SHBC1266

W.Y.GOH¹, Y.M.A.HUM¹

Tan Tock Seng Hospital¹

Older adults with cancer are at risk of under or over-treatment due to inadequate assessment. Survival estimation forms the cornerstone in oncological treatment decisions. However, there is dearth of evidence in geriatric oncology on short term (weeks to months) survival prognostication. The Prognostic Model for Advanced Cancer (PRO-MAC) was recently developed and validated for 90 days survival prognostication. We present a subgroup analysis of the PRO-MAC study focusing on older adults (age ≥ 65) with cancer to address the gap in short term survival prognostication.

This is a prospective study of patients with incurable cancer seen by palliative care service of a tertiary university hospital from June 2013 to March 2015. Pre-identified predictive variables were collected and patients followed up for 2 years or until death. Subjects were categorized into PRO-MAC groups and median survival compared by Kaplan-Meier survival curves with the log-rank test. Area under curve was calculated from the receiver operating characteristics curve for death at 30 days and 90 days.

592 subjects were included in this subgroup analysis. The median survival was 67 (95% CI: 52.1 – 82.0) days in group 1, 38 (95% CI: 32.0 – 44.0) days in group 2 and 14 (95% CI: 9.8 – 18.2) days in group 3 (p<0.001 by log-rank test). The AUC at 30 days and 90 days were 0.706 and 0.659 respectively.

PRO-MAC retained good discriminatory performance in geriatric oncology population for 30 days and 90 days survival. Further studies to externally validate PRO-MAC will help to ensure generalizability.