The Role of Cytotoxic Chemotherapy in Prostate Cancer
The Role of Cytotoxic Chemotherapy in Prostate Cancer
Approximately 20 years ago, the lead author of this paper co-wrote an extensive paper on the experience with cytotoxic chemotherapy in prostate cancer. Frequently cited in subsequent publications on the subject, this comprehensive review of the literature served as the basis for discussing this treatment modality at the first National Institutes of Health consensus conference on prostate cancer in 1987. While the paper reflected state-of-the-art knowledge about nonhormonal treatment approaches, the natural history of prostate cancer was only beginning to undergo systematic study. Disease-oriented research programs were first being organized, and data on outcomes relied primarily on long-term experience from a few large-scale clinical trials. The importance of clinical trial methodology was just beginning to be emphasized. The selection of endpoints to determine therapeutic efficacy was frequently based on semi-intuitively derived, and often biased, observations that were never properly validated.
The original publication discussed the methodologic challenges involving clinical trials in prostate cancer, and placed the issue in perspective. A unique aspect of that review was its emphasis on survival analyses in phase III trials as the endpoint for conclusively defining efficacy. The paper stressed the lack of evidence to support routine use of any treatment regimen based on survival observations from randomized studies in patients with hormone-resistant prostate cancer (HRPC). Perhaps most critically, in 1987, prostate-specific antigen (PSA) was only a "promising marker" awaiting clinical application.
Twenty Years Later
In the current review, we indicate what has evolved since that report from 20 years ago. We will consider the current status of nonhormonal chemotherapy for patients with hormone-refractory disease and provide a more global view of the clinical paradigms for which chemotherapy may be developed. The current armamentarium of cancer treatment includes a rapidly growing class of compoundsgenerally classified as noncytotoxic treatmentsdesigned to target specific biologic mechanisms that control the growth of cancer at the molecular level. These agents differ substantially from the empiric cytotoxic chemotherapy that represented the only approach available 20 years ago.
The clinical development of noncytotoxic treatment requires a completely different methodology from that employed for the development of conventional cytotoxics. Today, this methodology is the focus of extensive study, not only in prostate cancer, but in other malignancies as well. It is becoming increasingly clear that the development of new treatments requires the identification of clinical paradigms and endpoints that are specific for the treatment modality being tested. For example, the most adequate paradigm for developing bone-targeted approaches may differ substantially from the approach taken with immunotherapies or antimetastatic compounds designed to interfere with the process of metastasis at earlier stages of the disease.
In this review, we will describe how we addressed some of the challenges posed to clinical investigators 20 years ago and what has been accomplished since that time. In addition, we will outline the new challenges facing us.
HRPC in the Pre-PSA Era
Until 1987, chemotherapy trials in prostate cancer were largely limited to patients with evidence of radiologic and/or symptomatic disease progression after conventional androgen suppression. Results of different clinical trials were difficult to compare in view of the variability in eligibility criteria and criteria employed to define therapeutic benefit. The hormone-refractory patient population of the pre-PSA era was commonly characterized by patients with extensive disease involvement, frequent incidence of significant symptoms such as pain, anemia, and reduced functional status, and a history of extensive prior palliative radiation therapy. In such patients, tolerance to conventional cytotoxic treatments is often reduced, frequently requiring attenuation of doses of myelotoxic agents, which constituted the majority of regimens available at that time.
In the extensive 1985 review, we emphasized a number of issues. First, response rates reported in phase II trials with most regimens were generally low. Clearly, in most cases, clinical responses did not represent convincing evidence to support therapeutic efficacy. A common argument was that evidence of tumor shrinkagea commonly relied upon parametercould not be reliably assessed in most patients with metastatic prostate cancer because of the pattern of metastasis in this disease (ie, osteoblastic bone metastasis). Because of the infrequent incidence of bidimensionally measurable disease, various groups employed a "new" category-stable disease. This response category was popularized by a clinical trials group known as the National Prostate Cancer Project (NPCP), which was the most active National Cancer Institute-funded clinical trials group in prostate cancer at that time.
Studies conducted by the NPCP reflected a sequence of relatively small randomized phase II trials with various chemotherapeutic regimens utilized during the 1970s and early 1980s. NPCP investigators defined stable disease as "no evidence of disease progression for 12 weeks." This was based on the observation that patients without progression at 12 weeks from the start of the trial had a longer survival than those showing progression by that time. This highly biased and unvalidated criterion was vigorously disseminated and employed in clinical trials for several years and, extraordinarily, it was employed as a key criterion in the US Food and Drug Administration (FDA) approval of estramustine phosphate (Emcyt), the first chemotherapeutic compound approved for prostate cancer in the United States.
The first two NPCP studies (study 100 and 200, Figures 1 and 2), conducted in two different patient populations based on extent of prior radiation, compared various regimens against a control arm designated as "standard treatment." Patients in the control arm received a variety of palliative therapies, including radiation for pain control, analgesics, and alternative hormonal approaches such as corticosteroids.[3,4] NPCP 100 compared weekly fluorouracil vs cyclophosphamide vs standard treatment, whereas NPCP 200 compared estramustine phosphate vs streptozotocin (streptozocin, Zanosar) vs standard treatment.
The results of these two studies indicated very similar median survivals in all treatment arms, despite the higher "response" rates reported with the chemotherapy regimens, most of which were in the "stable disease" category. In both studies, crossover from standard treatment to chemotherapy arms was permissible. Given the advanced stage of these patients at the time of crossover, and based on the infrequent "responses" reported with treatment in those initially assigned to the chemotherapy arms, it is unlikely that crossover treatment resulted in any meaningful benefit that would affect survival.
These results served as the foundation for testing other regimens without further comparisons to a "no chemotherapy" control arm. Additional regimens evaluated on NPCP studies included dacarbazine, procarbazine (Matulane), prednimustine (Sterecyt), semustine (methyl-CCNU), vincristine, methotrexate, cisplatin, hydroxyurea, and combinations thereof. Besides their historical value, these trials provided no evidence to support the routine use of any of these regimens for the treatment of patients with hormone-refractory prostate cancer.
Table 1 illustrates the results of prospectively randomized studies conducted by the NPCP and other investigators before 1987. Figure 3 represents a composite figure for all prospectively randomized studies containing 20 or more patients per arm. After its original publication in 1985, this figure became known as the "spaghetti curve."