Bone is a rigid yet dynamic organ that is continuously shaped and repaired by a process termed remodeling. Bone remodeling is the predominant metabolic process regulating bone structure and function during adult life. Bone remodeling involves the breakdown and formation of bone by the coordinated action of osteoclasts and osteoblasts (Fig. 14A.1). Bone remodeling occurs in discrete microscopic units throughout the skeleton. Remodeling of each unit is geographically and temporally separated from other units. The process is initiated by migration of osteoclasts to these sites, resorption of bone, apoptosis of osteoclasts, and new bone formation by osteoblasts.

Osteoclasts are tissue-specific macrophages derived from hematopoietic stem cells from the monocyte/macrophage lineage. Bone resorption is a multistep process initiated by proliferation of immature osteoclast precursors, commitment of these cells to the osteoclast phenotype, and degradation of the bone matrix by mature multinucleated osteoclasts. Osteoclasts attach to the bone matrix and develop a specialized cytoskeleton or ruffled border. A process involving proton transport degrades the matrix in the isolated microenvironment at the ruffled border. The bone matrix is a storehouse of growth factors, and normal bone resorption releases a variety of factors that activate osteoblasts including transforming growth factor β(TGFβ), basic fibroblast growth factor (bFGF), platelet-derived growth factor (PDGF), and insulin-like growth factor (IGF) I and II. Osteoblasts are derived from stromal stem cells. Osteoblasts synthesize and secrete the organic matrix. Mineralization begins soon after the organic matrix is secreted and is completed within weeks.

The RANK signaling pathway regulates the activation, differentiation, proliferation, and apoptosis of osteoclasts (1). This pathway consists of receptor activator of NF-κB ligand (RANKL), its receptor RANK, and its decoy receptor osteoprotegerin (OPG). RANKL binds and activates RANK, a transmembrane receptor expressed on hematopoietic stem cells and osteoclasts. RANK expression on stem cells is required for osteoclast differentiation and activation. Hormones and other factors that stimulate bone resorption induce the expression of RANKL by bone stromal cells. RANKL expression by osteoblasts coordinates bone remodeling by stimulating bone resorption by osteoclasts that in turn stimulates new bone formation by adjacent osteoblasts in a process termed coupling. RANKL signaling pathway is negatively regulated by OPG. OPG is a soluble protein produced by osteoblasts in response to anabolic agents including estrogen and bone morphogenic proteins (BMPs). OPG acts as a decoy receptor that blocks RANKL binding to RANK. OPG also acts as a decoy receptor for TNF-related apoptosis-inducing ligand (TRAIL).

Prostate cancer preferentially spreads to the skeleton. More than 80% of men who die from prostate cancer have bone metastases at autopsy (2). In contrast to most other cancers, prostate cancer forms predominantly osteoblastic metastases. The vertebral column, pelvis, ribs, and proximal long bones are the most common sites of skeletal metastases (Fig. 14A.2). These sites correspond to the most abundant regions of bone marrow in the skeleton.

Pain is the most common symptom of metastatic prostate cancer. Metastases to the vertebrae may cause spinal cord compression, nerve root compression, or cauda equina syndrome. Metastases to the base of the skull can impinge on cranial nerves. Clinical fractures are common, with most fractures involving the vertebral bodies. In contrast to osteolytic bone disease, pathological fractures of long bones due to prostate cancer bone metastases are unusual. Hypocalcemia due to excess deposition of calcium in newly formed bone is commonly observed but rarely associated with symptoms. Most men with metastatic prostate cancer have normochromic normocytic anemia.

Most bone metastases in men with prostate cancer appear osteoblastic by radiographic imaging. Osteolytic and osteoblastic lesions represent two extremes of a spectrum, however, and morphologic studies suggest that most bone metastases from prostate cancer are characterized by excess bone formation and excess bone resorption. This pathological acceleration of bone remodeling results in disorganized bone with impaired biomechanical properties.

