Multidisciplinary analysis of evolution based Abiraterone treatment for metastatic castrate resistant prostate cancer
Listed in
- Evaluated articles (eLife)
- Cancer Biology (eLife)
Abstract
Background
We present a multidisciplinary approach to clinical trial design and analysis in a pilot study ( NCT02415621 ) in which evolution-based mathematical models guide patient-specific dosing for Abiraterone treatment in men with castrate resistant metastatic prostate cancer.
Methods
Abiraterone plus prednisone were administered intermittently based on an evolutionary mathematical model. Outcomes are compared to historical controls and a matched contemporaneous cohort who met trial eligibility but received SOC dosing. Longitudinal cohort data allowed modification of pre-trial model parameter estimates. Model simulations of each patient using updated parameters critically evaluated trial design.
Results
Trial patients, on average, received no abiraterone during 59% of time on treatment. Median Time to Radiographic Progression (TTP) was 30.4 months compared to 14.3 months in the contemporaneous SOC group (p<0.001). All patients in the SOC group have progressed but 4 in the adaptive cohort remain on treatment at >1800 days. Longitudinal trial data found the competition coefficient ratio ( α RS/ α SR) of sensitive and resistant populations, a critical factor in intratumoral evolution, was 2 to 3-fold higher than pre-trial estimates. Computer simulations using the corrected parameter unexpectedly demonstrated optimal cycling can reduce the resistant cells. Longitudinal data from 4 trial patients who remain on treatment are consistent with model predictions. Modeling results predict protocol changes that will allow similar outcomes in most patients.
Conclusions
Administration of abiraterone using evolution-based mathematical models decreased drug dosing and increased radiographic TTP. Integration of mathematical models into trial design identifies novel insights into key treatment parameters and provides optimization strategies for follow-up investigations.
Article Summary
supplemental sections outline the methodology for parameter estimates from trial data, computer simulations, and comparison of simulation results and actual clinical data in every patient in both cohorts.
Statement of Translational Relevance
Integration of evolution-based mathematical models significantly increased TTP in abiraterone therapy for mCRPC. This multidisciplinary approach represents a novel clinical trial strategy in which the treatment protocol is framed mathematically, clinical data then refines model parameterization, and simulations using the updated model predict alternative strategies to improve outcomes. Here we demonstrate the mathematical models used to design the trial can also produce novel analytic approaches. By using longitudinal trial data, key model parameters can be refined. Simulations using the updated model can then be applied to every patient in the trial. Finally, additional simulations demonstrate alternative protocols that could improve results. These analyses demonstrate evolution-based approaches may allow consistent long-term control in patients with metastatic prostate cancer.
Article activity feed
-
Evaluation Summary:
Zhang et al use evolution-guided mathematical models to guide the timing and dosing of arbiterone treatment in castrate-resistant prostate cancer. While the sample size is limited, the implications of the study outcome are broad and compelling, and the paper importantly highlights the transformative potential of deeply interdisciplinary research.
(This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)
-
Reviewer #1 (Public Review):
The authors presented updated results for a clinical trial described in a previous publication (Zhang J et al 2017). With the updated results, the authors were able to further support the validity of their evolution-based model proposed before. These datasets also allow the authors to fit individual-level evolution models and examine critical parameters in their models.
The concept of adaptive therapy is critical and has previously attracted broad attention in the field. The earlier work (Zhang J et al 2017) showed promising results in improving prognosis in prostate cancer patients. In this paper, the follow-up data for this clinical trial clearly confirms its previous findings that adaptive therapy was able to improve TTP and OS.
The authors also went on to infer an evolution model of treatment sensitive …Reviewer #1 (Public Review):
The authors presented updated results for a clinical trial described in a previous publication (Zhang J et al 2017). With the updated results, the authors were able to further support the validity of their evolution-based model proposed before. These datasets also allow the authors to fit individual-level evolution models and examine critical parameters in their models.
The concept of adaptive therapy is critical and has previously attracted broad attention in the field. The earlier work (Zhang J et al 2017) showed promising results in improving prognosis in prostate cancer patients. In this paper, the follow-up data for this clinical trial clearly confirms its previous findings that adaptive therapy was able to improve TTP and OS.
The authors also went on to infer an evolution model of treatment sensitive and resistant cells for each individual patient. With a small number of parameters, the authors can fit most patients' longitudinal data tightly. The authors found some parameters are important to determine the outcome of adaptive therapy. These results are interesting and could have clinical implications, but some model assumptions are strong (like assuming a shared competition coefficient across patients) and some claims need more explicit analysis.
One particularly interesting result from the modeling analysis is that failure of adaptive therapy is caused by overtreatment. However, the readers need to keep in mind that this conclusion is under the simple model described in the paper. More complicated clone composition, interaction and evolution paths will affect this conclusion. -
Reviewer #2 (Public Review):
In this study, Zhang et al. expand on their previous work on using mathematical modelling to guide the timing and dosing of arbiterone treatment in castrate-resistant prostate cancer. The study presents the results of a follow-up pilot trial with 33 patients and adapts an updated mathematical model to fit longitudinal patient data. While the sample size is limited, the implications of the study outcome are broad and compelling. The manuscript can be strengthened by showing that there are no statistically significant differences between the two treatment groups in terms of additional clinical features, such as prior therapies.
-
