Natural History of Metastatic Prostate Cancer
Occult Metastases
An important variable in “early hormonal” therapy is the definition of “early.” Since the intent is to treat systemic disease effectively to prolong survival and reduce morbidity, it could be argued that the start of adjuvant hormonal therapy in carefully selected high-risk patients will be the earliest point in time for hormonal therapy. This position could be justified by the current data indicating that despite a steady decline in the incidence of newly diagnosed metastatic disease and microscopic lymph node metastases (from 20% in 1972 to 1988 to 3.4% in 1988 to 1991), the risk of extraprostatic disease in patients with clinically organ-confined prostate cancer is still significant (40 to 60%). Depending on the local/regional extent and histologic grade of the disease, approximately 50% of patients with prostate cancer will progress despite definitive local therapy, verifying the presence of early but occult systemic disease.
Prostate-Specific Antigen Relapse and Metastatic Disease
Serial prostate-specific antigen measurements following definitive local therapy have proven that the rate of permanent prostate cancer eradication is not as high as previously believed. Current data suggest that in most cases a rising prostate-specific antigen level following local therapy heralds clinical progression. The number of patients with a rising prostate-specific antigen level after local therapy is increasing; the time of their failure represents the second point in time of relevance to early hormonal therapy. Earlier data suggested that lack of prostate-specific antigen normalization following local irradiation correlated with subsequent prostate-specific antigen failure and that up to 44% of patients with prostate-specific antigen failure would develop systemic metastases within 3 years. Subsequent reports indicated that in patients treated with local irradiation, a post-therapy prostate-specific antigen doubling time of less than 6 months is highly associated with systemic relapse. On average it takes 4.5 times prostate-specific antigen doubling time to clinical recurrence. In 100 patients with a rising prostate-specific antigen level after treatment, Pollack et al. calculated that the median time from prostate-specific antigen elevation to clinical relapse was 40 months. One report indicated that the prostate-specific antigen doubling time differed in the first 2 years following surgery or irradiation but became similar after 2 years. The authors concluded that the growth rate of recurrent cancer after radiation and radical prostatectomy is similar.
Recent observations from Lee et al. shed more light on the natural history of this disease. Overall, the estimated rate of distant relapse is 47% at 5 years following prostate-specific antigen failure. Predictors of distant metastases in a multivariate analysis were time to prostate-specific antigen elevation (0 to 11.9 months versus > 12 months, p < .001) and post-treatment prostate-specific antigen doubling time (0 to 11.9 months versus > 12 months, p = .01). Seventy-five percent of patients with an increasing prostate-specific antigen profile within 12 months of treatment developed distant metastases within 5 years from the time of prostate-specific antigen elevation as opposed to 25% of those having a prostate-specific antigen rise at least 12 months after treatment (p < .001). Among surgically treated patients, it has been reported that 50 to 100% of patients whose prostate-specific antigen becomes detectable within the first year following radical prostatectomy will progress rapidly, with either local recurrence or distant metastasis.
The effects of prostate-specific antigen-induced stage migration have also been noted in patients with newly diagnosed metastatic prostate cancer. This is reflected by a decline in the percentage of patients presenting with metastatic disease and is illustrated by data from the Metropolitan Detroit Cancer Surveillance System (member of the National Cancer Institute [NCI] SEER Program). Trends in stage-specific incidence were evaluated during a 22-year period (1973 to 1994). Incidence rates for stage of local and regional prostate cancer reached a maximum in 1992 and 1993, and distant stage disease has declined steadily since 1989 (p < .001). Despite the decline in its incidence and the perceived shift toward a less severe form of disease at presentation, metastatic prostate cancer remains incurable despite the best available therapy.
Biology of Hormonally Treated Prostate Cancer
Prostate cancer is a slowly growing tumor with low cellular turnover. There is a trend toward increased heterogeneity and genetic instability as tumor volume increases, reflected by increased incidence of aneuploidy and molecular changes. The frequency of aneuploidy increases with stage. The aggressive nature of nondiploid tumors is illustrated by the higher mortality associated with them compared to diploid tumors, despite appropriate therapy There have been a variety of chromosomal alterations and loss/mutations of tumor suppressor functions associated with the metastatic and androgen independent phenotype.