Osteoblastic metastases from prostate cancer feature increased osteoblast number and activity, increased bone volume, and increased bone mineralization rate (3). Osteoclast number and activity are increased in osteoblastic metastases, at bone adjacent to metastases, and distant uninvolved bone (3,4). Biochemical markers of osteoclast activity are also elevated in men with osteoblastic metastases from prostate cancer (5). Although osteoclast activity is increased in men with prostate cancer, it is unclear whether osteoclast activation precedes bone formation as in normal bone remodeling or is secondary to excessive osteoblast activity in the metastases. Markers of osteoclast activity independently predict the risk for subsequent skeletal complications (6), suggesting that cancer-mediated osteoclast activation not only accompanies bone metastases but also contributes to the clinical complications of metastatic disease.

Hyperparathyroidism is common in men with metastatic prostate cancer. Serum concentrations of calcium are typically normal or low, suggesting that excessive parathyroid hormone secretion is secondary to increased calcium phosphate deposition in osteoblastic metastases (7). Secondary hyperparathyroidism may propagate a vicious cycle involving parathyroid hormone activation of osteoclasts, release of bone-derived growth factors, tumor cell proliferation, osteoblast activation, and further calcium phosphate deposition.

Androgen deprivation therapy, the cornerstone of treatment for recurrent and metastatic prostate cancer, also contributes to osteoclast activation in men with prostate cancer. Androgen deprivation therapy accelerates normal bone turnover and increases biochemical markers of osteoclast activity by approximately twofold in men without bone metastases (8). The unintended skeletal effects of androgen deprivation therapy increase fracture risk in men with prostate cancer (9). In addition, treatment-related increases in osteoclast activity may promote progression of bone metastases by increasing the release of bone-derived growth factors.

Zoledronic acid is a potent bisphosphonate that inhibits osteoclast attachment, differentiation, and survival (10). In a randomized placebo-controlled trial, zoledronic acid decreased the risk of skeletal complications in men with bone metastases and progressive disease after first-line hormonal therapy (11). The efficacy of zoledronic acid supports the concept that excessive osteoclast activity has a causal role in skeletal complications from metastatic prostate cancer.

Recent work has led to a growing understanding of the cellular changes that allow prostate cancer to travel to bone. Cancer cells must acquire the ability to detach from the primary tumor and survive outside of their native tissue, a process known as the epithelial-mesenchymal transition. PI3K/Akt, Wnt/β-catenin, Notch, Ras, integrin-linked kinase, and other integrin signaling pathways are all thought to be involved in cancer epithelial-mesenchymal transition (12).

The concept that interactions between “seed” (circulating tumor cells) and “soil” (metastatic site) determine the metastatic potential of a cancer was first proposed by Stephen Paget more than a century ago (13). Consistent with the “seed and soil” hypothesis, reciprocal interactions between prostate cancer cells and bone stroma appear to account for both the predominant skeletal localization of metastases and the characteristic osteoblastic response (Fig.14A.3).

The bone extracellular matrix supports cellular adhesion and may contribute to the preferential skeletal localization of prostate cancer metastases. Integrin-mediated adhesion to the extracellular matrix increases androgen responsiveness and increased expression of androgen-responsive genes (14,15). In addition, adhesion to the extracellular matrix increases androgen-independent expression of some androgen-responsive homeobox genes (16), suggesting that interactions with the extracellular matrix contribute to androgen-independent growth.

Several bone-derived growth factors, including TGFβ, epidermal growth factor (EGF), and bFGF, promote prostate cancer growth and differentiation. Bone-derived growth factors may also contribute to the preferential skeletal localization of prostate cancer metastases. TGFβ facilitates adhesion to the extracellular matrix (17). EGF promotes migration of prostate cancer cells (18). Stromal cell-derived factor-1 (SDF-1) can bind to its receptor CXCR4 on prostate cancer cells to promote both adhesion and migration (19).

Many tumor-derived factors have been implicated in the pathogenesis of osteoblastic metastases and have led to the concept that prostate cancer cells acquire a bone-like phenotype. Cancer-produced factors include endothelin-1 (ET-1), BMPs, IGFs, OPG, TGFβ, and serine proteases prostate-specific antigen (PSA) and urokinase-type plasminogen activator (uPA).

