|Home | About | Journals | Submit | Contact Us | Français|
The major goals of the Southwest oncology genitourinary (SWOG-GU) committee in the area of advanced prostate cancer are to improve the survival and quality of life (QOL) of patients with advanced prostate cancer. SWOG trials have examined the role of combined androgen blockade, intermittent androgen deprivation, and the early application of chemotherapy in castration-naïve disease. In addition, they have contributed to advancing the current chemotherapy standard of docetaxel plus prednisone, and ongoing trials seek to improve upon that standard. Finally, surrogate endpoints have been identified and markers of treatment response or resistance with novel technology are under active investigation. This review highlights findings from recent SWOG clinical trials for advanced prostate cancer, emphasizing the clinical impact and future applications of the data.
Over 192,000 cases of prostate cancer are diagnosed each year in the United States [Jemal et al. 2009]. Although only 12–20% of men initially present with advanced stage disease [Oakley-Girvan et al. 2003], they will be joined by up to 30% of men who are initially treated with curative intent for localized prostate cancer, but who go on to suffer relapse. In this group of men, for whom prostate cancer threatens survival, the Southwest Oncology Group (SWOG) has completed landmark clinical trials answering significant questions which have helped both to define and refine the standards of care.
The major focus of research for the SWOG-genitourinary (GU) committee in the area of advanced prostate cancer is to improve the survival and quality of life (QOL) of patients with advanced prostate cancer. The successful fulfillment of the GU committee’s vision to impact the treatment of metastatic prostate cancer has been facilitated by a synergistic interaction among clinical and scientific thought leaders, a multidisciplinary focus, and collaborative strategies, while providing mentorship for the next generation of prostate cancer researchers.
This review will highlight some of the most important findings from recent SWOG clinical trials for advanced prostate cancer, emphasizing the clinical impact and future applications of the data. A summary of the cited SWOG trials is provided in Table 1.
Dramatic reductions in prostate cancer tumor mass are achieved through removal of testosterone [Huggins, 1942]; however, the cancer ultimately progresses toward the lethal castration-resistant phenotype. Recognition that castration reduces plasma testosterone by 90% but only reduces tissue levels of the more potent driver of prostate cancer – dihydrotestosterone (DHT) – by 75% [Geller et al. 1978] inspired the development of maximal androgen blockade protocols. Early reports of improved survival using an anti-androgen in conjunction with castration [Labrie et al. 1985] prompted several trials nationally and internationally. Two randomized trials to formally investigate the hypothesis were conducted by SWOG: SWOG 8494 and SWOG 8894. In the first trial, the gonadotropin-releasing hormone (GnRH, also known as LHRH) analogue, leuprolide, was used alone or in combination with flutamide, and a significant survival advantage was detected favoring the combination [Crawford et al. 1985]. In the second trial, men who had undergone surgical castration were randomized to receive the anti-androgen, flutamide, or placebo but the difference in survival did not reach statistical significance [Eisenberger et al. 1998].
These trials, in conjunction with data from smaller randomized trials published concurrently [Keuppens et al. 1993; Tyrrell et al. 1993], placed combined androgen deprivation within the realm of standard practice when a LHRH-analogue is used and led to its incorporation in subsequent clinical trials, such as SWOG 9346 and JPR7. With newer technology, the importance of maximally starving the prostate cancer cell of androgens has become even clearer. Gene expression studies show that androgen-regulated genes, such as the androgen receptor and prostate-specific antigen (PSA), remain activated during castration [Mostaghel et al. 2007], potentially due to the persistent presence of androgens. The concept of maximal androgen deprivation continues to move forward, with a panoply of agents designed to block the androgen receptor or prevent intra-tumoral production of DHT.
