Hundreds of clinical trials in oncology already use biomarkers to identify patients who have a higher or lower risk of disease progression, as well as help predict how patients will respond to different treatments. However, there has been no systematic overview on the landscape of biomarker use in oncology trials.
In this week’s Science Translational Medicine Robert Sikorski and Bin Yao present the results of their laudable and laborious task to analyse the public database ClinicalTrials.gov for this kind of information.
Their findings can be summarized as follows:
(1) More biomarker work is done in less frequent tumors (such as leukemias) than in more frequent types such as prostate cancer, so it seems that not the big cancer indications will be the first to become segmented into smaller populations with etter treatment options.
(2) There are relatively few selection biomarkers for the major solid tumor indications that will enter clinical practice through current Phase III trials. However, as the database does not include retrospective analyses of completed trials, this point is difficult to assess.
(3) Implementation of biomarkers in clinical trials adds a substantial layer of complexity, increasing costs and making intertrial comparisons more difficult.
(4) Next-generation approaches that apply whole-cancer genome analysis to identify changes associated with therapeutic response will increasingly serve as a major disruptive force reshaping the cancer biomarker landscape.
They forecast that the future of oncology trials will be the study of biomarker-defined patients in smaller, randomized Phase III trials. In addition, they conclude that it should be feasible soon to obtain the complete profile of DNA alterations, DNA copy number changes, and even DNA methylation patterns within a tumor for all subjects in phase I and II studies.
“The resulting ability,” the authors write, “to target new treatments by tumor molecular signatures early in drug development is transformative and offers the promise of demonstrating significantly greater clinical utility with smaller study populations.”