INTRODUCTION — For men with newly diagnosed metastatic prostate cancer, androgen deprivation is highly effective, providing disease control in over 80 percent for a median duration of 18 to 24 months. Although second line hormone therapy (eg, antiandrogen withdrawal, glucocorticoid therapy) may temporarily improve symptoms and diminish disease burden in some men, the vast majority eventually fail treatment with disease that is resistant or refractory to further hormone manipulation. The median survival of men with hormone-independent or hormone-resistant prostate cancer (HRPC) is approximately 12 months, and has not been prolonged by any therapy, until recently

Chemotherapy was previously considered to be relatively ineffective in HRPC. In early trials (many of which included disease stabilization or lack of progression in the definition of objective response, objective response rates were 10 to 20 percent, and median survival did not exceed 12 months. However, newer regimens, particularly those that include docetaxel, are associated with higher rates of both objective and biochemical (prostate specific antigen [PSA]) response, and importantly, median survival durations that approach two years.

The use of cytotoxic chemotherapy for the treatment of HRPC will be reviewed here. The discussion will focus on mitoxantrone and related compounds, the taxanes, estramustine, and the vinca alkaloids. While other drugs, such as antimetabolites, cyclophosphamide, platinum analogs, and topoisomerase I inhibitors, have some activity in HRPC as single agents, they will only be discussed here in the context of combination regimens. Second-line hormonal options, specific issues regarding management of bone metastases, and novel treatment strategies for advanced disease are discussed elsewhere.

ASSESSING RESPONSE TO THERAPY — The vast majority of men with advanced HRPC have bone metastases (over 97 percent in CALGB trial 9583 ). These lesions are poorly amenable to standard oncologic response criteria, which are geared toward measurement of soft tissue masses. As a result, surrogate endpoints (ie, decline in serum PSA, improvement in bone pain or analgesic consumption, quality of life [QOL]) have been increasingly relied upon as indicators of clinical benefit in therapeutic trials.

Serum PSA — For men undergoing chemotherapy for HRPC, changes in serum PSA correlate with objective disease progression, treatment response, and survival. In one multivariate analysis that included trials of various therapies in men with HRPC, median survival was twice as long for men with a posttherapy decline in serum PSA of 50 percent at 8 weeks compared to those with a lesser or no decline (23.6 versus 12.5 months). Moreover, a preliminary report of a prospective analysis of PSA data from Southwest oncology Group (SWOG) trial 9916 supports using the velocity of the PSA decline in the first three months of therapy as a surrogate endpoint for mortality.

However, others have shown that the beneficial impact of modern taxane-containing chemotherapy on survival is only partially explained by the PSA response Thus, PSA response cannot replace overall survival as a primary endpoint in phase III trials to confirm overall clinical benefit.

In 1999, a Prostate Specific Antigen Working Group was convened by the National Cancer Institute to propose eligibility and response guidelines for clinical studies in HRPC . They recommended the following as valid treatment endpoints for trials conducted in men with HRPC:

   Objective response (partial and complete), if soft tissue measurable disease is present;
   Survival;
   A minimum PSA decline of at least 50 percent, occurring in the absence of clinical or radiographic evidence of disease progression, and confirmed by a second PSA value four or more weeks later.

In this review, a 50 percent decline from baseline PSA will be used synonymously with "PSA response."

One caveat to using a decline in serum PSA as an indicator of clinical response is that it cannot be used in those 5 to 10 percent of cancers that are associated with low PSA levels. Some prostate cancers do not make PSA, or make only negligible amounts. Some of these are neuroendocrine tumors, that may respond to cisplatin-based chemotherapy rather than initial hormonal manipulation, while others lack neuroendocrine features, but are associated with a PSA level <10 ng/mL, even when metastatic. Objective responses in such patients are difficult to measure but the duration of their response to androgen ablation and median survival tend to be shorter than that of men with PSA-producing prostate adenocarcinomas

Survival — Despite the increasing reliance upon surrogate measures of clinical benefit, prolongation of survival is the ultimate gold standard for proving the superiority of one treatment over another. Demonstration of a survival benefit in men with HRPC with any cytotoxic agent or combination regimen was elusive until recently

Probably the most important reason for the difficulty in establishing a survival benefit is the heterogeneity that is inherent in any population of men with HRPC. This concept was nicely illustrated in a prognostic model for estimating survival in men with HRPC that was derived from 1101 such men who participated in therapeutic trials conducted by the Cancer and Leukemia Group B (CALGB). The model, which incorporated serum levels of lactate dehydrogenase, PSA, hemoglobin, and alkaline phosphatase, the disease Gleason score, Eastern Cooperative Oncology Group (ECOG) performance status and presence or absence of visceral disease, identified four risk groups with significantly different predicted median survival durations: 8.8, 13.4, 17.4, and 22.8 months, respectively. These variables were used to derive a nomogram that could be used to predict survival in individual men with HRPC

