Nolan A Wages1, Camilo E Fadul2. 1. Division of Translational Research & Applied Statistics, Department of Public Health Sciences, University of Virginia, Charlottesville, VA, USA. 2. Division of Neuro-Oncology, Department of Neurology, University of Virginia, Charlottesville, VA, USA.
Abstract
BACKGROUND/AIMS: Dose feasibility is a challenge that may arise in the development of adoptive T cell therapies for cancer. In early-phase clinical trials, dose is quantified either by a fixed or per unit body weight number of cells infused. It may not be feasible, however, to administer a patient's assigned dose due to an insufficient number of cells harvested or functional heterogeneity of the product. The study objective becomes to identify the maximum tolerated dose with high feasibility of being administered. This article describes a new dose-finding method that adaptively accounts for safety and feasibility endpoints in guiding dose allocation. METHODS: We propose an adaptive dose-finding method that integrates accumulating feasibility and safety data to select doses for participant cohorts in early-phase trials examining adoptive cell immunotherapy. We sequentially model the probability of dose-limiting toxicity and the probability of feasibility using independent beta-binomial models. The probability model for toxicity borrows information across all dose levels using isotonic regression, allowing participants infused at a lower dose than his or her planned dose to contribute safety data to the dose-finding algorithm. We applied the proposed methodology in a single simulated trial and evaluated its operating characteristics through extensive simulation studies. RESULTS: In simulations conducted for a phase I study of adoptive immunotherapy for newly diagnosed glioblastoma, the proposed method demonstrates the ability to identify accurately the feasible maximum tolerated doses and to treat participants at and around these doses. Over 10 hypothesized scenarios studied, the percentage of correctly selecting the true feasible and maximum tolerated dose ranged from 50% to 90% with sample sizes averaging between 21 and 24 participants. A comparison to the only known existing method accounting for safety and feasibility yields competitive performance. CONCLUSION: We have developed a new practical adaptive dose-finding method to assess feasibility in early-phase adoptive cell therapy trials. A design that incorporates feasibility, as a function of the quantity and quality of the product manufactured, in addition to safety will have an impact on the recommended phase II doses in studies that evaluate patient outcomes.
BACKGROUND/AIMS: Dose feasibility is a challenge that may arise in the development of adoptive T cell therapies for cancer. In early-phase clinical trials, dose is quantified either by a fixed or per unit body weight number of cells infused. It may not be feasible, however, to administer a patient's assigned dose due to an insufficient number of cells harvested or functional heterogeneity of the product. The study objective becomes to identify the maximum tolerated dose with high feasibility of being administered. This article describes a new dose-finding method that adaptively accounts for safety and feasibility endpoints in guiding dose allocation. METHODS: We propose an adaptive dose-finding method that integrates accumulating feasibility and safety data to select doses for participant cohorts in early-phase trials examining adoptive cell immunotherapy. We sequentially model the probability of dose-limiting toxicity and the probability of feasibility using independent beta-binomial models. The probability model for toxicity borrows information across all dose levels using isotonic regression, allowing participants infused at a lower dose than his or her planned dose to contribute safety data to the dose-finding algorithm. We applied the proposed methodology in a single simulated trial and evaluated its operating characteristics through extensive simulation studies. RESULTS: In simulations conducted for a phase I study of adoptive immunotherapy for newly diagnosed glioblastoma, the proposed method demonstrates the ability to identify accurately the feasible maximum tolerated doses and to treat participants at and around these doses. Over 10 hypothesized scenarios studied, the percentage of correctly selecting the true feasible and maximum tolerated dose ranged from 50% to 90% with sample sizes averaging between 21 and 24 participants. A comparison to the only known existing method accounting for safety and feasibility yields competitive performance. CONCLUSION: We have developed a new practical adaptive dose-finding method to assess feasibility in early-phase adoptive cell therapy trials. A design that incorporates feasibility, as a function of the quantity and quality of the product manufactured, in addition to safety will have an impact on the recommended phase II doses in studies that evaluate patient outcomes.
Authors: Lawrence G Lum; Archana Thakur; Zaid Al-Kadhimi; Gerald A Colvin; Francis J Cummings; Robert D Legare; Don S Dizon; Nicola Kouttab; Abby Maizel; William Colaiace; Qin Liu; Ritesh Rathore Journal: Clin Cancer Res Date: 2015-02-16 Impact factor: 12.531
Authors: Daniel W Lee; James N Kochenderfer; Maryalice Stetler-Stevenson; Yongzhi K Cui; Cindy Delbrook; Steven A Feldman; Terry J Fry; Rimas Orentas; Marianna Sabatino; Nirali N Shah; Seth M Steinberg; Dave Stroncek; Nick Tschernia; Constance Yuan; Hua Zhang; Ling Zhang; Steven A Rosenberg; Alan S Wayne; Crystal L Mackall Journal: Lancet Date: 2014-10-13 Impact factor: 79.321
Authors: Lawrence G Lum; Archana Thakur; Qin Liu; Abhinav Deol; Zaid Al-Kadhimi; Lois Ayash; Muneer H Abidi; Cassara Pray; Elyse N Tomaszewski; Patricia A Steele; Dana L Schalk; Hiroshi Yano; Alice Mitchell; Melissa Dufresne; Joseph P Uberti; Voravit Ratanatharathorn Journal: Biol Blood Marrow Transplant Date: 2013-03-22 Impact factor: 5.742