While biologically based therapeutic cancer vaccines and other cancer immunotherapies have great promise in treating certain cancers, getting to “yes” for FDA approval is likely to remain challenging. And, as knowledge and clinical trial experience about the nature of effective antitumor immune response to immunotherapeutics has emerged, it has become clear that valid endpoints and trial designs for conventional chemotherapeutic drugs need rethinking with respect to immunotherapy.
Despite enthusiastic market potential forecasts, with some predicting that the commercial market for cancer vaccines could increase to over $7 billion by 2015, only one therapeutic cancer vaccine has been approved to date in the U.S. In 2010, Dendreon’s Provenge, an autologous prostate cancer vaccine, became the first such therapeutic approved by the FDA.
The FDA had declined to approve it in 2007, against the recommendations of its own advisory committee, and required that the company undertake another clinical trial. In the first trial, of 127 patients, 115 patients had progressive disease at the time of data analysis, and all patients were followed for survival for 36 months. The median for time to disease progression (TTP) for sipuleucel-T was 11.7 weeks compared with 10.0 weeks for treatment remained a strong independent predictor of overall survival; the median ratio of T-cell stimulation at 8 weeks to pretreatment was eightfold higher in Provenge-treated patients (16.9 v 1.99; P < .001). Provenge therapy was well tolerated.
While the improvement in the primary end point TTP did not achieve statistical significance, this study suggests that Provenge might have provided a survival advantage to asymptomatic hormone-refractory prostate cancer (HRPC) patients.
A second trial resulted in approval of the dendritic cell-based vaccine for the treatment of metastatic, asymptomatic HRPC. Results of the pivotal IMPACT clinical trial had shown that Provenge extended median survival by 4.1 months to 25.8 months from 21.7 months, sufficient for FDA approval in 2010.
Provenge’s approval pathway has been characterized by some as having “all the features of a heavyweight healthcare fight—desperate patients demanding access to a promising therapy with a very expensive drug that extends life only a few months and efficacy data open to interpretation.”
Michael D. Becker, founder and senior partner of MD Becker Partners, says the FDA does a “good job” with cancer immunotherapeutics. “The approval of Provenge was a watershed event,” he said, and required a lot of time and discussion on the agency’s part.
“I think the FDA in general has been very consistent in their mandate requiring two Phase III randomized clinical trials.” To Becker’s point, lymphoma autologous vaccine developer Biovest International announced that it had conducted a formal clinical guidance meeting with the FDA to determine the most expeditious U.S. registration pathway for BiovaxID last June.
At that meeting, the FDA required that Biovest conduct a second Phase III trial to confirm the clinical data generated in a successful Phase III trial (BV301). Results from that eight-year pivotal, randomized, multi-center, double-blind, controlled trial showed that its BiovaxID autologous cancer vaccine prolonged disease-free survival in follicular non-Hodgkin’s lymphoma, and with little or no toxicity.
Clinical Trial Design
A key challenge of clinical trial design for cancer immunotherapeutics is that experience has shown efficacy of therapeutic cancer vaccines “should not be measured” by the standards used for chemotherapeutic agents. Chemotherapy, which directly counteracts tumor growth, can be evaluated on tumor response, or the basis of tumor size. But cancer immunotherapy, experts say, relies on an immune response that may develop only over the course of time.
In some cases, tumors may actually increase in size in the short term. In other cases, the immune response results in prolonged “stable disease,” making overall survival rate a better endpoint than tumor size.
One practical result of this is that patient survival may not be affected until some months after treatment start. For the survival endpoint, Kaplan-Meier curves from randomized immunotherapy trials may show a delayed separation after months, which directly influences the statistical power to determine treatment effects observed over the entire length of the curves.
As a remedy, Hoos and Britten have suggested immunotherapy response criteria derived from RECIST and WHO—immune-related response criteria—be developed. These criteria would capture all clinical activity patterns for a reliable assessment of activity signals in early trials.
Alternative statistical models tailored to address the delayed separation of curves have to consider that all events prior to the separation do not contribute to the differentiation between study arms after the separation, thus leading to loss of statistical power. Such new statistical models need to compensate for this loss of power. They may split the hazard ratio into an early and a late component before and after the separation of curves.
Importantly, the loss of power and absence of early effects should be carefully considered when designing randomized trials with early interim and futility analyses.
This phenomenon has been observed in multiple randomized immunotherapy trials and is often not apparent until 4–8 months or more after random assignment, as in the initial Provenge trial.
Yervoy, an anti-CD4 monoclonal antibody approved in 2011 for treatment of metastatic melanoma, provides an object lesson in the development of novel immunotherapeutics.
The antibody produces durable objective responses and/or stable disease in patients with metastatic melanoma. As more patients were treated with the drug during the course of clinical trials it became clear that “the kinetics of responses are heterogeneous and significantly different from those of chemotherapy and other immunotherapy.”
Though objective response or stable disease were observed within conventional time frames, responses were observed weeks to months after therapy initiation. Response or stable disease may be preceded by apparent early disease progression, or may occur simultaneously with different progressing lesions within the same patient (a “mixed” response).
According to Jedd Wolchok, director of immunotherapy clinical trials at Memorial Sloan-Kettering Cancer Center, “There are some people who we have treated with ipilimumab whose scans look just as abnormal now as they did five years ago, so it has turned it into a chronic disease.
“It changed the situation from something they were dying from into something they are living with. That really does show you that the immune system can restore an equilibrium between the person and the tumor.”
As the NCI announced that the first clinical trials will be launched under its Cancer Immunotherapy Trials Network (CITN) including 27 U.S. cancer centers and universities, Dr. Wolchuk, a CITN investigator, noted that the impact of the Provenge and Yervoy approvals was enormous. “For pharmaceutical companies and the industry to see a successful immunotherapy that can be administered in a doctor’s office has really changed the field.”
While multiple complex challenges exist in designing cancer immunotherapy trials, cancer vaccine developers, researchers note, are well advised to use a homogenous patient population, focus on containing minimum residual disease where vaccines seemed most effective, not use endpoints related to tumor shrinkage, and provide sufficient time for a patient response as there are “no perfect surrogate markers” for efficacy. And, they noted, it takes time to see results.
As more successes accrue, investors may lose some of their insecurities related to cancer immunotherapeutics. “In general, institutional investors remain on the sidelines with regard to interest in immunotherapy companies; they are waiting for additional positive outcomes from clinical trials and commercial success with autologous approaches given the Provenge experience,” Becker aptly remarked.