492P - Patient (pt) selection for immunotherapeutic early-phase clinical trials (ieCTs): A single phase I unit experience

Abstract

BackgroundSelecting which pts may benefit from ieCTs is challenging. Few prognostic indexes have been proposed so far and none qualifying as a predictor of response. The Gustave Roussy Immune Score (GRImS) identifies two prognostic categories based on three objective variables (albumin<3.5g/dL=1; LDH>ULN=1; NLR>6=1). MethodsWe retrospectively reviewed data of all pts enrolled into ieCTs at the Humanitas Cancer Center Phase I Unit between 2014 and 2019. A large series of demographic and clinical variables were correlated with overall survival (OS) and objective response rate (ORR) through univariate and multivariate analysis (UVA; MVA). Laboratory parameters were calculated either as baseline values and as dynamic six-weeks (6wks) changes. Finally, we explored the performance of the GRImS in our cohort. ResultsA total of 111 pts (M/F:63/48; median age: 62) with advanced solid tumors treated into ieCTs have been selected. The most frequent histologies were hepatocellular carcinoma (34%), lung carcinoma (22%), glioblastomas (13%). With a median follow-up (FU) of 14.3 months (mos), the OS was 12.9 mos, and the ORR 12.6%. In the UVA ECOG PS<1 (HR=0.53; IC 0.30-0.94; p=0.030) and higher baseline value of albumin (HR=0.40; IC 0.21-0.77; p=0.006) were significantly associated with a better OS, while baseline total protein level (OR=1.47; IC 1.24 - 2.10; p=0.030), and increase in lymphocytes count at 6wks (OR=1.02; IC 1.00-1.04; p=0.029) were predictive of response. In the MVA only higher baseline value of albumin seems to confirm its independent prognostic value (HR=0.48; IC 0.23-1.01; p=0.054), as well as baseline total protein level its predictive role (OR=1.52; IC 1-2.33; p=0.052). The GRImS resulted prognostic, with pts at low-risk (≤1) having a significant better OS compared to pts with a high-risk score (>1) (14.3 mos vs 7.3 mos; p=0.029), but not predictive. ConclusionsWe assessed the prognostic accuracy of GRImS in our ieCTs cohort. With limitations due to small sample size, short FU and few events recorded we identified additional static and dynamic variables with a potential prognostic and predictive relevance to be further explored in larger series and to be eventually included in new scores. Legal entity responsible for the studyThe authors. FundingHas not received any funding. DisclosureM. Simonelli: Advisory / Consultancy: AbbVie. A. Santoro: Advisory / Consultancy: Bristol-Myers-Squibb; Advisory / Consultancy: Servier; Advisory / Consultancy: Gilead; Advisory / Consultancy: Pfizer; Advisory / Consultancy: Eisai; Advisory / Consultancy: Bayer; Advisory / Consultancy: Merck Sharp & Dohme; Speaker Bureau / Expert testimony: Takeda; Speaker Bureau / Expert testimony: Bristol-Myers-Squibb; Speaker Bureau / Expert testimony: Roche; Speaker Bureau / Expert Testimony: Abbvie; Speaker Bureau / Expert testimony: Amgen; Speaker Bureau / Expert testimony: Celgene; Speaker Bureau / Expert testimony: Servier; Speaker Bureau / Expert testimony: Gilead; Speaker Bureau / Expert testimony: AstraZeneca; Speaker Bureau / Expert testimony: Pfizer; Speaker Bureau / Expert testimony: Arqule; Speaker Bureau / Expert testimony: Lilly; Speaker Bureau / Expert testimony: SANDOZ. All other authors have declared no conflicts of interest.