I-SPY Trials Seek Effective New Drug Combinations Using Adaptive Trial Design

Evidence-Based Oncology, May/June, Volume 19, Issue SP4

Clinical trials usually test 1 drug at a time by comparing it to a similar existing drug or a placebo, but a new trial is trying to determine whether the responses that women with breast cancer have with certain investigational agents early on in a clinical trial can guide treatment decisions later on within the same trial, as well as expedite further clinical trials involving these new agents.

That is the goal of the I-SPY trials, a series of novel trials designed to test several different experimental therapies at once to determine which ones work best in which patients. Laura J. Esserman, MD, MBA, director of the Carol Franc Buck Breast Care Center and professor of surgery and radiology at the University of California San Francisco, discussed the ISPY 1 and I-SPY 2 trials at the 30th Annual Miami Breast Cancer Conference (MBCC).

Esserman said that women who are at risk for systemic recurrence will not be cured by surgery alone, and the order of surgery and systemic therapy has no impact on survival outcomes. To this end, the I-SPY trials are designed to determine which patients benefit the most when they receive experimental drugs, and among those who perform well, patients with similar disease characteristics can be stratified to receive the same drugs.

“We’re looking for the people who have high proliferative disease, if they’re hormone-positive, and the triple-negative and the HER2-positive patients,” Esserman said. “Those are the ones who are going to get the biggest benefit from chemotherapy, and they’re the ones where the ability to respond to therapy or not is ultimately going to translate into a difference in event-freesurvival.”

In the I-SPY 1 trial, 3 important factors were involved when determining how to structure further trials. First, patients who did well regardless of the chemotherapy they received were considered

to have a good prognosis. Second, patients who were not in the good prognosis group who had a better pathologic complete response to neoadjuvant therapy had longer disease-free survival. Third, imaging, specifically with MRI, to determine the size of the tumor during treatment was also a good predictor of how patients would respond to a particular treatment.

Using this information, researchers designed the I-SPY 2 trial to test several different agents in women with newly diagnosed, locally advanced breast cancer. Specifically, I-SPY 2 is evaluating whether the addition of investigational agents to standard chemotherapy improves outcomes in the neoadjuvant setting.

Unlike many studies that focus on how patients with metastatic disease respond to new and investigational agents, the I-SPY 2 trial is enrolling patients with primarily stage II and stage

III disease. Esserman explained that even if the drugs are effective, it’s possible that if the drugs are tested in the metastatic setting that they will not yield a signal or be of sufficient strength to overcome metastatic disease.

“A perfect example of this is chronic myelogenous leukemia, where the drug Gleevec targets the ABL pathway, and if you give that drug to someone in the accelerated phase of CML, or early disease, you can cure most people now,” Esserman said. “But if you give it in the blast phase or what is the equivalent of metastatic disease, you can’t cure anybody.”

Esserman also explained the importance of testing multiple agents in ISPY 2. “The reason we have more than 1 agent is we don’t know for which subtype the agents are going to work,” Esserman said. “But we let the trial sort that out. We may have an idea of who it’s going to benefit, but we really don’t know. The data have got to sort that out for us.”

Additionally, the design of the trial allows for only 1 control arm despite the myriad of agents being tested in the study. Each new agent is compared with the same control arm. In her presentation at MBCC, Esserman provided a brief summary of agents involved in the trial thus far (Table). A total of 5 agents are currently active in the trial, and certain therapies have already proved more effective in certain subtypes of cancers than others. For example, neratinib, a pan-ErbB inhibitor, is active regardless of HER2 or hormone receptor status, whereas ABT-888 plus carboplatin appears to only be effective in HER2-negative patients, though hormone receptor status does not seem to matter.

The end point for each one of these agents is pathologic complete response. In order to advance, each agent must meet a threshold of having an 85% predicted likelihood of success in a phase III trial that would enroll at least 300 patients. Esserman said that these drugs have to double the log odds of having a pathologic complete response.

Additionally, I-SPY 2 is not designed to compare 2 drugs that target the same pathway. This way, it is known which patients with a particular mutation will specifically benefit from a drug, rather than having to design a series of trials for that 1 drug and find a large group of patients with that particular mutation to enroll.

“It’s in everybody’s interest to figure out how to be more efficient, work together, bring down the cost and the time to get these agents screened,” Esserman said. “One of the things we’re talking about doing now is, if drugs graduate from I-SPY, how can we start working seamlessly to get the drug into a confirmatory trial relatively quickly? That would actually benefit patients the most, and that’s very exciting.”