Quality of life — QOL parameters, including pain scales, are increasingly used as a measure of chemotherapy efficacy In fact, the United States Food and Drug Administration approval of mitoxantrone for HRPC was based almost exclusively on its beneficial impact on QOL While it is widely recognized that palliation of bone pain is a meaningful clinical endpoint, the use of other QOL parameters is less well standardized

MITOXANTRONE — Mitoxantrone was the first cytotoxic agent to be tested in a phase III trial in which QOL parameters were incorporated as a measure of response It is one of only two cytotoxic agents approved by the United States Food and Drug Administration for palliative treatment of HRPC.

Mitoxantrone versus corticosteroids — At least three trials have directly compared mitoxantrone to oral corticosteroids for the treatment of men with HRPC; two were performed in predominantly symptomatic men, while the third trial enrolled only asymptomatic men

  Symptomatic disease — An early trial randomly assigned 161 symptomatic men with metastatic HRPC to mitoxantrone (12 mg/m2 IV every three weeks) plus prednisone (10 mg daily) or the same dose of prednisone alone. The primary study endpoint was pain control, while secondary endpoints included duration of palliative response, PSA response, and overall survival.

Although combined therapy was associated with pain relief in significantly more men (29 versus 12 percent), there were no significant differences in either PSA response rate (33 versus 22 percent) or overall survival (less than 12 months in both arms). Nevertheless, palliative benefit was achieved by significantly more men with a PSA response that those without (53 versus 29 percent). Toxicity of combined therapy included grade 3 or 4 neutropenia in 45 percent of all treatment courses, although only 1 percent were complicated by fever.

A confirmatory CALGB study randomly assigned 242 men with HRPC (65 percent of whom were taking analgesics for bone pain) to mitoxantrone (14 mg/m2 IV every three weeks) plus hydrocortisone (40 mg daily) or the same dose of hydrocortisone alone. The primary endpoint was survival, while secondary endpoints included QOL (as assessed by the change from baseline analgesic use), and PSA response. As in the earlier trial, median survival was similar (approximately 12 months in both groups), and pain control was significantly better with combination therapy. Although the greater PSA response with combined therapy (38 versus 22 percent) achieved statistical significance, it was quantitatively similar to the earlier study, and the median time to disease progression was short in both groups (3.7 and 2.3 months, respectively).

  Asymptomatic disease — A lack of survival benefit from mitoxantrone was also seen in a third trial in which 120 asymptomatic men with metastatic HRPC were randomly assigned to prednisone (5 mg twice daily) with or without mitoxantrone (12 mg/m2 IV every three weeks) Although the median time to progression (8.1 versus 4.1 months) and PSA response rate (48 versus 27 percent) were significantly higher with combined therapy, survival was similar (23 and 19 months, respectively), approaching two years in both arms.

The most likely explanation for the apparently longer duration of survival in this study compared to the other two cited above is the degree of disease burden, as reflected in symptomatology and the baseline PSA values at study entry (71 and 59 ng/mL for men assigned to mitoxantrone with and without prednisone, respectively in the asymptomatic study compared to 150 and 158 ng/mL in the other two studies with predominantly symptomatic men). These data underscore the impact of disease heterogeneity when assessing the results of clinical trials in HRPC.

Mitoxantrone/prednisone versus docetaxel with or without estramustine — These results led to the adoption of mitoxantrone/prednisone as the standard against which other treatments for HRPC were compared. However, two subsequent randomized trials demonstrated the superiority of a docetaxel-based regimen over mitoxantrone and prednisone in men with metastatic HRPC, with both showing higher response rates, and a significantly longer median survival duration. These trials are discussed in detail in the following sections.

