Comparison of prognostic scores in early phase clinical trials: a 10-year single centre Australian experience

Background/Aim: Early phase clinical trials (EPCTs) assess the tolerability of novel anti-cancer therapeutics in patients with advanced malignancy. Patient selection is important given the modest clinical benefit and time commitments for trials. Prognostic scores have been developed to facilitate identification of high-risk patients. This study aimed to compare five prognostic scores to predict survival for patients on an EPCT. Patients and Methods: We performed a retrospective review of patients enrolled in EPCT at Liverpool Hospital, Sydney, from 2013 to 2023. Demographic, biochemical, and survival data were collected from electronic medical records. The score from five prognostic scoring systems (Royal Marsden hospital, MD Anderson Cancer centre, Gustave Roussy Immune, MD Anderson Immune Checkpoint Inhibitor and Princess Margaret Hospital Index) were calculated. Overall survival was measured using the Kaplan-Meier method and predictive discrimination was assessed using Harrell's c-index. Results: A total of 218 patients across 36 EPCTs were included. The median overall survival was 9.8 months with 22% of patients dying in less than 90 days. Seventeen to thirty-four percent of patients were categorised as high-risk. The MDACC score obtained the highest predictability for overall survival for the whole cohort (c-index=0.67, 95%CI=0.62-0.72) and the immunotherapy-based cohort (c-index= 0.65, 95%CI=0.59-0.71). However, all scores performed similarly with a significant overlap in the confidence intervals. Conclusion: Our retrospective audit confirms the utility of prognostic scores to predict survival in an Australian EPCT cohort, with similar predictive discrimination across various scoring systems. Integration of these prognostic tools into EPCT screening processes may optimise benefits and reduce risks associated with EPCTs.