Background: “Precision medicine” is the approach to drug design and treatment that is intended to ensure that the right treatment is delivered to the right patient at the right time. Despite the promise of this concept, the science has yet to live up to expectations. In prior work, we evaluated the research portfolio for the excision repair cross-complimentation group 1 (ERCC1) as a predictive biomarker for precision medicine in lung cancer (see appendix B). Despite more than a decade of research and dozens of studies, the scientific community is still uncertain about whether or not this is a useful biomarker in clinical practice. In mapping the ERCC1 portfolio, we identified some plausible explanations for this lack of progress: We found considerable waste from uninterpretable study reports, little replication and follow-up of positive findings, and no evidence of systematic validation of the diagnostic technologies. Epidermal growth factor receptor (EGFR) is an even more-extensively tested mechanism and pathway in cancer therapy. In contrast to ERCC1, EGFR is officially recognized by the FDA as a useful pharmacogenomic biomarker (with 5 on-label indications for approved drugs). The research portfolio for EGFR may therefore provide an instructive model for how the many uncertainties surrounding a particular cancer biomarker should be studied and resolved across many of its oncology applications.
Aim: To document the entire research portfolio of EGFR testing, providing a dynamic model of biomarker development in oncology and a new modality for evaluating the efficiency of precision medicine research. We hypothesize that despite FDA recommendations, the aggregated research activities supporting EGFR diagnostics may still demonstrate considerable inefficiency—characterized by little effort directed at replication or systematic exploration and validation of assay parameters. If so, then there may still not be adequate evidence to support effective and reliable use of EGFR technologies in the clinic.
Approach: We will conduct a systematic literature search of Medline, Embase, and Clinicaltrials.gov for clinical trials and retrospective biomarker studies that examine EGFR as a predictive biomarker for cancer chemotherapy. The resulting reports and registration records will be extracted for demographics, clinical characteristics, study design properties, biological assay materials and procedures, and outcome data. We will then perform a descriptive analysis of research and reporting activities, statistical meta-analysis of outcomes and data trends, and AERO graphing to chart diagnostic validation efforts and the evolving estimates for EGFR’s predictive value across different chemotherapies and cancer types.
A pilot search for this protocol identified 18 EGFR-targeting agents and 2,577 registered trials of these in Clinicaltrials.gov. However, in our prior work on pharmacogenomic biomarker evidence development, the number of studies of molecularly-targetted agents that include a biomarker assessment has been a small fraction (approximately 7%) of the total number of registered trials of the agent. Therefore, we estimate that the number of studies for data extraction in this project module will be closer to 250.