The proper selection of patients in phase I studies is crucial in order to guarantee not only a safer approach to patients but also to obtain reliable results. Recently, Gustave Roussy Immune Score (GRIm-score) predicted overall survival for patients involved in phase I immune-oncology (IO) trials in a more accurate way when compared to other classical scores. The goal of our study is to validate the GRIm-score in our institution.
We retrospectively studied 171 patients (pts) treated in phase IO trials at our institution between January 2012 and August 2017. Variables such as haemoglobin, neutrophil-to-lymphocyte ratio (NLR), albumin level, LDH, gender, number of previous lines and tumour type were assessed. GRIm score was calculated as follows: LDH > ULN (+1), Albumin < 35 g/dl (+1), NLR>6 (+1). According to GRIm score patients were stratified into low risk (0-1) or high risk (2-3). Overall survival (OS) and OS at 90 days were studied.
A total of 171 pts (M/F: 115/56) were evaluated. The median age was 59.53 (95%CI 57.68-61.38) Mean number of previous lines were 2.25 (95%CI 2.03-2.48). The most frequent tumour histologies were colorectal carcinoma (28.24%), non-small cell lung cancer (15.29%), renal carcinoma? (11.18%) and melanoma (8.24%). Patients with 0 – 1 risk factors were classified as low-risk (n = 114) whereas patients with 2-3 risk factors were considered high-risk (n = 57). Low-risk patients presented an OS of 13.8 months (95% CI 8.51-20.44) whereas high-risk patients presented an OS of 3.02 months (95% CI 2.3-4.3, HR = 2.66; 95% CI1.76-4.03, p < 0.001). When predicting overall survival at 90 days, we found that patients with low-risk GRIm score punctuation were more prone to be alive at 90 days when compared to patients with high-risk score (OR = 6.61; 95% CI 3.06-14.29, p value <0.001).
We have externally validated the GRIm score in an unselected population. Our data suggests that GRIm score predicts overall survival in patients enrolled in phase I IO trials and may be useful when selecting patients. The correct selection of the patients to be enrolled in clinical trials can help to improve the quality of the trials and improve toxicity profile and survival outcomes.
Clinica Universidad de Navarra
Has not received any funding
I. Melero: Consultant for Roche, BMS, Lilly, Biotech, Alligator and TurkTurk; Research grant from Roche -Genentech BMS, Alligator. All other authors have declared no conflicts of interest.