Validation of the modified Glasgow Prognostic Score (mGPS) in recurrent ovarian cancer (ROC) : Analysis of patients enrolled in the GCIG Symptom Benefit Study (SBS)

BACKGROUND: Modified Glasgow Prognostic Score (mGPS) is predictive of survival in many advanced cancers, but has not been evaluated in recurrent ovarian cancer (ROC). The aim was to determine validity of mGPS in ROC, investigate its associations with health related quality of life (HRQL) and ECOG performance status (PS). METHODS: mGPS is based on serum C reactive protein (CRP) and albumin, with scores ranging from 0 (least) to 2 (most). HRQL was measured with EORTC QLQ C-30 and OV-28. χ2 tests for trend were used to examine the relationship between HRQL, PS and mGPS. Cox proportional hazards regression was used to assess associations between mGPS, HRQL, clinicopathological factors, and overall survival (OS). RESULTS: Inflammatory markers were available in 516 of 948 patients in GCIG SBS. 200(39%) had potentially platinum sensitive ROC with ≥3 lines of chemotherapy, 316(61%) had platinum resistant ROC. 282(55%), 123(24%), 111(22%) had mGPS of 0, 1, 2, respectively. Median OS (months) was 18.1, 9.6, and 6.6 for mGPS 0, 1, and 2 respectively. mGPS was an independent predictor of OS after adjusting for PS and platinum sensitivity (p<0.001). mGPS remained a predictor of OS after adjusting for physical function, role function, global health status, abdominal/GI symptoms, and multiple clinicopathologic factors (p=0.02). Worse PS and higher mGPS were associated with poorer HRQL (p<0.001). Higher mGPS was associated with worse HRQL, independent of PS. CONCLUSION: The mGPS is an independent predictor of OS in ROC after adjusting for HRQL and clinicopathological factors. Higher mGPS is associated with worse HRQL independent of PS. mGPS is simple, inexpensive and may be suitable for clinical practice, clinical trial patient selection and stratification.