The emergence of castration resistance occurs during ADT via multiple mechanisms in addition to residual tissue androgens [Feldman and Feldman, 2001]. Preclinical data culminated in two theories, advanced by Isaacs and Coffey (1991), which postulated that intermittent exposure to ADT may delay the development of resistance: clonal selection and molecular adaptation [Isaacs and Coffey, 1991]. The clonal selection hypothesis suggests that in a testosterone-depleted environment, the sensitive epithelial cells undergo apoptosis, while the basal cells, a source of pre-existing castration resistance, can flourish. Bruchovsky and colleagues found, using a Shionogi mouse model that after a period of androgen deprivation there was an increased proportion of castration-resistant stem cells, but that when testosterone was re-introduced, the predominant surviving stem cell clones shifted to an androgen-dependent population [Bruchovsky et al. 1990]. Molecular adaptation describes the various androgen receptor (AR) escape mechanisms which have been documented in castration-resistant cells, including AR amplification, activation by alternative ligands, and independent AR pathway signaling facilitated by co-activators or activation of downstream pathway targets [Grossman et al. 2001]. SWOG put these theories to the test by designing trial 9346, a randomized comparison of intermittent versus continuous ADT for metastatic prostate cancer. Survival is the primary endpoint. Accrual of a total cohort of 3,040 and randomization of over 1,704 men has been completed, and data are currently awaited. This trial has ample power to define the role of intermittent hormone therapy, hence the potential to impact the standards of care.
In addition to the survival endpoint, the data regarding toxicity and formal quality-of-life assessment from S9346 will be very important, given the recent identification of serious ADT side-effects, including osteoporosis, weight gain, insulin resistance, and lipid abnormalities. Data from small phase II trials have suggested better reported quality of life with intermittent therapy due to reduced time spent in a castrate state, reviewed by Bhandari [Bhandari et al. 2005]. If survival with the intermittent arm is not inferior, a positive outcome from the QOL evaluations in this trial will likely impact whether intermittent ADT becomes adopted as a standard in community practice.
Transformation from the castration-sensitive to the castration-resistant phenotype is due to expansion of a resistant clone that is either present at diagnosis, or emerges during the course of the disease. SWOG’s early attempts at targeting this resistant clone began over two decades ago with combined chemohormonal therapy using cyclophosphamide and 5-fluorouracil, which gave encouraging results [Mukamel et al. 1980]. SWOG 8219 was designed to compare concurrent versus sequential chemotherapy, utilizing doxorubicin and cyclophosphamide, with either diethylstilbestrol (DES) or castration for androgen deprivation. With 143 patients, the study was powered to detect a 50% survival advantage, but no difference was seen between the treatment groups despite a higher initial response rate in the combined chemohormonal therapy group [Osborne et al. 1990].
Given that no survival advantage had been shown for anthracycline therapy in castration-resistant prostate cancer (CRPC), the concept of early chemotherapy was revisited after taxanes were found to have activity in this disease. For SWOG S0032, paclitaxel, estramustine, and etoposide were selected, based on phase II data showing significant response rates [Smith et al. 1999]. With the emergence of a survival advantage with docetaxel this concept remains vital, especially as data from cell culture (LNCaP) and xenograft models suggest that concurrent chemohormonal therapy is more effective at reducing tumor burden than a sequential approach [Eigl et al. 2005]. The intergroup ECOG trial E3805 (CHAARTED) is currently investigating combined chemohormonal therapy, incorporating docetaxel plus prednisone, the combination which has been demonstrated to improve survival in CRPC [Sweeney, 2006].
Until the late 1990s, prostate cancer was generally considered to be a chemoresistant disease, although some selected phase II studies suggested clinical benefit [Logothetis et al. 1983; Torti et al.1983]. Ultimately, mitoxantrone received FDA approval in 1999 due to its clear palliative benefit, even in the absence of a documented survival advantage, notwithstanding the crossover design [Kantoff et al. 1999; Tannock et al. 1996]. However, interest in chemotherapy for CRPC was limited until a survival advantage was documented for docetaxel as compared to mitoxantrone [Petrylak et al. 2004; Tannock et al. 2004]. Now that chemotherapy has become widely utilized, SWOG will strive to improve upon the docetaxel backbone. Molecular understanding of docetaxel resistance has advanced interest in novel microtubule inhibitors, and broader thinking about the importance of the tumor microenvironment has guided the incorporation of biologic agents into the chemotherapy platform for treatment of CRPC.
