Matthew Schipper, PhD

• PhD, Biostatistics, University of Michigan, 2006
• MS, Western Michigan University, 2001
• BS, Western Michigan University, 2000

Research Interests:
Clinical Trial Design: Optimal design of early phase clinical trials must make efficient use of small patient sample sizes to address study objectives and inform subsequent trials while also protecting the safety of trial subjects. We have proposed new designs for optimally incorporating dose expansion cohorts in early phase trials, an extension of the popular phase I CRM design to accommodate uncertainly in toxicity assessment.  We have also proposed an extension of small n and standard SMART designs to include continuous tailoring variables for second stage randomization. I have also participated in development of guidelines for content of statistical analysis plans for early phase trials. 

Data Integration: Large observational datasets are well suited for development of prognostic models while treatment effect estimates generally come from randomized clinical trials to minimize risk of confounding.  Integrating these two sources of information allows for estimation of treatment effects for individual patients. We have applied these methods in prostate cancer to estimate the survival benefit of hormone therapy (ADT) for individual patients.  The resulting clinical decision aid is implemented in a R Shiny web app. 

Development and Application of Statistical Methods for Personalized Medicine: One of my major research interests is the use of statistical models and data analysis to inform and optimize treatment selection and other clinical decisions for individual patients.  Our group is actively working on development and application of new statistical methods to inform clinical decisions including optimal RT dose in liver cancer, RT heart dose constraints in lung cancer, benefit of ADT in prostate cancer and whether to order a brain CT for patients presenting in the emergency room with traumatic brain injury. 

Research Projects:
Androgen Deprivation Therapy has been shown to result in improved overall survival in multiple randomized trials. At the same time it is associated with significant toxicity and its benefit to individual patients varies with their baseline health and prostate cancer risk factors. We have developed a modeling approach that provides treatment benefit estimates for individual patients to enable informed decision making.

Optimal treatment planning in Radiation Therapy requires tradeoffs between efficacy and toxicity and between multiple potential toxicities. We have developed a Utility based approach that makes these tradeoffs explicit and quantitative.

Use of intermediate clinical endpoints for futility stopping: We and others have shown that biochemical recurrence (BCR) does not meet surrogacy criteria in localized prostate cancer. However it is being used in several ongoing master protocols to stop treatment arms if they fail to show a benefit in BCR. We are performing a meta-analysis to assess correct and incorrect stopping probabilities in this setting.

When patients present in an emergency department with traumatic brain injury (TBI) an important clinical decision is whether to obtain brain computed tomography (CT) imaging. There are costs associated with both false positive (exposure to radiation) and false negative (patients with bleeding in the brain may need surgery to stop the
bleeding and remove blood clots) decisions. We are building models for the probability of a positive brain CT using both biomarker and clinical factors to inform this decision.

Li P, Taylor JMG, Boonstra PS, Lawrence TS, Schipper MJ: Utility based approachin individualized optimal dose selection using machine learning methods. Stat Med: 2022. PM35343595 

Chase EC, Bryant AK, Sun Y, Jackson WC, Spratt DE, Dess RT, Schipper MJ: Development and Validation of a Life Expectancy Calculator for U.S. Prostate Cancer Patients. BJU Int: 2022. PM35373440 

Hartman H, Tamura RN, Schipper MJ, Kidwell KM: Design and analysis considerations for utilizing a mapping function in a small sample, sequential, multiple assignment, randomized trials with continuous outcomes. Stat Med 40(2): 312-326, 2021. PM33111381

McFarlane MR, Hochstedler KA, Laucis AM, Sun Y, Chowdhury A, Matuszak MM, Hayman J, Bergsma D, Boike T, Kestin L, Movsas B, Grills I, Dominello M, Dess RT, Schonewolf C, Spratt DE, Pierce L, Paximadis P, Jolly S, Schipper M, Michigan Radiation Oncology Quality Consortium as part of the Blue Cross Blue Shield of Michigan and Blue Care Network of Michigan Value Partnerships Program.: Predictors of Pneumonitis After Conventionally Fractionated Radiotherapy for Locally Advanced Lung Cancer. Int J Radiat Oncol Biol Phys: 2021. PM34314815 

Schipper MJ, Yuan Y, Taylor JM, Ten Haken R K, Tsien C, Lawrence TS: A Bayesian dose-finding design for outcomes evaluated with uncertainty. Clin Trials: 17407745211001521, 2021. (In Press) PM33884907 

Li P, Taylor JMG, Spratt DE, Karnes RJ, Schipper MJ: Evaluation of predictive model performance of an existing model in the presence of missing data. Stat Med: 2021. (In Press) PM33843085 

Gharzai LA, Jiang R, Wallington D, Jones G, Birer S, Jairath N, Jaworski EM, McFarlane MR, Mahal BA, Nguyen PL, Sandler H, Morgan TM, Reichert ZR, Alumkal JJ, Mehra R, Kishan AU, Fizazi K, Halabi S, Schaeffer EM, Feng FY, Elliott D, Dess RT, Jackson WC, Schipper MJ, Spratt DE: Intermediate clinical endpoints for surrogacy in localised prostate cancer: an aggregate meta-analysis. Lancet Oncol 22(3): 402-410, 2021. PM33662287 

Soni PD, Hartman HE, Dess RT, Abugharib A, Allen SG, Feng FY, Zietman AL, Jagsi R, Schipper MJ*, Spratt DE*. *Contributed equally.: Comparison of Population-Based Observational Studies With Randomized Trials in Oncology. Journal of Clinical Oncology: 2019. May 10;37(14):1209-1216. PMID: 30897037; PMC7186578.

Email: [email protected]
Office: 734-232-1076

Address:
M4015 SPH II
1415 Washington Heights
Ann Arbor, MI 48109