TAXANES — Both taxanes in clinical use, paclitaxel and docetaxel, bind to tubulin subunits and inhibit the disassembly of microtubules that normally occurs during cell cycle progression. Taxanes also inactivate the antiapoptotic protein bcl-2 by phosphorylation, thereby promoting apoptosis; whether this mechanism underlies their cytotoxic effect is unclear

Docetaxel — As a single agent, the efficacy of docetaxel in men with HRPC has been evaluated on an every three week and a weekly schedule

Two trials administering therapy every three weeks used the same dose (75 mg/m2), and outcomes were similar. In the larger of the two, which included 35 men with HRPC, a PSA response was noted in 46 percent, objective responses in 28 percent of those with measurable disease, and median survival was an impressive 27 months. Toxicity included a pulmonary embolus in one patient, grade 4 neutropenia in 43 percent, and grade 3 hyperglycemia in 50 percent. Median survival was not reported in the second study

Although weekly administration of docetaxel is associated with less myelotoxicity, it produces cumulative increases in hyperlacrimation, skin, and nail toxicity Moreover, it is unclear whether the therapeutic results are comparable. In three available phase II studies totaling 116 men with HRPC, using a similar docetaxel dose (35 to 40 mg/m2 per week for 6 of every 8 weeks), PSA response rates ranged from 34 to 64 percent. Median survival was 20 months in one report  but only approximately 9 months in the other two. The superiority of an every three week schedule was further suggested in a randomized trial in which prednisone plus docetaxel administered on an every three week but not a weekly schedule provided a survival advantage when compared to mitoxantrone/prednisone

Docetaxel/prednisone versus mitoxantrone/prednisone — In the landmark multicenter TAX-327 trial, 1006 men with chemotherapy-naive metastatic HRPC (either asymptomatic or symptomatic with stable pain scores and analgesic requirements) were randomly assigned to docetaxel in one of two schedules (75 mg/m2 every three weeks, or 30 mg/m2 weekly) or mitoxantrone (12 mg/m2 every three weeks), with all patients receiving prednisone 5 mg orally twice dailyThe primary endpoint was overall survival.

In a report with median 21 month follow-up, compared to mitoxantrone/prednisone (M/P), men receiving every three week (D/P) but not weekly docetaxel/prednisone (wD/P) had a significantly longer median survival (18.9 and 17.4 versus 16.5 months for D/P, w D/P, and M/P, respectively). Moreover, both D/P arms were associated with a higher PSA response rate (45 and 48 versus 32 percent), and a higher pain response rate (35 and 31 versus 22 percent). As expected, grade 3 or 4 neutropenia during therapy was most common with D/P (32 versus 1.5 and 22 percent with D/P, wD/P, and M/P, respectively), although rates of neutropenic infection were low (3 versus 0 and 2 percent, respectively), and few patients discontinued therapy because of adverse effects (11, 16, and 10 percent with D/P, wD/P, and M/P, respectively). There were no thromboembolic events.

These data have established every three week docetaxel plus daily prednisone as a new standard of care for men with HRPC, and further support the use of the every three week compared to weekly schedule of docetaxel administration. Largely on the basis of this trial, docetaxel has been approved in the United States for the treatment of men with HRPC, in combination with prednisone.

Docetaxel/EMP — Estramustine phosphate (EMP) is rapidly dephosphorylated in vivo, producing estramustine, which is thought to exert its cytotoxic effects by dysregulation of normal microtubule assembly, and binding to the nuclear matrix. As a single agent, EMP has minor activity in HRPC. In a contemporary multiinstitution series, rates of objective response, PSA response, and subjective pain relief were less than 5, 21, and 31 percent, respectively

EMP appears to have greater potential in combination regimens. In vitro studies suggest that EMP potentiates the activity of other cytotoxic agents that interact with the microtubule network, such as taxanes, vinca alkaloids, and etoposide. Docetaxel has a higher affinity for tubulin than does paclitaxel, and exhibits greater potential for synergy with EMP

  Thromboembolic events — However, one of the risks inherent to the use of EMP, particularly in combination with taxanes, is thromboembolic events. The reported incidence is consistently around 10 percent Although both venous and arterial events have been reported (ie, deep venous thrombosis, pulmonary embolism, stroke, and myocardial infarction), the risk appears higher for venous thromboembolic disease than for arterial thrombosis (5 to 6 versus 1 percent or less in a meta-analysis of 23 studies enrolling a total of 896 men with HRPC). Daily asiprin or low dose warfarin have been proposed as prophylaxis however, a decrease in thromboembolic complications using this strategy has not been tested in a randomized trial.

  Phase II studies — The antitumor activity of EMP plus docetaxel was initially demonstrated in two phase I trials and subsequently confirmed in several phase II trials using weekly, every other week, and every three week docetaxel administration, and different EMP doses In many of these uncontrolled studies, median survival durations approached two years.

   In a representative multicenter CALGB trial, docetaxel (70 mg/m2 every three weeks) and EMP (10 mg/kg per day, days one to five) were combined with low dose hydrocortisone (40 mg daily) in 47 men with HRPC  The lack of contribution of corticosteroids to the efficacy of this regimens has been demonstrated . The PSA response rate was 68 percent, 12 of 24 men with measurable disease had an objective response (including three complete responses), and the median survival was 20 months.