The docetaxel era began with reports of rather striking preclinical activity of docetaxel, followed quickly by phase II studies showing PSA response rates of 65–70% and measurable disease responses with docetaxel-based combination chemotherapy [Savarese et al. 2001]. These data provided the rationale for phase III trials comparing docetaxel-based chemotherapy to the standard of mitoxantrone plus prednisone. SWOG 9916, an intergroup trial, was rapidly completed and demonstrated that docetaxel plus estramustine conferred a survival advantage compared to mitoxantrone [Petrylak et al. 2004]. These data, published concurrently with the industry-sponsored trial TAX327 [Tannock et al. 2004], provided supportive data for FDA approval of docetaxel plus prednisone to treat metastatic CRPC in 2004. This chemotherapy standard also improved QOL and pain palliation. SWOG rigorously evaluated these endpoints, with excellent compliance on the validated instruments. These robust data confirmed the palliative benefits of docetaxel compared to mitoxantrone, reassuring oncologists that the dual goals of prolonging life and minimizing suffering in advanced prostate cancer patients were achieved [Berry et al. 2008].
Having recognized the particular success of microtubule targeted therapy in prostate cancer (vinca alkaloids, estramustine, taxanes), SWOG has been instrumental in rapidly evaluating novel microtubule agents. SWOG S0111 investigated the efficacy of the epothilone compound, ixabepilone (BMS-247550), in chemotherapy-naïve, CRPC patients. The PSA response rate (33%) and objective response rate (15%), confirmed the substantial activity of epothilones in this disease [Hussain et al. 2005]. Another class of agents, the mitotic kinesin spindle protein (KSP) inhibitors, appeared promising for taxane-refractory disease due to its unique mechanism of action, as well as its lack of neuropathy, which limits other agents in this setting. SWOG S0418 investigated the response rate of the KSP inhibitor, ispinesib (SB-715992), in men with taxane-refractory metastatic prostate cancer. While the results were disappointing, with no responses seen [Beer et al. 2008], the question was answered rapidly and definitively.
Unfortunately, even with the new gold standard chemotherapy, docetaxel, fewer than half of men with metastatic CRPC have responses, and survival is still limited. Given that metastatic prostate cancer predominantly affects bone, designing trials incorporating our emerging understanding of the proteins and pathways which allow CRPC cells to flourish in the bone microenvironment is a high priority. RANK ligand, interleukin-6 (IL6), insulin-like growth factor, bone morphogenic protein, endothelin-1, epidermal growth factor, and platelet-derived growth factor have all been identified as biological targets of interest [Pinski and Dorff, 2005], though initial clinical investigations with available therapeutic inhibitors for the latter two pathways failed to yield significant clinical responses [Chung et al. 1999].
Atrasentan targets endothelin-1, a key molecule involved in the unique affinity of prostate cancer cells to grow in bone marrow. Early investigations confirmed suppression of prostate cancer activity in bone using serum and urine biomarkers such as alkaline phosphatase and n-telopeptide [Nelson et al. 2003]. The initial randomized phase II trial found a prolonged time to PSA progression compared to placebo in men with asymptomatic metastatic CRPC [Carducci et al. 2003]. Unfortunately, the single-agent randomized phase III trials failed to identify a statistically significant improvement in outcome [Carducci et al. 2007] in men with castration-resistant metastatic disease. Design flaws, particularly the inclusion of patients with no bone metastases and rapid removal of subjects with progression of PSA only, may have interfered with the adequacy of these trials to detect a benefit. In fact, subset analysis of patients with only bony metastatic disease suggested a statistically significant therapy effect. However, it is also quite plausible that inhibiting or modulating only one bone-specific target is unlikely to impact a disease with multiple redundant pathways driving the bone metastases. Therefore, combinations targeting multiple relevant pathways should be more effective. This hypothesis was tested preclinically, yielding evidence for synergy between taxanes and endothelin inhibition [Banerjee et al. 2007; Rosano et al. 2003; Del Bufalo et al. 2002]. SWOG S0421, which compares docetaxel, prednisone plus bone-targeted therapy using atrasentan and zoledronic acid to docetaxel, prednisone, Zoledronic acid plus placebo, builds upon the docetaxel foundation by adding this rationally derived targeted therapy. As of July, 2009, over 660 of 930 planned patients have been accrued to S0421, with the first interim analysis planned after 465 patients. This trial is uniquely designed with coprimary endpoints of progression-free survival (PFS) and overall survival (OS). Blood samples are being collected for assessment of several markers of bone formation (osteocalcin, pro-collagen I and III amino-terminal propeptides) as well as markers of bone resorption (deoxypyridinoline and n-telopeptide) at baseline and set timepoints, in order to assess their prognostic value and determine whether changes over time correspond to treatment response. A recent addition to the study has been the addition of collection of circulating tumor cells (CTCs), a purported new prognostic and predictive factor [De Bono et al. 2008]. Using a novel technology, the CTCs will be assayed for endothelin receptors and other potentially biologically relevant endpoints.
