Developing Immunotherapy As Legitimate Therapy for Patients With Prostate Cancer

Abstract

Although significant progress has been made in the immunotherapy of prostate cancer, the challenges involved in the successful development of immunotherapy for prostate cancer are many. First, from an immunologic perspective, prostate cancer is almost certainly not one disease, and the progression from one disease state to another and development of castration-resistant prostate cancer (CRPC) are likely associated with antigenic drift. Thus, the development of seemingly efficacious therapy in patients with advanced metastatic CRPC may or may not be relevant to the treatment of patients with earlier stage or androgen deprivation–responsive (hormone-sensitive) prostate cancer. Second, under the best of circumstances, defining clinical responses is challenging in prostate cancer, with considerable debate and confusion over how to use prostate-specific antigen (PSA) kinetics and, in the case of immunotherapy, the apparent disconnect between the intermediate outcomes such as response proportion and time to progression (TTP) traditionally used in oncology and survival. Third, the difficulty in demonstrating a clear relationship between the induction of antigen-specific immune responses and clinical outcome has fueled skepticism regarding immunotherapies. Finally, there is inherent tension between the imperative to have a rapid and realistic drug development timeline (hence the instinct to test new agents in patients with high disease burden and a shorter time to events such as progression or death) and the selection of patients with a lower likelihood of immunosuppressive mechanisms (hence the instinct to test new agents in patients with minimal disease burden, but in whom time to clinically meaningful events such as progression or death can be prohibitively long). In this issue of Journal of Clinical Oncology, Kantoff et al report the results of a randomized phase II trial of a poxviral-based vaccine approach targeting PSA in men with metastatic CRPC. One hundred forty-two men were randomly assigned in a ratio of two to one to receive either PROSTVAC or control vectors. The primary end point, progression-free survival, was not met. However, and somewhat unexpectedly, an overall survival advantage favoring the investigational agent was observed. This report in many ways highlights some of the challenges that confront investigators in seeking to develop immunotherapy in prostate cancer. Prostate cancer is not one disease. Like most immunotherapeutic studies undertaken to date, the present study clearly defined the target population, an important element of any clinical investigation, but in so doing, it also limited the ability to generalize these results. In particular, the group of patients with CRPC was intentionally limited to patients believed to have favorable risk features. Thus, patients with visceral metastases or pain requiring opioid analgesics were excluded, and only patients with Eastern Cooperative Oncology Group performance status of 0 or 1 and Gleason score (GS) of less than or equal to 7 were included. These are reasonable eligibility criteria, because each one of these features is associated with worse survival in a multivariable predictive model developed by Halabi et al on the basis of 1,100 patient with CRPC. Furthermore, an earlier study of sipuleucel-T (Dendreon, Seattle, WA) suggested that only patients with low GSs benefited from such therapy, although this preliminary observation subsequently was shown to be incorrect, with all GS subsets benefiting. However, the use of these eligibility criteria raises important questions. For example, should the development of PROSTVAC proceed only in patients with GSs of 7 or less? Is there biologic plausibility to segmenting the CRPC population in this fashion? With regard to patient selection and balance between the two treatment groups, the survival predicted by Halabi et al in such patients was 22.5 months for those treated with PROSTVAC and 20.4 months for controls. Although the difference in median survival predicted by Halabi et al was not statistically significant, this difference nevertheless favored the experimental treatment group. Furthermore, although the small sample size precludes definitive conclusions, there was also an imbalance in the percentage of lymph node–only patients (a group with favorable prognosis) that favored the PROSTVAC arm (9.8% v 0%) and that conceivably could have had an impact on outcome. PSA kinetics and the apparent disconnect between response proportion or TTP and survival. Kantoff et al indicate that PSA responses were infrequent, although they describe a single case of a patient with a drop in PSA and prostatic acid phosphatase of more than 80%. It is not clear if this was the only case of a PSA decline in the 84 patients assigned to receive PROSTVAC, but the lack of PSA declines has previously been described in immunotherapy trials with some agents (eg, sipuleucel-T, GVAX). In an academic community that has grown accustomed to screening for anti–prostate cancer activity with PSA declines, it will take a concerted effort to move away from PSA end points to define activity. The fact that some apparently active immunotherapeutic agents, such as ipilimumab, do have significant PSA-lowering effects makes this adjustment all the more difficult. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 28 NUMBER 7 MARCH 1 2010