Reviewing the issue of early hormonal therapy in 1991, Kozolowski et al. stated that “there is no experimental or clinical evidence that early androgen ablation is biologically deleterious.” Although early but limited exposure may not be deleterious, the long-term sequelae of early and life-long hormonal therapy are not well known.
The androgen dependence of prostate cancer is well established, as is the inevitable and irreversible progression to androgen independence. The latter heralds the terminal step in the natural history of prostate cancer. Clear understanding of the mechanism of progression is crucial not only for the development of better systemic therapy but also for clinicians considering early, long-term hormonal therapy. The mechanisms characterizing progression from localized to metastatic to androgen-independent prostate cancer are unclear. Of increasing relevance to the issue of early hormonal therapy are the genetic alterations underlying progression to androgen independence. Two theories have been proposed to explain the phenomenon: the adaptation model and the clonal selection model. Central to both is an androgen-deprived environment that may foster development or expansion of the androgen-independent phenotype.
Experimental observations in the Dunning rat prostatic adenocarcinoma model may support treatment by early hormonal therapy. Tumor regression and increased survival were more achievable if the tumor volume was relatively small at the time of castration. On the other hand, studies using poorly differentiated, fast growing tumors demonstrated a significant reduction in androgen sensitivity and lack of response to castration in the later transplant generations. These data suggest the potential for mutation selection and spontaneous progression to androgen independence and the possible role for early castration in preventing progression.
Results in the Dunning rat model must be contrasted with the observations of Bruchovsky et al. in their use of the Shionogi androgen-dependent mammary carcinoma. In this tumor model, androgen withdrawal resulted in the elimination of differentiated cells and a decrease in the number of tumorogenic stem cells. Despite continued androgen deprivation, tumor progression ensued and was associated with a 20-fold increase in the proportion of stem cells as well as a massive 500-fold increase in androgen-independent stem cells. Progression was delayed but not prevented by periods of androgen exposure and deprivation. The authors suggested that progression in this model may be due to the activation and enhanced expression of androgen-repressed genes coding for autocrine and paracrine growth factors that replace androgens in maintaining the viability of tumor cells.
Progression to androgen independence has been clinically and experimentally associated with multiple changes at basic levels of cellular control, including increased expression of bcl-2 and C-myc, TRPM-2, altered P53 status, and androgen receptor-gene amplification and mutations. The bcl-2, a negative regulator of apoptosis, may form a hetrodimer with a related protein bax, a positive regulator of programmed cell death. The ratio of bcl-2 and bax determines cell survival or death. Experimentally, bcl-2 overexpression protected human prostate cancer cells from apoptosis in vitro and conferred resistance to androgen deprivation in vivo. Further, expression of bcl-2 has been linked to the androgen-independent property of prostate cancer recurrences. Enhanced expression of C-myc protooncogenes is observed following castration; it is involved in the regulation of mitosis and apoptosis. In the presence of critical growth stimuli such as overexpression of bcl-2, C-myc capacity to induce apoptosis may be abrogated. Expression of C-myc oncogene is demonstrated in both LNCap cells (hormone sensitive) and DU 145 cells which are not dependent on hormones for survival. Elevated levels of C-myc are often associated with cell proliferation and metastases. The tumor suppressor gene Mxil located at chromosome 10q22.1 is related to C-myc activity; less of this suppressor may lead to C-myc activation. These observations may in part help explain the clinical experience with hormone therapy. They should raise concerns about the potential deleterious effects of long-term therapy.
The Case for and against Early Hormonal Therapy
There has been a debate regarding the timing of androgen deprivation since it was introduced. At issue is whether early androgen deprivation improves survival. This has not been resolved as no prospectively randomized studies are available to answer the question stage by stage. Support for early hormone therapy was originally based on several reports that invariably involved comparisons with historic controls.” Historic data on untreated metastatic prostate cancer reflect a very high mortality, with a median survival of less than 10 months. Arguments for early hormonal therapy based on comparisons with historic data are flawed for several reasons, including the retrospective nature of these studies, lack of randomization, changing patient characteristics, impact of progress in medical care on comorbidities, and on the diagnosis and management of prostate cancer with lead-time bias offered by early diagnosis/improved staging.