CNTO 328 is a chimeric antibody which inhibits IL6, another key facilitator of prostate cancer proliferation within the bone microenvironment. Preclinical studies documented inhibition of growth when prostate cancer cells were exposed to the anti-IL6 antibody [Chung et al. 2005] as well as substantial reduction in tumor volumes in xenograft models [Smith and Keller, 2001]. SWOG S0354, which was recently completed, was a phase II trial of this antibody in CRPC patients with progressive disease after taxane therapy. Results will be presented this year.
One of the greatest strengths of data from groups like SWOG is the multi-institutional experience, which is typically more applicable to the general patient population than data derived from a single institution. For instance, the combination of estramustine and etoposide was found to be highly promising in CRPC in several series [Berruti et al. 2005; Sumiyoshi et al. 2000; Dimopoulos et al. 1997; Pienta et al. 1997], however, SWOG 9407 found the response rate was dramatically lower at 14% [Pienta et al. 2001]. Toxicity was also noted to be higher than reported in the single-institution experiences. These data provide a more accurate depiction of real-world experience before such regimens become integrated into clinical practice.
Similarly, the phenomenon of anti-androgen withdrawal response was fully described and clarified in SWOG 9426 after a number of case series and single-institution experiences had been reported. Retrospective reports had suggested a 30% response rate for all anti-androgens in general [Caldiroli et al. 2001] and some case series had noted radiographic responses [Huan et al. 1997]. When analyzed specifically by anti-androgen agent, 15–50% of patients were noted to have a PSA response after flutamide withdrawal, lasting a median of 3.5 months in one report [Schellhammer et al. 1997; Small and Srinivas, 1995]. For bicalutamide, a 29% withdrawal PSA response was noted [Schellhammer et al. 1997]. The prospective study 9426, which enrolled 210 patients, documented PSA responses in 24% of flutamide-withdrawal patients, 13% of bicalutamide patients, and 25% of nilutamide patients [Sartor et al. 2008]. Median PFS was 3 months, and there were no radiographic responses. These data provide a reliable and realistic expectation, allowing clinicians to be confident in decision making as well as better information for discussions with their patients.
A second major impact of randomized cooperative group trial data is the ability to investigate the utility of intermediate endpoints thus leading to strong hypothesis-generating data to be prospectively tested. For instance, in the setting of newly diagnosed hormone-sensitive metastatic disease there have not been any strong, simple and consistent prognostic biomarkers that can be used at an individual patient level for decision making and prognostication regarding risk of death. In SWOG 9346 (intermittent versus continuous androgen deprivation) the absolute PSA value after 7 months of ADT was shown to be a significant predictor of mortality risk. Men with a PSA value greater than 4ng/dL at 7 months had a median survival from that point of only 13 months, compared to 44 months for men with a value between 0.2 and 4ng/dL, and 75 months for those with a PSA value less than 0.2ng/dL [Hussain et al. 2006]. Future trials, which are being formulated both in and outside SWOG to improve upon ADT, are using the 7-month PSA value data as an intermediate endpoint because of this powerful association with survival, and will also target the population with the suboptimal response as reflected by PSA levels for additional therapeutic interventions.
Objective response assessment has been a major problem in determining therapeutic efficacy in single-arm trials in CRPC; therefore, a simple, reproducible measure of response is highly desirable. SWOG 9916, in which docetaxel was compared to mitoxantrone for CRPC, provided a platform to further analyze whether or not PSA reductions ranging from 5–90% or changes in PSA velocity could predict which patients received the greatest overall survival benefit from treatment. While a 50% PSA reduction had been the consensus for uniformity of reporting of PSA response, in fact a 30% reduction in PSA after four cycles of docetaxel chemotherapy was shown to have optimal surrogate correlation with survival [Petrylak et al. 2006]. This is now being validated prospectively in S0421.
