(W-099) Constructing a virtual control arm and evaluating operating characteristics using TGI metrics to support go/no-go decisions for single-arm Phase Ib/II combination studies

Rene Bruno, PhD – Senior Fellow, Genentech; Dan Lu, PhD – Distinguished Scientist, Genentech; Kenta Yoshida, PhD – Distinguished Scientist, Genentech

Objectives: Model-based tumor growth inhibition (TGI) metrics have been demonstrated to have superior operating characteristics (OC) than traditional endpoints such as objective response rate and progression free survival in supporting early decisions for randomized Phase Ib/II studies [1]. We aimed to extend this approach to inform the decision making of single-arm studies, by constructing a virtual control (VC) arm from patients in a separate historical clinical trial.

Methods: The VC mimicking a small Phase Ib/II study with early data cut was generated using data from the control arm of carboplatin and nab-paclitaxel combination (CnP) in IMpower130 or IMpower131 trials [2, 3]. First, considering a specified enrollment duration for 40 patients, a subset of patients were identified who enrolled within 10 months from the CnP arm. Then 40 patients were selected from this subset via propensity score matching (PSM) to baseline covariates of 40 patients resampled either from Atezolizumab, Bevacizumab, Carboplatin and Paclitaxel combination (ABCP) arm or BCP arm of IMpower150 trial [1], and the follow-up duration was cut to 24 weeks after the last enrolled patient. Next, TGI metrics were derived based on Stein and simplified TGI (sTGI) models [4, 5], and the TGI metrics geometric mean ratios (GMRs) between VC and the resampled data in either ABCP or BCP arm were calculated. The above process was repeated 500 times. Correct and incorrect go decisions were evaluated as the probability to achieve desired effect sizes in ABCP versus VC and BCP versus VC, respectively. Impacts of which study to construct the VC, matching strategy, type of TGI metrics on the OC curve were assessed.

Results: Among all scenarios explored, the OC curve using the VC generated with IMpower131 CnP arm is similar to the resampled control arm generated from the randomized BCP arm of IMpower150 study, when using the predicted tumor size ratio at 24 weeks to baseline (TR24) as the tumor metric from the Stein model. A correct go decision of 86.4% and incorrect go decision of 24.8% was computed when the TR24 GMR was < 0.9, which is the threshold value for a positive treatment effect. Using the IMpower131 CnP arm, with or without PSM, and the number of covariates (7 or 14) used for matching appear to have a relatively small impact. The TGI metrics of time to tumor growth (TTG) did not have a better OC than TR24 based on the CnP arm from the IMpower131 study. The results from the IMpower130 CnP arm showed an increased power but also increased Type I error.

Conclusions: We developed a methodology to generate VC that has a reasonable performance to inform go/no-go decision making using TGI metrics for single-arm Phase Ib/II studies. This method will be further evaluated in other studies and has the potential to be applied prospectively for on-going studies.

Citations:

[1] Bruno R, Marchand M, Yoshida K, et al. Tumor Dynamic Model-Based Decision Support for Phase Ib/II Combination Studies: A Retrospective Assessment Based on Resampling of the Phase III Study IMpower150 [published correction appears in Clin Cancer Res. 2023 Oct 13;29(20):4314]. Clin Cancer Res. 2023;29(6):1047-1055. doi:10.1158/1078-0432.CCR-22-2323

[2] West H, McCleod M, Hussein M, et al. Atezolizumab in combination with carboplatin plus nab-paclitaxel chemotherapy compared with chemotherapy alone as first-line treatment for metastatic non-squamous non-small-cell lung cancer (IMpower130): a multicentre, randomised, open-label, phase 3 trial. Lancet Oncol. 2019;20(7):924-937. doi:10.1016/S1470-2045(19)30167-6



[3] Jotte R, Cappuzzo F, Vynnychenko I, et al. Atezolizumab in Combination With Carboplatin and Nab-Paclitaxel in Advanced Squamous NSCLC (IMpower131): Results From a Randomized Phase III Trial. J Thorac Oncol. 2020;15(8):1351-1360. doi:10.1016/j.jtho.2020.03.028

[4] Stein WD, Gulley JL, Schlom J, et al. Tumor regression and growth rates determined in five intramural NCI prostate cancer trials: the growth rate constant as an indicator of therapeutic efficacy. Clin Cancer Res. 2011;17(4):907-917. doi:10.1158/1078-0432.CCR-10-1762

[5] Claret L, Gupta M, Han K, et al. Evaluation of tumor-size response metrics to predict overall survival in Western and Chinese patients with first-line metastatic colorectal cancer. J Clin Oncol. 2013;31(17):2110-2114. doi:10.1200/JCO.2012.45.0973