Background: ASCO guidelines recommend that geriatric assessment (GA) should be performed in all older adults with cancer. However, GA is labor- and time-intensive, hence the uptake is poor, especially in crowded resource poor-settings. There are no data correlating GA with overall survival (OS) outcomes from the Indian subcontinent. Methods: A prospective observational study in the geriatric oncology clinic of the Tata Memorial Hospital in Mumbai, India. Patients aged 60 years and above, with cancer who underwent a GA were enrolled. The domains assessed included: function (basic and instrumental activities of daily living, timed-up-and-go), nutrition (body mass index, unintentional weight loss, mini-nutritional assessment), comorbidities, cognition, psychological (depression, anxiety), social support, and medication (polypharmacy and potentially inappropriate medications). Patients with > 2 deranged GA domains were considered frail. Results: Between June 2018 and January 2022, 909 patients were enrolled. The median age was 69 (IQR, 60-88) years. Common malignancies included lung (40%), esophagus (21%) and head and neck (11%); 53% had metastatic disease. 80% had > 2 impaired domains in GA patients had vulnerabilities in a median of 3 (IQR, 0-5) domains. Median OS in fit patients based on the GA was 17.5 (95% CI, 13.9-21.0) months vs 12.1 (95% CI, 10.1-14.0) months in frail patients, (HR 0.66; 95% CI, 0.49-0.88, p = 0.005), which remained significant after adjusting for age, sex, and stage (HR, 0.71; 95% CI: 0.53-0.94, p = 0.021). In the multivariate analysis (Table), the domains that were predictive of survival were nutrition (HR: 0.65, 95% CI: 0.47-0.92, p = 0.014), cognition (HR: 0.65; 95% CI: 0.46-0.91, p = 0.012) and fatigue (HR: 0.74, 95% CI: 0.56-0.98, p = 0.038). Conclusions: In older Indian patients with cancer, GA is a powerful prognosticator of survival. In settings where a complete GA is not possible, nutrition, cognition, and fatigue should be the minimum domains assessed. Clinical trial information: CTRI/2020/04/024675.

HR: hazard ratio; CI: confidence interval