Determining early when there is disease progression based on PSA values can be clinically relevant. Using the data from SWOG 9346 and 9916, several definitions for PSA progression were evaluated for correlation with survival. PSA-P defined as an increase of ≥25% over nadir plus an absolute increase of at least 2 or 5ng/mL predicted overall survival in hormone-sensitive prostate cancer and CRPC, and may be a suitable endpoint for phase II studies in these settings. Specifically in S9346 patients, this definition predicted a 2.4-fold increased risk of death and over four times the risk of death if PSA progression occurred in the first 7 months. In S9916 patients, it predicted a 40% increase in the risk of death and a two-fold increased risk of death if PSA progression occurred at 3 months. [Hussain et al. 2009].
Prognostic variables for androgen-sensitive prostate cancer patients may be identified and validated retrospectively in the large data sets generated by SWOG trials. For instance, the presence of anemia at baseline as well as the decline in hemoglobin during ADT were both found to have prognostic value in SWOG 8894, with decline corresponding to inferior survival [Beer et al. 2006]. Similarly, body mass index (BMI) was examined in SWOG 8894 and 9916, and was found to be correlated with improved prognosis in androgen-naïve patients, with overweight but not obese subjects having better progression-free and overall survival [Montgomery et al. 2007].
During SWOG 9346, the trial of intermittent versus continuous androgen deprivation, blood was collected for correlation with polymorphisms in candidate genes involved in steroid synthesis, androgen-receptor signaling and the insulin-like growth factor pathway. In addition to the aforementioned analysis involving markers of bone formation and resorption in SWOG S0421, promising predictors of docetaxel response will be prospectively tested: CYP450, MDR and β-tubulin [Goh et al. 2007]. These investigations, which include application of innovative technology, will contribute to the understanding of treatment resistance by identifying specific gene polymorphisms as well as broader gene and protein expression profiles of responsive versus resistant tumors, crucial preliminary steps toward the development of individualized therapy.
SWOG has been a leading innovator in prostate cancer therapy, answering significant questions which have helped optimize the fundamental treatments. SWOG-GU committee investigators are currently testing new strategies and novel agents, focusing on extending survival without overlooking quality of life. We are actively involved in validating biomarkers of therapeutic efficacy and outcome. Ongoing trials will continue to enhance the quality of care for men with advanced prostate cancer.
The authors wish to acknowledge the leadership and vision of Dr E. David Crawford, the chairman of the GU committee for over 25 years. We also wish to acknowledge the contributions and leadership of Dr Derek Raghavan the vice chair of the GU committee (1998–2007)
The various studies discussed in this paper were supported in part by the following PHS Cooperative Agreement grant numbers awarded by the National Cancer Institute, DHHS: CA32102, CA38926, CA58348, CA27057, CA20319, CA68183, CA42777, CA14028, CA58882, CA46441, CA35192, CA46282, CA128567, CA45907, CA46113, CA58416, CA04919, CA58861, CA76132, CA58686, CA12644, CA35261, CA35431, CA22433, CA46368, CA63848, CA67575, CA76447. CA46136, CA86780, CA35281, CA45560, CA63844, CA37981, CA67663, CA11083, CA35178, CA95860, CA16385, CA12213, CA35119, CA35090, CA63845, CA74647, CA45461, CA45377, CA45808, CA76462, CA35128, CA35262, CA45807, CA35176, CA76426, CA58658, CA76429, CA63850, CA13612, CA21076, CA45450, CA58723, CA04920, CA32734, CA52650, CA28862, CA52386, CA35200, CA45466, CA52420; CA52654, CA35117, CA52772, CA35283, CA073590, CA35421, CA74811; CA60138 (CALGB); CA25224 (NCCTG); CA21115 (ECOG); 2U10CA11488-28 through 5Y10CA011488-38 (EORTC); NCI/CTEP sponsored Translational Research Initiative contract (#24XS146), and in part by AstraZeneca Pharmaceuticals, Sanofi-Aventis US, Inc., Abbott Laboratories, Pharmacia, Inc., Schering-Plough Research Institute, Immunex Corporation, and Centocor, Inc.
David Quinn has received speaker honoraria from Sanofi Aventis, a sponsor of SWOG 9916 and S0421. Daniel Petrylak has received research support from Sanofi Aventis. All other authors have declared no conflicts of interest relevant to the manuscript.