In view of all these variables in the assessment of new treatments, one must ensure that there is rigor and structure in the assessment of progress in this field, which is best done through the use of randomized trials. Non-comparative phase I-II trials have been essential in identifying the potential role of the conventional and novel cytotoxic agents and combination regimens for advanced prostate cancer [1, 2, 21–26]. The following conventional drugs have been shown to have anti-cancer effect against bone-dominant, hormone refractory prostate adenocarcinoma when used as single agents: doxorubicin, cyclophosphamide, cisplatin, carboplatin, mitoxantrone, paclitaxel, docetaxel and mitomycin C [1, 2, 21]. Each produces objective response rates of about 15–20 %, sometimes with reduction of tumor-related symptoms and/or improvement in measured quality of life.

The problem is that the newer endpoints of assessment could easily lead to misinterpretation of the benefit from novel agents, as compared to the utility of some of the older drugs studied in the trials of the 1980s and 1990s. Softer end points (PSA response, reduction in circulating tumor cell numbers, application of waterfall plots, and a new emphasis on progression-free survival) or changes in the population of patients treated may have led us into the trap of making it too easy to attribute “patient benefit” to novel agents. There is a relatively new fashion of expressing tumor reduction through the vehicle of “waterfall plots”. In a waterfall plot, reductions of tumor mass are graphically expressed as a bar below the baseline, and reductions as small as 5 % can be interpreted to indicate treatment benefit, a much lower bar than the 50 % mass reductions required in earlier classification systems. We may thus have begun to over-emphasize the utility of some novel agents [27].

Ultimately the most reliable way to prove that novel agents are achieving more than older compounds is through randomized clinical trials. Through this vehicle, direct comparison has shown that mitoxantrone improves QOL compared to non-cytotoxic treatment [13, 14], and that docetaxel is more active against prostate cancer than is mitoxantrone (at the expense of more toxicity) [3, 4]. This paradigm will continue to be essential to the accurate evaluation of novel agents for prostate cancer.

Although single agent chemotherapy has been shown to improve QOL and yield a modest survival benefit, randomized trials have not yet proven the superiority of combination chemotherapy over single agents. Some promising combination regimens were reported in non-randomized trials in the past decade, suggesting that combination chemotherapy may increase patient benefit, despite the problems of stage migration and altered endpoints. However, few confirmatory randomized trials have proven this benefit.

The combination of paclitaxel or docetaxel plus estramustine has been reported to produce subjective response in approximately 50–60 % and PSA response in 40–75 %. Objective tumor responses occurred in up to 30 %, depending on whether patients with soft tissue disease dominated the population of patients [22–24]. Others have reported that the addition of carboplatin to the paclitaxel-estramustine doublet adds to clinical activity, claiming higher objective and PSA response rates [25, 26]. A meta-analysis of randomized trials comparing chemotherapy versus chemotherapy plus estramustine concluded that there was a survival benefit from the addition of estramustine [28]. However, once again, a decade after initial reporting of estramustine-based combinations, these data seem to remain controversial. There are significant concerns regarding the thrombo-embolic effects of this combined therapy, and many believe that the gold standard remains single-agent chemotherapy for castrate-resistant prostate cancer.

More recently, cabazitaxel, formerly known as XRP6258, a new generation semi-synthetic taxane with low affinity for multidrug resistance protein, was demonstrated in a phase I trial to show efficacy against docetaxel-treated castrate-resistant prostate cancer [29]. In an international randomized trial, known as “TROPIC”, 755 patients were allocated to treatment with mitoxantrone or cabazitaxel after prior failure of docetaxel [30]. A statistically significant median survival benefit of 3.4 months was noted with cabazitaxel. In several eastern European centers, toxic deaths were experienced in patients treated with this novel agent, emphasizing the hazards of myelosuppression in an elderly population with potential renal dysfunction and bone marrow compromise. These outcomes highlight the importance of subspecialty experience when applying novel therapies to an at-risk population [30]. Nonetheless, this agent was approved by the Food and Drug Administration for use in North America for castration-resistant prostate cancer.

Studies are being implemented to test cabazitaxel versus docetaxel in front line therapy for castrate-resistant disease and to assess utility and safety of novel combination regimens that employ this new agent.

Novel compounds that have some functional similarity to the taxanes, the epothilone B analogues, have been tested against prostate cancer. One of these agents, ixabepilone, has recently been shown to induce objective response in up to 30 % of cases, and a 50 % PSA reduction in 48 % of patients [31]. Similar patterns of response have been recorded with a structural analogue, patupilone [32]. This concept will require further testing, with a focus on randomized clinical trials, to compare activity with that of the taxanes, and to evaluate utility in combination regimens. To our knowledge, this class of compounds has not been approved for use in North America by the Food and Drug Administration, and they thus remain investigational.