Bone metastases are a major cause of morbidity from prostate cancer. Prostate cancer spreads preferentially to the skeleton. Although most bone metastases in men with prostate cancer appear osteoblastic by radiographic imaging, “osteoblastic metastases” are characterized by abnormal activation of both osteoblasts and osteoclasts. Pathological acceleration of bone remodeling results in disorganized bone with impaired biomechanical properties. Complex reciprocal interactions between tumor and bone account for both the preferential spread to skeleton and the characteristic osteoblastic phenotype. Several tumor-derived factors have been implicated in the pathogenesis of osteoblastic metastases, including ET-1, BMPs, IGFs, OPG, and TGFβ. Several bone-derived growth factors, including TGFβ, EGF, bFGF, and RANKL, appear to promote prostate cancer growth. Greater understanding of these tumor-bone interactions may provide future therapeutic opportunities.

Androgen deprivation therapy (ADT) is well established as the initial treatment of metastatic prostate cancer. Surgical bilateral orchiectomy was the earliest method of ablating the production of androgens. John Hunter noted in 1786 that castration decreased prostate mass in bulls (1). Castration was later utilized as a therapy for benign prostatic hypertrophy causing urinary obstruction (2). In 1941, Huggins and Hodges reported on the effects of castration on prostate cancer metastatic to the bone while delineating the role of testosterone in promoting prostate cancer (3). They also noted symptomatic improvements and corresponding decreases in prostatic acid phosphatase and prostatic alkaline phosphatase in the serum of such patients and thus established the role of hormone therapy in the management of metastatic prostate cancer. Castration-resistant prostate cancer (CRPC) is inevitable after a variable period of response to ADT. Strategies to delay CRPC are an active area of investigation with promising novel agents emerging as candidates for trials in first-line therapy of metastatic cancer.

A review of the hypothalamic-pituitary-gonadal axis as it pertains to the regulation of the synthesis of androgens from the testes, the major site of androgen production, is in order (4). GnRH (gonadotropin-releasing hormone) secreted by the hypothalamus is transported to the anterior pituitary by the hypophyseal portal system where GnRH acts on specific GnRH receptors to stimulate secretion of LH (luteinizing hormone) and FSH (follicle stimulating hormone). GnRH is therefore also known as LHRH (luteinizing hormone-releasing hormone). LH and FSH are then secreted into the systemic circulation. In the testes, LH stimulates testosterone production by Leydig cells, while FSH causes inhibin (important in feedback control of FSH production) secretion from Sertoli cells. The HPG axis is regulated by negative feedback loops of testosterone and inhibin on the anterior pituitary and hypothalamus.

Testosterone is converted by the enzyme 5-alpha reductase in tissues to a more potent androgen, dihydrotestosterone (DHT). DHT has greater affinity for the androgen receptor (AR) in target organs than testosterone itself. AR is kept inactive in the cytoplasm of normal prostate cells by heat shock proteins (HSP). The binding of DHT to AR frees it from HSP. The DHT-AR complex translocates to the nucleus and forms a dimer with another DHT-AR complex. The dimerized molecule binds to androgen-response elements on target genes. AR thus modulates expression of several downstream genes (including PSA) and pathways in the prostate cell that promote survival and growth.

Orchiectomy is a relatively simple surgical procedure with the resultant advantages of low cost and immediacy of attaining castrate testosterone levels (5). These merits are however counterbalanced by the psychological impact of the surgery and the irreversibility particularly given the younger age at diagnosis in the PSA era and the options for intermittent therapy. In a questionnaire-based study, only 24% of eligible men opted surgical castration compared to medical alternatives (6). Advancements such as subcapsular orchiectomy and testicular prostheses have not appeared to change this scenario.

Estrogens: The earliest means of medical castration were synthetic estrogens that blocked the hypothalamic-pituitarygonadal axis via feedback inhibition. The Veterans Administrative groups conducted the initial trials in metastatic prostate cancer with the nonsteroidal estrogen, diethylstilbestrol (DES) (7). DES causes peripheral blockade of the AR and direct cytotoxic effects on the cancer cells in addition to blocking LHRH release. Prostate cancer-specific survival was equivalent in patients receiving 1 mg DES and 5 mg DES. However, the 5-mg dose was associated with markedly increased cardiovascular deaths. Cardiovascular toxicity was particularly prominent at daily doses >3 mg that are needed for reliable castrate levels of testosterone and in older men with cardiovascular risk factors. Conjugated estrogens, ethinyl estradiol, and medroxyprogesterone acetate have all been tested and are occasionally used in the second-line setting but seldom in the initial management of prostate cancer (8).

