Meta-Analysis Shows Biomarker-Based Trials Have Better Outcomes

Maria Schwaederle, PharmD

Clinical trials studying a biomarker-based treatment showed substantially better results than non-personalized treatment methods in a meta-analysis, presented in advance of the 2016 ASCO Annual Meeting, that examined 13,203 patients from phase I trials.

In the analysis, there was a response rate of 31% in the biomarker driven arm compared with just 5% in the non-personalized group (P<.0001). The median progression-free survival (PFS) in the biomarker arm was 5.7 months compared with 3 months without biomarker selection (P= .0002). Findings from the study could reshape the intent behind early phase clinical trials, according to the researchers.

"In the past, it was heard that phase I trials were not about efficacy and their only purpose was to understand side effects," lead investigator Maria Schwaederle, PharmD, Center for Personalized Cancer Therapy and Division of Hematology and Oncology, University of California, San Diego, said during an ASCO press briefing. "Our analysis really showed that this belief is completely outdated, and with biomarker selection, and especially genomic biomarkers, we can reach high response rates even in phase I trials."

For the analysis, findings from 351 single-agent phase I study arms were selected from the PubMed index from between January 2011 and December 2013. For the purpose of the study, personalized therapy was defined as a treatment that was selected using a biomarker test or a trial conducted in a population in which at least 50% of individuals harbored a known biomarker.

The analysis focused exclusively on response rates and PFS. Overall survival was not explored, since this information was rarely reported from these early phase studies. Response rates were available for all 13,203 patients from 58 biomarker-driven trials and 293 non-personalized trials. PFS data were only available for 1700 patients from 7 biomarker trials and 38 standard trials.

Targeted therapies that were utilized in the absence of a biomarker test demonstrated low response rates that were similar to clinical trials exploring cytotoxic agents. In the non-biomarker driven targeted therapy trials (n = 177), the response rate was 5.1% versus 31.1% in those treated with a biomarker-based targeted therapy (n = 57).

"Targeted drugs in and of themselves were not effective. They absolutely need to be given to the right patients," said Schwaederle. "A biomarker-based approach was the most significant independent predictor of improved outcomes in phase I studies."

Certain types of biomarkers were associated with even higher response rates for patients treated using personalized approaches. Those selected for therapy using a genomic biomarker had a response rate of 42% versus 22.4% in those selected using protein expression (P= .0001).

"Both genomic and protein biomarkers led to improved outcomes, but genomic biomarkers performed better," said Schwaederle. "Indeed, the response rate in this heavily refractory population was over 40% with the use of a genomic biomarker."

In recent years, there has been an increasing trend toward accelerated FDA approvals based upon early findings from phase I studies. An early example of this success was characterized in the approval of crizotinib, which was based on early phase studies for patients withALK-rearranged non—small cell lung cancer (NSCLC). One of these instrumental studies was included in the analysis, according to Schwaederle.

In this phase I study, 119 patients withALK-rearranged NSCLC received crizotinib. The response rate was 61%, which included 2 complete responses. The median duration of response was 48.1 weeks.

Other examples of success seen with targeted therapies included BRAF inhibition for patients withBRAF-mutant metastatic melanoma and imatinib for BCR-ABL chronic myelogenous leukemia. For the timeframe of the study, studies exploring BRAF inhibitors were explored as well as second-generation BCR-ABL inhibitors, such as dasatinib, Schwaederle explained. All of these agents have gone on to gain FDA approval.

&ldquo;Precision medicine is not the future of cancer care, it is the present," said press briefing moderator Don S. Dizon, MD, ASCO spokesperson. "This study reinforces that the more we personalize treatment to the patient and the tumor, the better the outcomes — even in the earliest phases of research.&rdquo;

Schwaederle MC, Zhao MM, Lee JJ, et al. Impact of precision medicine in refractory malignancies: a meta-analysis of 13,203 patients in phase I clinical trials. J Clin Oncol. 2016 (suppl; abstr 11520).