Objective

There is a need to develop and validate biomarkers for treatment response and survival in tubo-ovarian high-grade serous carcinoma (HGSC). The chemotherapy response score (CRS) stratifies patients into complete/near-complete (CRS3), partial (CRS2), and no/minimal (CRS1) response after neoadjuvant chemotherapy (NACT). Our aim was to review current evidence to determine whether the CRS is prognostic in women with tubo-ovarian HGSC treated with NACT.
Methods

We established an international collaboration to conduct a systematic review and meta-analysis, pooling individual patient data from 16 sites in 11 countries. Patients had stage IIIC/IV HGSC, 3–4 NACT cycles and >6-months follow-up. Random effects models were used to derive combined odds ratios in the pooled population to investigate associations between CRS and progression free and overall survival (PFS and OS).
Results

877 patients were included from published and unpublished studies. Median PFS and OS were 15 months (IQR 5–65) and 28 months (IQR 7–92) respectively. CRS3 was seen in 249 patients (28%). The pooled hazard ratios (HR) for PFS and OS for CRS3 versus CRS1/CRS2 were 0·55 (95% CI, 0·45–0·66; P < 0·001) and 0·65 (95% CI 0·50–0·85, P = 0·002) respectively; no heterogeneity was identified (PFS: Q = 6·42, P = 0·698, I2 = 0·0%; OS: Q = 6·89, P = 0·648, I2 = 0·0%). CRS was significantly associated with PFS and OS in multivariate models adjusting for age and stage. Of 306 patients with known germline BRCA1/2 status, those with BRCA1/2 mutations (n = 80) were more likely to achieve CRS3 (P = 0·027).
Conclusions

CRS3 was significantly associated with improved PFS and OS compared to CRS1/2. This validation of CRS in a real-world setting demonstrates it to be a robust and reproducible biomarker with potential to be incorporated into therapeutic decision-making and clinical trial design.