Biopsy-Derived Biomarkers in Phase I Trials: Building Confidence in Drug Development.
T. Yap,U. Banerji,J. D. de Bono,P. Workman
DOI: https://doi.org/10.1200/JCO.2016.67.0075
IF: 45.3
2016-05-02
Journal of Clinical Oncology
Abstract:TO THE EDITOR: The article by Sweis et al in a recent issue of Journal of Clinical Oncology raises important issues on the use of nondiagnostic biopsies for pharmacodynamic (PD) studies in phase I trials. We respectfully disagree with the authors that the addition of biopsy-derived biomarkers to phase I trials has a low yield in enhancing drug development and that such studies should be limited only to trials that primarily assess the effects of the drug(s) on the biomarker. On the contrary, we believe that PD studies are critical in first-in-human phase I trials of novel antitumor drugs to confirm and quantify target engagement, biochemical pathway modulation, and biologic effects. Ideally, these studies should define the extent and duration of PD effects and the dose-response relationship. Relating such parameters to corresponding preclinical data helps build confidence in interpreting anticipated or observed antitumor responses and/or toxicities. Without such PD studies in phase I trials, we risk returning to the practice of empirical drug development and increasing the likelihood of late-phase trial attrition and ultimately costly failure. For example, biopsy-derived PD biomarkers are important in phase I trials when non–mechanism-based toxicities are observed with no evidence of antitumor activity. In such scenarios, a tumor PD assay demonstrating a lack of the desired mechanistic activity would support a no-go drug development decision and avoid going forward with ineffective drugs or subtherapeutic doses and schedules. Importantly, such scenarios are also unlikely to lead to late-phase trial publications, which is the measure the authors use to assess the impact of biopsy-derived biomarkers, because progression to phase II studies would probably not occur. Ultimately, for drug development decisions, PD data should be considered alongside pharmacokinetic (PK) and efficacy/toxicity findings within the context of the pharmacologic audit trail. A pertinent example to illustrate these points is our phase I trial of the poly (ADP-ribose) polymerase (PARP) inhibitor olaparib (AstraZeneca, London, United Kingdom), a study highlighted by Sweis et al for which a significant biomarker result was cited in subsequent publications. Integrated with PK, efficacy, and toxicity data, the PD results provided investigators with the necessary confidence to proceed to phase II trials. In addition, such PD data remain informative in daily clinical practice by providing the range of active doses to use in patients who are poorly tolerant of olaparib. In contrast, no biopsy-derived PD studies were undertaken in early-phase trials for iniparib (BiPar Sciences, Brisbane, CA), which was developed as a purported PARP inhibitor. Despite promising phase II trial data, the large phase III study of iniparib with chemotherapy in patients with triple-negative breast cancer was negative. Critically, these results led to concerns that PARP inhibition may not be an effective therapeutic strategy and negatively impacted the development programs of bona fide PARP inhibitors. However, iniparib was eventually found to be a poor inhibitor of PARPactivity, which should have been determined much earlier. We agree with the authors that it is critical that any tumor PD biomarkers used for decision making should undergo rigorous scientific and technical validation, and all key assay parameters should be clearly presented or cited in peer-reviewed publications. To facilitate this, it is important that both drug development and biomarker validation processes occur in parallel so as to produce a robust and reproducible biomarker test that is progressively validated and fit for purpose. The authors raise issues of safety concerns and patient acceptability of tumor biopsies, but published data do not support these as major challenges. Despite the recent exponential growth in immuno-oncology therapeutic strategies, our understanding of the complex mechanisms of these novel agents remains poor. The authors noted that none of the immune checkpoint inhibitor trials identified included tumor biopsy–derived biomarkers, highlighting the dearth of such PD studies, which could prove useful to potentially explain differences in antitumor responses observed between patients. Ultimately, the importance of biopsy-derived PD biomarkers in phase I trials lies largely in the thorough pharmacologic characterization of novel drugs to build confidence in the relationship between PK, PD, and effect. This was evident in the early clinical development of the US Food and Drug Administration–approved BRAF inhibitor vemurafenib (Genentech, South San Francisco, CA), where PD studies confirmed that near-complete inhibition of ERK activation, which is required for significant tumor response, was achieved at the recommended phase II dose. Although only a minority of phase I trials between 2003 and 2010 included posttreatment tumor biopsies, we are encouraged that such studies increased steadily over this period. We are also optimistic that with improvements in modern technologies, this rise in biomarker use will continue to improve decision making in drug development and help drive improvements in cancer care.