Adaptive Randomization Pairs Two Treatments With Two Breast Cancer Subtypes

July 7, 2016

Results of two studies taken from the I-SPY 2 trial showed that the use of adaptive randomization in a phase II trial was able to successfully identify agents that would be most effective at treating certain molecular subtypes of breast cancer

Results of two studies taken from the I-SPY 2 trial showed that the use of adaptive randomization in a phase II trial was able to successfully identify agents that would be most effective at treating certain molecular subtypes of breast cancer.

According to the studies, published in the New England Journal of Medicine, “The adaptive randomization algorithm uses the molecular characteristics of the cancers and incorporates accumulated outcome data to efficiently identify the biomarker signatures of tumor subtypes-combinations of molecular subtypes-in which specific agents are most effective.”

In this case, I-SPY 2 identified neratinib for HER-2 positive, hormone receptor–negative disease and veliparib with carboplatin for triple-negative breast cancer as having met prespecified criteria for testing in large phase III trials.

“As more new targets and drugs are discovered, traditional statistical designs, at best cumbersome and inefficient today, will be wholly insufficient for matching patients with effective drugs,” David Harrington, PhD, of Dana-Farber Cancer Institute, and Giovanni Parmigiani, PhD, of Harvard T.H. Chan School of Public Health, wrote in an perspective that accompanied these articles. “We applaud the use of I-SPY 2 described here and urge continued innovation in trial design, especially in both earlier phase I and later phase III settings.”

In the first study, by John W. Park, MD, of the University of California San Francisco, and colleagues, treatment with neratinib, an inhibitor of ErbB and the HER kinase family, was highly likely to produce higher rates of pathologic complete response in patients with HER2-positive, hormone-receptor–negative disease compared with standard chemotherapy with trastuzumab.

The researchers used adaptive randomization to compare standard chemotherapy with or without neratinib. Women in the trial had stage II or III disease and were categorized according to eight biomarker subtypes based on HER2 status, hormone-receptor status, and risk according to a 70-gene profile.

Neratinib reached the prespecified efficacy threshold in patients with HER2-positive, hormone receptor–negative signature. These patients had a mean estimated rate of pathologic complete response of 56% when treated with neratinib compared with 33% in the control patients. According to the researchers, the likelihood that neratinib was superior to standard chemotherapy was 95% and the final predictive probability of success in phase III testing was 79%.

“Although neratinib reached the prespecified threshold of efficacy in this trial only with regard to patients with HER2-positive, hormone receptor–negative cancer, there was some evidence of superior activity over control with regard to several other biomarker signatures,” Park and colleagues wrote. These groups included participants with HER2-positive, hormone receptor–positive disease and HER2-positive disease, regardless of hormone receptor status.

In the second study, by Hope S. Rugo, MD, of the University of California, San Francisco, and colleagues, the addition of the PARP inhibitor veliparib plus carboplatin to standard chemotherapy resulted in higher rates of pathologic complete response in women with triple-negative disease. They randomly assigned 72 patients to veliparib/carboplatin and 44 patients to standard therapy.

After the end of treatment, the estimated pathologic complete response rate was 51% in women with triple-negative disease assigned to veliparib/carboplatin compared with 26% for control patients. There was a 99% predicted probability that this drug combination was superior to standard chemotherapy and an 88% predicted probability of success in a phase III trial.

In their perspective, Harrington and Parmigiani suggested that the value of the data from the I-SPY 2 trials may extend past the results from these two trials.

“Adaptive multigroup trials such as I-SPY 2 have the potential to answer several questions simultaneously and more efficiently than traditionally designed trials,” they wrote. “Which of several promising therapies appear best suited for larger, confirmatory trials? Which patients should be asked to participate in those trials? Is the chance of success in subsequent larger trials sufficient to justify the expense and time needed? Therapies designed to target molecular subtypes of cancer may increase the chances of good responses and, equally important, may be useful in allowing patients to avoid treatments when meaningful benefit is unlikely.”