Frailty and treatment outcome in advanced gastro-oesophageal cancer: An exploratory analysis of the GO2 trial

Abstract

IntroductionResearch into the optimal management of frail patients with cancer is limited and treatment decision-making in this cohort can be difficult. A number of measures have been developed to assess frailty, but few studies explore the correlation between frailty measures and cancer treatment outcomes. MethodsThis retrospective cohort study is an exploratory analysis of the GO2 randomised controlled trial. GO2 recruited both older and frail younger patients commencing first-line palliative chemotherapy for advanced gastro-oesophageal (aGO) cancer. This analysis aims to explore the correlation between baseline frailty and treatment outcome. Baseline frailty measures were derived from clinical data and included ECOG Performance Status (PS), the GO2 Frailty Score (GO2FS), Geriatric-8 (G8), Cancer and Aging Research Group (CARG) toxicity score and a ‘modified’ Rockwood Clinical Frailty Scale (mCFS). Novel patient-centred composite measure Overall Treatment Utility (OTU) was the primary endpoint. Ordinal logistic regression was undertaken to give odds ratios for poor vs good/intermediate OTU. Secondary endpoints were progression-free and overall survival. Models were adjusted for age, sex, histology, metastases, Trastuzumab and renal/hepatic function. ResultsIn GO2, 514 patients were randomised between three chemotherapy dose-levels; all of these patients were assessed for OTU and are included in this analysis. Worse GO2FS, mCFS and G8 scores all had a statistically significant association with poor (vs good/intermediate) OTU, progression and death, which persisted after adjustment. Adjusted odds ratios for poor OTU amongst those with the worst GO2FS and mCFS and best G8 scores were as follows: 1.85 (95% confidence interval [CI] 1.20–2.88) for GO2FS ≥3 (‘severely frail’), 1.72 (1.19–2.50) for mCFS 5+ (‘frail’) and 0.57 (0.32–1.00) for G8 > 14 (‘normal’). Worse ECOG PS and CARG scores did not have a statistically significant association with poor OTU/progression/death. ConclusionIn this study, frailty identified via GO2FS, mCFS and G8 conveyed a statistically significant increased risk of worse treatment outcome in older and frail younger patients with aGO cancer. Frailty assessment provides information over and above PS and should be integrated alongside routine assessments in research and clinical practice. In the absence of prospective data, frailty measures can be derived retrospectively to build the evidence base around optimal care of frailer patients.