   The more favorable myelotoxicity profile of weekly docetaxel (35 mg/m2 per week) in conjunction with EMP (280 to 420 mg PO TID for three days each week) and dexamethasone (4 mg PO three times daily for three days weekly, two weeks of every three) was shown in a study of 18 men with HRPC. In a preliminary report, 72 percent had a PSA response, objective responses occurred in four of eight with measurable disease (one complete, three partial), and median survival was 16 months.

  Docetaxel/EMP versus mitoxantrone/prednisone — These results led to a randomized trial comparing 21 day cycles of docetaxel (60 mg/m2, increased to 70 mg/m2 for subsequent cycles in the absence of toxicity) plus EMP (280 mg orally three times daily days 1 to 5) versus mitoxantrone (12 mg/m2, increased to 14 mg/m2 for subsequent cycles in the absence of toxicity) plus prednisone (5 mg twice daily continuously) in 777 men with HRPC. Warfarin (2 mg daily) and aspirin (325 mg daily) were administered to the docetaxel/EMP group as prophylaxis against thromboembolic events. The primary endpoint was overall survival.

In a report with average 32 month follow-up, median survival was modestly but significantly better with docetaxel/EMP (17.5 versus 15.6 months) as was median progression-free survival (6.3 versus 3.2 months) and PSA response rate (50 versus 27 percent). Docetaxel/EMP was also associated with significantly more grade 3 or 4 gastrointestinal, cardiovascular, metabolic, and neurologic toxicity, although this did not translate into a higher rate of toxic deaths or study withdrawal compared to the mitoxantrone/prednisone arm. Prophylactic anticoagulation did not protect against treatment-related thromboembolic events.

Although these results support the superiority of docetaxel/EMP over mitoxantrone/prednisone, it is difficult to endorse the continued use of EMP in view of the similar survival benefit and better tolerability of docetaxel plus prednisone compared to mitoxantrone/prednisone in the TAX-327 study, and the elevated risk of venous and arterial thromboembolism in patients receiving EMP

Other docetaxel combinations under study — Several other regimens that include docetaxel show promising results, although randomized trials will be needed to demonstrate their superiority over docetaxel plus prednisone.

  Docetaxel plus calcitriol — Calcitriol, a natural ligand to the vitamin D receptor, may enhance the cytotoxicity of docetaxel In one early trial, in which 37 men received weekly therapy with Rocaltrol (0.5 µg/kg in four divided oral doses on day 1) followed by docetaxel (36 mg/m2 on day two), 30 (80 percent) achieved a PSA response, while 8 of 15 with measurable disease had a partial response, and median survival was 19.5 months

These intriguing results led to a randomized multicenter phase II study sponsored by two pharmaceutical firms (the ASCENT trial), in which 250 men with metastatic HRPC, castrate testosterone levels, and either clinical or biochemical progression after antiandrogen withdrawal were randomly assigned to weekly docetaxel (36 mg/m2 for three of every four weeks with standard dexamathasone) and either 45 mcg of a high-dose oral formulation of calcitriol (DN-101) or placebo, given on day one of each cycle prior to docetaxel. In a preliminary report with a median eight month follow-up, there was a trend towards a higher PSA response (58 versus 49 percent) and objective reponse rate in soft tissue (28 versus 20 percent) with DN-101; neither reached the level of statistical significance, and survival endpoints were immature.

  Docetaxel plus vinorelbine — Single agent vinorelbine is of marginal benefit as in men with HRPC However, at least two reports suggest that weekly administration of both docetaxel and vinorelbine is active and well-tolerated], prompting an ongoing phase III trial comparing this regimen with docetaxel plus EMP.

Paclitaxel — The antitumor efficacy of single agent paclitaxel in HRPC is schedule-dependent. In an early ECOG study of 23 men with HRPC with measurable disease receiving paclitaxel monotherapy (135 or 170 mg/m2 by continuous infusion over 24 hours every three weeks) an objective response (accompanied by an 80 percent decline in PSA for nine months) was noted in only one patient, and hematologic toxicity was prominent (grade 3 or 4 neutropenia, and febrile neutropenia in 75 and 26 percent of patients, respectively)

The efficacy of weekly therapy was illustrated in a study of 18 men with HRPC who received paclitaxel (150 mg/m2 over one hour) weekly for 6 of every 8 weeks. A PSA response was noted in 39 percent, and four of eight men with measurable disease had a partial response. Despite a favorable hematologic toxicity profile, grade 3 peripheral neuropathy developed in six men (36 percent).

