The addition of veliparib–carboplatin to standard chemotherapy appeared likely to improve rate of pathologic complete response among women with triple-negative breast cancer, according to results of I-SPY 2 published in The New England Journal of Medicine.
Researchers calculated an 88% predictive probability that the veliparib (ABT-888, AbbVie) and carboplatin combination would remain effective when added to standard chemotherapy in phase 3 confirmatory trials.
Hope S. Rugo
The I-SPY 2 trial is the adaptive second phase of a clinical trial aimed at assessing safety and efficacy of multiple new agents which are added to standard chemotherapy for patients with high-risk clinical stage II or III breast cancer and a tumor size of 2.5 cm or larger in diameter. The objective of the trial is to reduce the cost, time and number of patients needed to identify effective drugs for locally advanced breast cancer.
“The goal of the trial from a drug-development perspective is to rapidly identify which disease subtypes (or “signatures”), if any, are sufficiently responsive to treatment with a given regimen to enable a small, focused and successful phase 3 trial,” Hope S. Rugo, MD, clinical professor in the department of medicine and director of the breast clinical oncology trials program at Helen Diller Comprehensive Cancer Center of University of California, San Francisco, and colleagues wrote. “Preclinical models have shown that veliparib, an oral, potent inhibitor of poly(ADP-ribose) polymerase, markedly potentiates the antineoplastic effect of carboplatin.”
Researchers categorized patients into eight biomarker subtypes based on HER-2 status, hormone receptor status and risk defined by a 70-gene assay.
The current exploratory group involves evaluation of veliparib–carboplatin among 116 patients with HER-2–negative breast cancer.
All patients received the standard 12 cycles of weekly 80 mg/m2 IV paclitaxel, followed by 60 mg/m2 IV doxorubicin and 600 mg/m2 cyclophosphamide every 2 to 3 weeks for four doses.
In this section of I-SPY 2, researchers used adaptive randomization to assign patients to receive standard therapy alone (control, n = 44; median age, 47.5 years) or with veliparib–carboplatin (n = 72; median age, 48.5 years).
Patients in the veliparib–carboplatin group received oral veliparib (50 mg twice daily for 12 weeks) with carboplatin at various doses (to achieve a pharmacologic area under the concentration-vs.-time curve of 60 mg/hr/liter at weeks 1, 4, 7 and 10) concurrently with paclitaxel.
Pathologic complete response served as the primary endpoint.
Meeting the prespecified Bayesian predictive probability of success (85%) or futility (10%) thresholds for a phase 3 trial triggered the end of enrollment.
Veliparib–carboplatin failed to reach the Bayesian efficacy threshold for the full cohort (53% predictive success probability).
In a subgroup analysis, veliparib–carboplatin failed to reach the threshold for patients with hormone receptor–positive, HER-2–negative cancer (8% predictive success probability). However, among patients classified as having triple-negative disease, veliparib–carboplatin yielded an 88% predicted probability of success in a phase 3 trial. The predicted pathologic complete response rate was higher among patients with triple-negative disease who received veliparib–carboplatin (51%; 95% Bayesian probability interval [PI], 36-66) compared with patients who received the control (26%; 95% Bayesian PI, 9-43).
However, patients in the veliparib–carboplatin group experienced higher rates of grade 3 or worse neutropenia (71% vs. 2%), thrombocytopenia (21% vs. 0%) and anemia (28% vs. 0%).
“An important objective is to reduce the number of patients needed to determine the clinical activity of an agent or regimen,” Rugo and colleagues wrote. “I-SPY 2 represents a new clinical trial model that is designed to facilitate rapid evaluation of new therapeutics and to identify biomarkers for definitive subsequent study. Predicting outcomes in a future trial that has a substantial chance of being successful establishes a high bar for continued development.”
Adaptive phase 2 models, like that used in I-SPY 2, may have a larger role in cancer research, David Harrington, PhD, and Giovanni Parmigiani, PhD, both professors at the T.H. Chan School of Public Health at Harvard University, wrote in an accompanying editorial.
“Oncology has been slow to adopt Bayesian designs even though they are often well suited to settings in which inference and decisions benefit from adaptation based on accruing information,” they wrote. “The fundamental tenets behind the I-SPY 2 platform and model — the multi-group platform, options for graduation and addition of drugs, adaptive randomization, and prediction of success in confirmatory trials — are important first steps toward the efficient use of clinical resources.” – by Nick Andrews
Rugo reports grant and other support from BioMarin, and support from AbbVie outside the submitted work. Please see the full study for a list of all other researchers’ relevant financial disclosures.