Precision Promise (PrP) Bayesian platform trial for metastatic pancreatic cancer (mPDAC): Results of the first experimental arm, SM-88 as second line therapy

• Published in: Journal of clinical oncology Vol. 42; no. 3_suppl; p. 675
• Main Authors: wang-gillam, Andrea, Varghese, Anna M., Ko, Andrew H., Hendifar, Andrew Eugene, Ocean, Allyson J., McGlothlin, Anna, Graves, Todd, Berry, Scott M., Detry, Michelle A., Fitzgerald, Mark, Bosse, Anna, Moravek, Cassadie, Feehan, Kelly, Howland, Carrie Meghann, Berry, Donald A., Simeone, Diane M., Picozzi, Vincent J.
• Format: Journal Article
• Language: English
• Published: American Society of Clinical Oncology 20.01.2024
• ISSN: 0732-183X; 1527-7755
• DOI: 10.1200/JCO.2024.42.3_suppl.675

675Background: PrP is a phase 3 Bayesian adaptive platform trial for efficiently testing multiple arms against a common control. Its design is what FDA calls a Complex Innovative Design (www.fda.gov/media/130897/download). Experimental arms in stage 1 are adaptively randomized against other arms. The maximum size of an arm's stage 1 is 100 subjects. Strong efficacy in stage 1 leads to a fixed randomization stage 2. Stages 1 and 2 are combined for the arm's final analysis and, if positive, would be submitted for drug registration. SM-88 (D,L-alpha-metyrosine) is a dysfunctional, racemic tyrosine mixture designed to interrupt PDAC cellular metabolism and induce oxidative stress. Methoxsalen, phenytoin and sirolimus are added to SM-88 to potentially enhance oxidative stress, and form a 4-drug oral regimen. SM-88 was selected for testing in 2nd line mPDAC using the PrP platform. Methods: Primary endpoint overall survival (OS) is analyzed based on modified intention to treat (mITT). Randomization is 7:3 to experimental and control arms. Efficacy is defined by OS hazard ratio (HR, experimental vs control). All statistical measures in PrP are Bayesian. Superiority (HR < 1) is assessed monthly and is claimed should the Bayesian probability of benefit be ≥ 98%. Futility analyses occur monthly once 50 subjects have accrued to the arm. Accrual ends for futility if Bayesian predictive power (PP) of eventual success in PrP is < 20% for all the arm's signatures. Follow-up continues for 12 months after arm accrual stops. PrP experimental arm designs are flexible: an arm's protocol appendix can supersede the master protocol. However, type I error is controlled at < 0.025, as shown by simulation. Results: SM-88 entered PrP in Apr 2020. As the first experimental arm in PrP, it was randomized 7:3 against control arms. Pooled controls for SM-88 were gemcitabine/nab-paclitaxel and mFOLFIRINOX at standard doses, since neither therapy was a backbone for SM-88. Between 4/2020 - 10/2021, 142 subjects were screened, 73 subjects started therapy and are included in the mITT cohort. 55 subjects were enrolled in the SM-88 arm and 18 in control arms. Median age was 65, 41% were female and 67% had ECOG-1. As per protocol, accrual stopped at its first futility analysis in Nov 2021 based on Predictive power (PP) 0.0001 (see table), which is less than the protocol bound 0.20. SM-88 toxicity was mild. Interim results at accrual stop and final results (12 month follow up) shown below. mOS was 4.1m for SM-88 vs 8.1m for control. Conclusions: Leveraging small sample sizes of SM-88 and controls, PrP efficiently and compellingly concluded SM-88 futility in 2nd line mPDAC. PrP continues to assess other therapies, utilizing time-adjusted, as well as concurrently randomized, controls. Clinical trial information: NCT04229004. AnalysisN (mITT)Mean HRSD HR95%PIP(HR<1)PPInterim54/172.791.051.32-5.32<0.0010.0001Final55/181.720.441.02-2.750.020NA