  EMP plus paclitaxel — In one of the earliest trials of combined antimicrotubule therapy for HRPC, in which paclitaxel (120 mg/m2 by continuous infusion for 96 hours, every three weeks) and EMP (600 mg/m2 daily, given continuously) were administered to 35 men, an objective response was noted in 4 of 9 with measurable disease, the PSA response was 53 percent, and the median survival for the entire group was 17 months

These encouraging results, and emerging data supporting the toxicity-sparing effect of weekly administration led to follow-up trials combining weekly paclitaxel and EMP in a variety of doses. In one study, in which 66 men received paclitaxel (90 mg IV over one hour weekly) plus EMP (140 mg three times daily for three days weekly), only 4 of 26 with measurable disease had an objective response, but the PSA response rate was 42 percent Grade 3 or 4 neutropenia, edema, and thrombosis occurred in 6, 5, and 6 percent, respectively, and median survival was 16 months.

Efficacy was not improved with higher paclitaxel/EMP doses. In a study of 28 men receiving paclitaxel (150 mg/m2 weekly) and EMP (280 mg three times daily, three days weekly), objective responses were noted in 5 of 13 with measurable disease, PSA response was 62 percent, and the median survival 13 months.

Out of concern for thromboembolic events, the importance of EMP to these combined antimicrotubule regimens has been questioned . One randomized phase II trial comparing weekly paclitaxel (100 mg/m2 weekly for three of every four weeks) with and without EMP (280 mg three times daily for three days every week) in 166 men with HRPC suggests the necessity of EMP. In a preliminary report, the PSA response rate was twofold higher with combined therapy (48 versus 25 percent), and there was a trend towards improved 12 month progression-free survival (29 versus 8 percent, p = 0.08); median survival was not reported.

However, as noted above, the contribution of EMP to docetaxel-based regimens has been called into question by the results of the TAX-327 trial

INITIAL CHEMOTHERAPY FOR HRPC: SUMMARY AND RECOMMENDATIONS — The optimal treatment for men with HRPC has become more clear with the recent demonstration of a survival benefit from docetaxel-based regimens as compared to mitoxantrone/prednisone. Based upon the results of the TAX-327 trial, every three week administration of docetaxel plus continuous daily prednisone was approved by the United States Food and Drug Administration for treatment of men with HRPC, and this combination should now be considered the standard against which other treatments must be compared

Although a survival benefit for docetaxel plus EMP compared to mitoxantrone and prednisone was also shown in a large randomized trial, it is difficult to endorse the continued use of EMP in view of the similar survival benefit of docetaxel plus prednisone compared to mitoxantrone/prednisone in the TAX-327 study, and the elevated risk of venous and arterial thromboembolism in patients receiving EMP.

Although significant gains are beginning to be realized in the treatment of men with HRPC, median survival remains less than two years, and there is much room for improvement. Men with HRPC should be encouraged to participate in clinical trials testing new therapeutic strategies.

CHEMOTHERAPY FOR HRPC WITH LOW PSA PRODUCTION — As mentioned above, rising serum PSA levels in men with HRPC generally suggest disease progression, while a decrease in PSA, particularly in response to cytotoxic therapy, correlates with both clinical response and survival However, some men with aggressive metastatic disease have low serum PSA values. One report evaluated 18 such patients with a median serum PSA of 1.6 ng/mL (0 to 9.5 ng/mL) All had elevations in at least one of several other tumor markers (eg, CEA, CA 19-9, CA 153, CA 125). Histologically, the patients could be divided into two patterns:

   Neuroendocrine (ie, small cell) features with variable elements of adenocarcinoma
   Poorly differentiated prostate adenocarcinoma

Men whose disease falls into these categories are more likely to have visceral and osteolytic bone lesions, and the prostate cancer is often resistant to hormone ablation. Furthermore, serum PSA is not suitable as a surrogate marker of treatment benefit

However, these tumors are relatively sensitive to chemotherapy regimens such as those used for small cell cancer involving the lung (eg, platinum/etoposide combinations), with response rates of over 50 percent  Nevertheless, the optimal treatment regimen is unknown, and clinical studies are sparse.

In one series, 30 men with neuroendocrine features based upon morphology alone, or by immunohistochemistry received cisplatin or carboplatin plus etoposide and EMP. Regression of measurable lesions was noted in four of nine cases of small cell carcinoma (44 percent), 6 of 13 poorly differentiated carcinomas (46 percent), 7 of 13 tumors with one neuroendocrine marker immunohistochemically detected (54 percent), and three of seven tumors without any positive staining (43 percent). Thus, the presence of neuroendocrine differentiation did not predict for a response to chemotherapy.