Life expectancy estimation is critical to select patients in clinical trials, notably in the phase I setting. However, most protocols use subjective criteria to meet this goal. Several scores have been developed from retrospective studies, but their utility in the immuno-oncology (IO) context remains unknown. The recent recognition of fast progressors under IO treatment further underscores the need for more objective prediction.
NSCLC and UBC patients (n = 972) prospectively enrolled in 3 phase I trials investigating durvalumab (anti-PDL1) ± tremelimumab (anti-CTLA4) across 160 centers were analyzed. We assessed the variability in the 8- and 12-week life expectancy rate across centers. Clinicopathological variables were used to train (n = 648) and validate (n = 324) a tool (FastProgIO) predicting 12-week life expectancy using a multivariate regression method. We compared FastProgIO to existing published scores (RMH, GRIM, LIPI). The performance was assessed by time dependent true positive rate (TPR) and false positive rate (FPR).
Substantial variability was observed across sites with 26% and 65% of centers having enrolled > 15% of patients with life expectancy < = 8 and 12 weeks, respectively. FastProgIO includes neutrophils, AST, alkaline phosphatase and hemoglobin as the predictive markers. Overall, the TPR of FastProgIO was superior across a 4 to 12 weeks range to RMH, GRIM and LIPI. At 12 weeks, the TPR for FastProgIO was 73%, (90% CI: 67-80%) versus 69%, 65% and 41% for RMH, GRIM and LIPI, respectively. Furthermore, the FPR of FastProgIO (11%, 90%CI: 9.3-13%) was comparable to LIPI (10%), but better controlled than RMH and GRIM (20%, 17%).
The use of subjective criteria to estimate the 12-week life expectancy of patients enrolled in IO trials is clearly suboptimal and can lead to ethical, scientific, medical, and public health conundrums. In this large cohort of patients, FastProgIO appears superior to RMH, GRIM and LIPI to define life expectancy of patients based on easily available baseline clinical characteristics, and can be used to better select patients for IO clinical trials. An extended validation of this model in other tumor types is ongoing and will be presented.
Clinical trial identification
Study 006: NCT02000947 December 4, 2013 Study 1108: NCT01693562 September 26, 2012 Study 10: NCT02261220 October 10, 2014.
Legal entity responsible for the study
C. Massard: Advisory board member: Amgen, Astellas, AstraZeneca, Bayer, Celgene, Genentech, Ipsen, Janssen, Lilly, Novartis, Pfizer, Roche, Sanofi, and Orion. N.H. Segal: Honoraria: MedImmune; Consulting or advisory role: Bristol-Myers Squibb, Pfizer, AstraZeneca, MedImmune, MacroGenics, Imugene, Roche, Genentech, Kyocera, Amgen, Calithera Biosciences; Research funding: MedImmune, Bristol-Myers Squibb, Pfizer, Roche, Genentech, Merck. D.C. Cho: Consultant or advisory board member: Pfizer, BMS, Genentech. Exelixis, Prometheus. V.A. Papadimitrakopoulou: Grant research support and consultancy or advisory board member role: AstraZeneca. N.A. Rizvi: Honoraria or consultants fees: Roche, AstraZeneca, Novartis, Merck, Pfizer, Lilly, and BMS; Equity ownership: Gritstone Oncology and Armo Biosciences. B.C. Cho: Research grant support: Novartis, AstraZeneca, Yuhan, Ono/BMS, MSD, and Bayer; Consultancy or advisory board member: AstraZeneca, Roche, Boehringer Ingelheim, Yuhan, BMS, MSD, and Novartis; Speaker's bureau: AstraZeneca, BMS, MSD, and Novartis; Honorarium: AstraZeneca, Roche, Boehringer Ingelheim, Yuhan, BMS, MSD, and Novartis. L. Yu H. Yang: Employment: MedImmune; Equity ownership: AstraZeneca. H-J. Hsieh: Employment: Genentech, AstraZeneca, and MedImmune; Equity ownership: Genentech. J. Zhang, W. Zhao, G. Gao, X. Guo, S. Abdullah, J. Englert, M. Dar, L. Roskos: Employment: MedImmune and AstraZeneca; Equity ownership: AstraZeneca. J-C. Soria: Employment: MedImmune and AstraZeneca; Equity ownership: AstraZeneca. C. Ferte: Employment: MedImmune and Astrazeneca; Honoraria: Roche, MSD, Merck, and BMS; Consultancy or advisory role: Roche, MSD, Merck, Amgen, and BMS; Travel expenses: BMS, MSD, Merck, Amgen. S.J. Antonia: Research grant support: Novartis; Consultancy or advisory role: BMS, Novartis, Merck; Honoraria: BMS, Novartis, Merck, CBMG, Boehringer Ingelheim, AstraZeneca, MedImmune, Memgen.