LHRH (GnRH) agonists (9) bind to GnRH receptors in the anterior pituitary, causing an initial transient increase in LH release and rise in testosterone levels. Subsequently, GnRH receptors are downregulated in the second week resulting in a decline in testosterone levels to the castrate range over the next 3 to 4 weeks. LHRH agonists include leuprolide, goserelin, buserelin, and triptorelin. Depot formulations of leuprolide are available as intramuscular injections every 1 month (7.5 mg), 3 months (22.5 mg), 4 months (30 mg), or 6 months (45 mg) though an even longer-acting formulation, leuprolide acetate, is available as a 1-year subcutaneous implant (72 mg). Goserelin is given as a 3.6-mg subcutaneous injection every month or 10.8-mg subcutaneous (sc) injections every 3 months. Buserelin is available as nasal spray or subcutaneous injections given every 8 hours. Triptorelin is available as 3.75-mg intramuscular injection every month and 11.5-mg injection every 3 months. The different LHRH formulations are considered equivalent in efficacy. In the initial studies, leuprolide was administered as daily subcutaneous injections of 1 mg and compared to oral DES dosed at 3 mg daily in the first-line treatment of metastatic prostate cancer (10). 86% of patients in the leuprolide group and 85% of patients treated with DES achieved objective responses. Moreover, side effects including major cardiac events occurred more frequently in the DES patients, although cardiovascular complications were not reached statistically higher after DES. Long-term administration of leuprolide either as daily doses or as long-acting depot preparations achieves comparable testosterone suppression and objective responses (11,12). Similar efficacy of other LHRH agonists such as goserelin has been demonstrated in clinical trials (13,14) and in meta-analyses (15).Other than the injection-related issues, side effects from LHRH agonists are from testosterone deprivation and include impotence, loss of libido, metabolic side effects that are increasingly being recognized such as dyslipidemia including reduction in HDL-cholesterol with elevation in cardiovascular risk, changes in body habitus due to altered fat distribution, gynecomastia both from breast tissue increase and from increased fat deposition on the chest, hair loss, loss of muscle mass and bone loss, hot flashes, and night sweats (Table 14B.1). The obvious advantages of LHRH agonists are ability to achieve effective castrate levels of testosterone, reversibility, and ease of administration though the cumulative economic costs over a lifetime are substantial. A unique phenomenon stemming from their LHRH agonist activity is the initial and transient upsurge in LHRH activity and consequent rise in testosterone effect. In situations where such a “flare response” could have deleterious effects clinically such as cord compression from substantial spinal metastases or ureteric or urethral outflow obstruction from enlarged lymph nodes or prostate, initiation of an AR blocker before an LHRH agonist is important. No studies have been done to determine the duration of the lead in therapy; however, it is reasonable to consider a 7 to 10 day therapy with an antiandrogen alone and continuing for at least 2 weeks after the LHRH agonist has been started.

LHRH antagonists are not associated with the flare response. Abarelix was withdrawn from the market in 2005 following reports of acute onset allergic reactions
including hypotension and syncope following intramuscular administration. Degarelix has recently been approved (240 mg sc, given as 2 injections of 120 mg each followed by maintenance of 80 mg sc every 28 days) in advanced prostate cancer based on a randomized study comparing degarelix to monthly leuprolide for a duration of 1 year (16). The primary end point of percentage of patients who achieved castrate levels of testosterone maintained for 1 year was similar in both groups. Declines in testosterone to castrate range were faster (>95% of patients at 3 days) and PSA levels were lower (at 14 and 28 days) with degarelix. There was a higher incidence of injection site reactions with degarelix but no systemic allergic or anaphylactic reactions.

