Abstract 4916: Improved statistical approaches that account for pseudoprogression in preclinical studies of RRx-001 with immunotherapies

Background and Objective Immuno-oncology agents are associated with apparent tumor growth due to immune infiltration leading, in some cases, to misinterpretation of radiological scans that can result in premature discontinuation of effective therapeutic agents. In the clinic, this resulted in the introduction of modified immune-related response criteria. Preclinically, in the absence of radiology or immunohistochemistry, traditional tumor growth delay may similarly lead to misinterpretation of progression even in the context of a true response. More powerful statistical strategies are therefore required to distinguish between true and pseudoprogression. RRx-001 is a novel immuno-epigenetic agent in Phase 2 with promising single agent and combination activity that is associated with clinical pseudo progression in some cases. Previously, we tested the hypothesis that RRx-001 + anti-PD-L1 would increase the complete response rate in a myeloma preclinical model. Time-to-event statistical analyses methods that account for the study design features and capture the longitudinal and panoramic aspect of pseudo progression were used to test superiority of the combination in lieu of traditional, mean-based mixed effects modeling approaches that did not show statistical significance. Nonparametric p-values for the difference of cumulative incidence rates of time to ≥50% tumor growth reduction and its associated restricted mean survival times are illustrated. Methods and Results Analysis of the raw data growth curves from a study comparing RRx-001, anti-PD-L1 alone and in combination in mice with J558L myeloma tumors showed an initial increase followed by a reduction in size suggesting a pseudoprogressive response. In light of this observation, the study-end probability of tumor volume reduction (≥50%) was chosen as a suitable endpoint. The difference in the Kaplan-Meier estimate for this endpoint between the combination and anti PD-L1 alone was 50% (95% CI: 10.8%, 89.2%). The two-sided nonparametric (Cox model) p-value was 0.0124, a statistically significant finding in favor of the combination group as compared to the PD-L1 group alone. Conclusion Tumor flare or pseudo progression describes a transient increase in tumor volume due to immune cell infiltration and inflammation/edema, which resolves over time. Pseudo progression is visualizable in the clinic with radiological scans; however, pre-clinically it is difficult, if not impossible, to distinguish between pseudo and true progression with traditional statistical methods. Kaplan-Meier description of time-to-volume reduction (≥50%) coupled with Cox9s proportional hazards model follows the data longitudinally and therefore permits an analysis of immune infiltration resolution, and make it an improved method for analysis of pre-clinical experiments with immuno-oncology agents. Citation Format: Nacer E. Abrouk, Jan Scicinski, Bryan Oronsky, Susan J. Knox, Donna Peehl, Shoucheng Ning, Gary R. Fanger. Improved statistical approaches that account for pseudoprogression in preclinical studies of RRx-001 with immunotherapies. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4916.