Development of a geriatric vulnerability score (GVS) in elderly advanced ovarian cancer (AOC) patients (pts) treated in first line: A prospective GINECO trial

Background: Two previous prospective GINECO studies in elderly patients have underlined the prognostic value of geriatric covariates (Co) on overall survival (OS) (Freyer G., et al. Ann Oncol. 2005 and Tredan O, et al Ann Oncol 2007). Methods: This open prospective trial was designed to confirm the impact of geriatric Co including psycho-geriatric Co on OS of elderly pts (≥70) with stage III-IV AOC treated in first line with 6 courses of carboplatin AUC5/3weeks. Geriatric Co were tested for their impact on OS in uni- and multivariate analyses. The best fitting proportional hazard model (GVS) was developed with the inclusion of major (MaC) and minor Co (miC). Results: From 08/2007 to 01/2010, 111 pts were included in 21 centres. A majority of pts displayed characteristics of vulnerability: age (median:78, range: 70-93, ≥80:41%), PS≥2: 43%, ≥3 major comorbidities: 27%, ≥4 comedications: 69%, ADL score<6: 55%, IADL score<25: 75%, HADS≥15:37%. A total of 74% of pts, however, completed planned chemotherapy. Gr3-4 haematological toxicity was observed in 50% of pts (thrombocytopenia [27%], anaemia [19%], and neutropenia [30%]) and gr3-4 non-haematological toxicity was fatigue (16%), anorexia (13%) and infection (10%). Median OS was 16.2 months (95%CI[14-21]). MaC were: albuminemia<35g/L; ADL score <6; IADL score <25 and miC: lymphopenia<1G/L; HADSμ14. The survival score=exp(0.320∗Number [MaC] +0.354∗Number[miC]) was validated upon a bootstrap analysis. Using a cutoff of 3, the simplified GVS score = Number[MaC] + Number[miC] discriminated two groups with significantly different OS :11.5 vs 21.7 months; HR=2.94; p<10-4, but also treatment completion rates: 65.4% vs 82.1%; OR=0.41; p=0.04; severe adverse events (SAE): 52.7% vs 28.6%; OR=2.8; p=0.009 and unplanned hospital admissions: 52.8% vs 30.3% OR=2.6; p=0.02. Conclusions: The GVS identified a group of pts at high risk of severe toxicity, early treatment stopping, unplanned hospitalization and poor outcome. GVS provides a useful tool to identify vulnerable pts in future elderly AOC trials.