Antiandrogen molecules were developed in the late 1960s to specifically block the DHT receptor (17). They block the effect of gonadal and extragonadal androgens (adrenal) at the target tissue level. As a result, LH and testosterone levels increase and sexual function and libido are preserved more commonly than with ADT. Based on molecular structure, antiandrogens are classified as steroidal or nonsteroidal. Nonsteroidal antiandrogens include flutamide, bicalutamide, and nilutamide. Flutamide has a short half-life of 5 hours and is usually dosed at 250 mg every 8 hours. Once a day dosing of flutamide at 500 mg has been demonstrated to be equally efficacious as thrice daily dosing. Half-lives of bicalutamide and nilutamide are longer (˜7 and 2 days) and are dosed once a day (bicalutamide at usual daily dose of 50 mg and nilutamide at 300 mg for 30 days followed by 150 mg daily). Side effects of nonsteroidal antiandrogens (Table 14B.2) as a class include nausea, vomiting, diarrhea, breast growth or enlargement with or without tenderness (due to higher testosterone levels and with incidence up to 74% associated with bicalutamide), and potentially fatal hepatotoxicity (most with flutamide and least with bicalutamide but requires periodic liver function tests with use of any antiandrogen). Serum transaminase levels should be measured prior to starting treatment, monthly for the first 4 months of therapy, and periodically thereafter. Flutamide should not be started or continued in patients whose ALT values exceed twice the upper limit of normal. Nilutamide has, in addition, specific toxicities of decreased dark adaptation, rare but serious interstitial pneumonitis and alcohol intolerance. In terms of efficacy, the nonsteroidal antiandrogens are considered equivalent. For example, a four-armed randomized trial of an antiandrogen (bicalutamide or flutamide) in combination with an LHRH-agonist (leuprolide or goserelin) showed equivalent overall survival (18).

Steroidal antiandrogens (not generally used in North America) such as cyproterone acetate, megestrol acetate, and medroxyprogesterone acetate also have the progestational adverse effects of loss of libido and impotence and the advantage of reducing hot flashes from ADT. Several randomized trials used cyproterone acetate (not approved in USA) as part of combined androgen deprivation (CAD). A meta-analysis (PCTCG) suggested a trend toward decreased overall survival with cyproterone acetate due to a greater number of nonprostate cancer deaths (19).

Several large meta-analyses have been conducted to address the issue (19,44,45,46,47). The Prostate Cancer Trialists’ Collaborative Group (PCTCG) published a meta-analysis using individual patient data from 8,275 men from 27 randomized trials in an intention-to-treat fashion. The data included 98% of all patients registered in randomized trials of prostate cancer to compare CAD versus monotherapy to that point in time (19). Of the 27 included trials, 8 studied nilutamide, 12 flutamide, and 7 cyproterone acetate. Eighty-eight percent of the patients had metastatic disease, fifty percent were older than 70 years, and average follow-up was 5 years. At the time of analysis, 70% had died (80% of deaths attributable to prostate cancer). Five-year survival was 25.4% with CAD versus 23.6% with monotherapy alone (p = 0.11). When the cyproterone acetate trials (20% of the patients) were excluded, there was an overall survival benefit to CAD (nilutamide or flutamide) at 5 years of 27.6% versus 24.7% (p = 0.005). For cyproterone acetate trials, overall survival was slightly but significantly worse with CAD with 5-year overall survival 15.4% versus 18.1% (p = 0.04). Non-prostate cancer deaths (did not reach statistical significance) seemed to explain the worse outcome with cyproterone acetate. The authors concluded that there was probably a 2% to 3% overall survival benefit to CAD (excluding cyproterone acetate) although the true benefit statistically could range from 0% to 5% at 5 years. Criticism of the metaanalysis included the fact that control groups in some of the included trials used short-term antiandrogens to prevent the flare response.

A Cochrane review analyzed 6,320 patients from 20 trials comparing combined therapy to monotherapy (45). Odds ratio (OR) for overall survival with combined androgen blockade was 1.29 (95% CI = 1.11-1.50; n = 3,550 from seven trials) at 5 years. The authors estimated a 5% improvement in 5-year survival rate (30% vs. 25%) with CAD. PFS was improved at 1 year (OR = 1.38; 95% CI = 1.15-1.67; n = 2278 from seven trials). Cancer-specific survival was improved at 5 years (OR = 1.58; 95% CI = 1.05-2.37; n = 781 from two trials). However, when analysis was limited to studies identified as being of high quality, the OR was not significant at any follow-up interval (1, 2, or 5 years).

