Prognostic Factors and Outcome of Patients with Hematological Malignancies in Phase I Trials: The Gustave Roussy Scoring System

Background: Phase I clinical trials are conducted to determine the recommended phase II dose through safety and pharmacokinetic assessments. Potential clinical benefit is also a key aim. To improve patient's selection and identify patients most likely to benefit from these studies in hematological malignancies (HM), we designed and then validated in an independent cohort, a prognostic score based on simple variables.

Patients and Methods: We retrospectively collected data from 82 consecutive patients enrolled in 14 phase I trials at Gustave Roussy (GR) between January 2008 and February 2012 (cohort 1). A simple scoring system was established through multivariate analysis (MVA) of multiple potential prognostic factors (age, gender, BMI, smoking status, comorbidities, histological subtype, WHO performance status (PS), prior autologous stem cell transplant, prior radiotherapy, LDH, serum albumin and protein levels). We then prospectively collected data from 88 consecutive patients treated in 17 phase I trials in the GR phase I department from February 2012 to May 2014, to validate this prognostic score on an independent cohort (cohort 2). Stratification was done according to histology given the heterogeneity of diseases.

Results: All patients had progressive HM after a median of 2 prior systemic standard therapeutic lines in cohort 1. The distribution of HM was: acute myeloid leukemia (AML) 23%, indolent non Hodgkin lymphoma (iNHL) 22%, myelofibrosis (MF) 21%, aggressive non-Hodgkin lymphoma (aNHL) 16%, chronic lymphoid leukemia (CLL) 10%, Hodgkin lymphoma (HL) 6% and multiple myeloma (MM) 2%. Median age was 73 years (range: 27-93) and 56 patients (67%) had more than one comorbidity. Best overall response rate (ORR) and disease control rate (DCR: objective response and stable disease rates) were 20% and 64 % respectively. Within a median follow up of 26 months, median progression-free survival (PFS) and overall survival (OS) were respectively 5 (CI95%: 4; 8) and 17 months (CI95%: 12; 26). One toxicity-related death (2%) occurred. Thirty-seven patients (45%) experienced grade 3 or 4 adverse events (AE) and 9 patients (11%) a dose-limiting toxicity (DLT). MVA identified PS > 0 at inclusion, baseline albumin ≤ 35 g/l and histological subtype as prognostic factors for OS. We defined the GR score using this MVA data: +1 if PS > 0, +1 if albumin ≤ 35 g/l. Patients with a GR score = 0 experienced significantly better OS compared to patients with a score = 1 and a score = 2 (37 months versus 17 months versus 8 months; p = 0.003).

Cohort 2 patients were characterized by a median age of 64 years (range: 23-84) and a similar comorbidity rate (61%). Distribution among hematological entities was: 22% iNHL, 21% AML, 19% aNHL, 14% MM, 13% HL, 7% MF and 5% CLL. Patients received a median of 3 systemic therapeutic lines prior to phase 1 inclusion. Best ORR and DCR were respectively 28% and 57%. Within a median follow up of 11 months, median PFS and OS were 3 (CI95%: 1.5; 5) and 20 months (CI95%: 12; NA) respectively. No toxicity-related death occurred. Forty-seven patients (53%) experienced grade 3 or 4 toxicity and 1 patient (1%) a dose-limiting toxicity. Pre-established GR score was predictive of OS in this validation cohort (p=0.033).

Main study discontinuation reason was progressive disease (75% and 80% in cohorts 1 and 2, respectively). Drug related toxicity was responsible for study discontinuation in 9% and 14% of cases in cohorts 1 and 2 respectively.

The GR score distinguishes patients most likely to remain on study for more than 6 weeks, considered as the minimum required period for response and toxicity evaluation (classical DLT and targeted therapies associated long term toxicities). Among patients who discontinued prematurely clinical trials, GR score was ≥ 1 in 88% and 89% of cases respectively in cohort 1 (Cochran-Armitage: p= 0.02 and p=0.06) and cohort 2 (p=0.036).

Conclusion: Our data demonstrate safety and efficacy of phase I trials in a significant number of relapsed/refractory HM patients with a clinical benefit achieved in more than half of patients. Response rates are higher than previously reported in solid tumors (Arkenau HT et al. 2008), thus encouraging HM patients inclusions in phase I trials. The simple GR score based on PS and albumin offers a valuable selection tool enabling OS and early trial discontinuation prediction.