly in platinum-sensitive disease 1
• Different assays and cutoffs were used across trials, making direct comparisons difficult 1
• The NCCN panel does not recommend any particular approach for determining HRD in non-BRCA mutated patients due to ongoing investigation 1

Histologic requirements:

• HRD testing is most validated for high-grade serous carcinoma 1
• High-grade endometrioid carcinoma is included in most trial populations 1
• Other histologic subtypes are unlikely to harbor HRD and should not routinely undergo testing 2