The individual trials and meta-analyses have been criticized for several reasons. Most of the studies included were early incomplete looks at the data or contained too few patients to reach statistical significance on their own. The lack of statistical power in several of the reported “negative trials” leading to incomplete information to the reader and consequent misinterpretation of “inconclusive trials” as “negative trials” has been highlighted (48). The trials also included various antiandrogens and modalities of gonadal suppression. In particular, few of the trials evaluated bicalutamide as the antiandrogen.

While it is likely that a small overall survival benefit indeed exists to CAD (2%-3% at 5 years) but is not certain, data on improvements in other outcomes such as TTP and PFS are not conclusive either. Data on quality of life have been assessed in some of the trials (including the Japanese trial and INT-0105) with conflicting results. The Japanese trial showed a trend toward improved QoL parameters while INT-0105 showed worse QoL in the CAD group.

Although there is no conclusive evidence-based consensus about the advantage of CAD as first-line therapy for metastatic prostate cancer over castration alone, treatment decisions have to be individualized, as always in medicine, weighing potential benefit with risk of increased toxicity.

Antiandrogens as monotherapy for initial hormonal therapy have the theoretical benefit of avoiding androgen deprivation and consequent side effects while possibly controlling disease from metastatic prostate cancer. Direct inhibition of the DHT receptor would allow for some hormonal effects
of circulating testosterone such as the maintenance of sexual interest and function. Initial comparisons of low-dose bicalutamide (50 mg daily) with castration showed inferior response and survival outcomes with antiandrogen monotherapy but improved QoL as expected (49,50,51,52). Large randomized trials with higher doses of bicalutamide have provided mixed results in patients with metastatic and locally advanced T3/4 disease (53,54). Tyrrell et al. used optimal-dose estimation arms and demonstrated improved PSA response at a dose of 150 mg (vs. 100 mg); thus, the higher dose was used for the remainder of the study. With over 800 patients enrolled with metastatic disease (of the total of 1,453 patients in the study), the study was stopped due to a 6-week median survival benefit to castration at a median follow-up period of approximately 100 weeks. Thus, bicalutamide at 150 mg daily was associated with poorer survival (HR of 1.30 for time to death). In patients with symptomatic metastases, bicalutamide was associated with a statistically significant improvement in subjective response (70%) compared with castration (58%). Analysis of a QoL questionnaire demonstrated an advantage for bicalutamide in sexual interest and physical capacity. Bicalutamide had a substantially lower incidence of hot flushes compared to castration (6%-13% compared with 39%-44%), but significantly more breast tenderness that did not lead to withdrawal from therapy. The authors concluded that bicalutamide monotherapy could be an option in a well-counseled patient with metastatic disease who wishes to avoid certain side effects of castration. A secondary analysis of the data suggested that patients with PSA <400 ng/mL or less than six metastatic sites had equivalent outcomes with castration or antiandrogen monotherapy (55). A larger total number of ARs in patients with extensive disease might explain the worse outcomes in that subset. Two smaller underpowered trials in Europe also evaluated bicalutamide at 150 mg against CAD. No significant outcome differences were noted though these trials are limited by small sample size and inclusion of nonmetastatic patients in one trial (53,56).

Conclusions regarding flutamide as monotherapy are somewhat more difficult to make. Flutamide has been randomized against CAD, orchiectomy, and DES (57,58,59). Survival time in the DES study with 92 patients only was much shorter in the flutamide group (43.2 months vs. 28.5 months) while survival was equivalent in the other two studies (against CAD with 319 patients and against orchiectomy in 104 patients). Thorpe et al. examined CPA monotherapy in a three-armed study of 525 patients comparing monotherapy with CPA versus goserelin versus combination therapy with both (40). CPA performed slightly worse than LHRH-agonist in time to progression. Two other trials with a total of 175 patients have examined CPA against DES (3 mg) and orchiectomy and found no difference in overall survival at up to 5 years (60,61). The combination of an antiandrogen with a 5-alpha reductase inhibitor has been tested but with very few metastatic patients among those enrolled (62).

ADT has well-accepted benefits in palliating painful bone metastases and urinary tract obstruction in patients with metastatic disease. However, ADT is not considered curative, and the response to therapy is temporary, with most patients progressing within 2 years (63). Since the introduction of ADT, controversy has existed over optimum timing. Many have advocated initiating treatment at the time of diagnosis with the aim of delaying disease progression and possibly prolonging survival. Others have argued that survival is not prolonged and that the treatment may be deferred until symptoms develop.

Some experimental evidence supports the early treatment of prostate cancer (not necessarily in the metastatic setting). Using the Dunning R3327 rat prostatic adenocarcinoma model, studies showed increased regression and survival after castration if the tumor volume was relatively small as opposed to when the tumor was allowed to grow larger (64). Using the same model, researchers were able to show that the mechanism responsible for the emergence of androgen resistance was the intrinsic heterogeneity of the tumor and not castration (65). They also demonstrated that poorly differentiated, fastgrowing androgen-sensitive tumors respond less to androgen withdrawal later in their lifespan.

Some nonrandomized data suggest a potential benefit from earlier therapy (66,67). On the other hand, the Veterans Administration Cooperative Urological Research Group study demonstrated that delaying hormonal therapy did not compromise overall survival (5,7,68,69) and demonstrated that many of the patients died of other causes while only 41% died from prostate cancer.

The Medical Research Council in the UK randomized 938 patients (934 patients had data available for analysis) with locally advanced prostate cancer or asymptomatic metastasis to either immediate treatment (orchiectomy or LHRH analog) or treatment deferred until symptoms occurred (70). Eighty patients died before treatment commenced (29 from prostate cancer and 51 from other causes). Progression from M0 to M1 disease (p < 0.001) and development of metastatic pain occurred more rapidly in deferred patients. One-hundred and forty-one deferred patients needed transurethral resection for local progression compared with sixty-five treated immediately (p < 0.001). Pathological fracture, spinal cord compression, ureteric obstruction, and development of extraskeletal metastases were twice as common in deferred patients. Three-hundred and sixty-one patients died in the deferred arm compared with three-hundred and twenty-eight in the immediate arm (p = 0.02). Sixty-seven of all deaths were from prostate cancer. Two-hundred and fifty-seven and two-hundred and three deaths were from prostate cancer in the deferred versus immediate treatment arms, respectively (p = 0.001). The difference was seen largely in locally advanced nonmetastatic patients, with 119 and 81 deaths from prostate cancer, respectively (p < 0.001).

The Cochrane databases compiled a meta-analysis of data from four randomized controlled trials in 2001 (71). The trials included were VACURG-I, VACURG-II, the MRC trial, and an ECOG trial studying early versus delayed androgen deprivation as adjuvant therapy in patients with pathologically involved lymph nodes after radical prostatectomy and lymph node dissection. The authors concluded that early hormonal therapy definitively reduces disease progression and the subsequent complications with a small improvement (6%) in overall survival at 10 years. Of note, 97% of the patients included in the 10-year analysis were extracted from the VACURG-I data, which makes this result sensitive to the weaknesses of that trial.

The clearest indication for immediate therapy for patients with new M1 disease is in the context of symptoms. Data from the MRC trial showing that early hormonal therapy can lower risks of significant complications such as cord compression and urinary tract obstruction and other data discussed above collectively provide the clinical rationale to treat early in the context of the commonly encountered scenario of the patient with asymptomatic metastasis.

The better understanding of the biology of progression from hormonal dependence to castration resistance has led to a variety of novel agents and experimental approaches intended to enhance the effectiveness of therapy for patients with metastatic disease. Several such approaches/agents, discussed below, are in different phases of development at this time.

ADT is the standard first-line therapy for metastatic prostate cancer. The recently developed novel potent inhibitors of enzymes involved in androgen synthesis and newer antiandrogens hold great promise as potential components of future first-line therapy in metastatic prostate cancer. Improving understanding of the molecular and cellular biology of development of androgen resistance should permit strategies to prolong duration of sensitivity to androgen deprivation and prevention of castration androgen resistance. Ongoing trials will demonstrate any potential value in adding chemotherapy to hormonal therapy in hormone sensitive metastatic disease while novel biomarkers in development such as circulating tumor cells or PSA will assist in individualizing therapy with the goal of maximizing benefit while minimizing toxicities. Critical to progress is the timely implementation and accrual to clinical trials in this disease.