|Home | About | Journals | Submit | Contact Us | Français|
Prostate cancer and its management have been intensely debated for years. Recommendations range from ardent support for active screening and immediate treatment to resolute avoidance of screening and active surveillance. There is a growing body of level I evidence establishing a clear survival advantage for treatment of subsets of patients with clinically localized prostate cancer. This chapter presents a review of these randomized controlled trials. We argue that an understanding of this literature is relevant not only to those considering active surveillance but also to those evaluating the merits of screening. In addition, a number of important evidence-based conclusions concerning what should and should not be done can be gleaned from these trials.
The argument for active surveillance (AS) in men with localized prostate cancer is critically dependent on whether a delay in the initiation of treatment adversely affects outcome and the degree to which treatment is proven to be safe and effective at prolonging survival. If treatment is not effective, then delays may be irrelevant. If on the other extreme, treatments are successful regardless of the extent of disease, then delays in treatment may also be irrelevant. Unfortunately, although moderately effective, the available treatments are frequently not curative in patients with advanced disease, and thus delays may have an adverse impact on outcome. This review concisely summarizes the data from the randomized treatment trials for localized prostate cancer. When appropriate and possible, we also comment on the number needed to treat (NNT) to render benefit (1). Our summaries focus on the degree to which various types of treatments are effective as well as on the subsets of patients who benefit. We not only demonstrate that there is a large and growing body of level I evidence, supporting the notion that there are populations of men with prostate cancer who clearly benefit from specific treatments, but also that certain treatments are ineffective and should be avoided. The modest impact of treatment on survival shown in some of these trials may provide support for the rationale of AS (particularly in the subsets for whom no benefit is shown). This conclusion can be explained by the fact that competing causes of death tend to attenuate benefits of treatment in patients with moderate- to high-risk disease and may overwhelm the potential beneficial impact of treatment in patients with very low-risk disease.
Using the outcome data from all of the randomized control trials (RCTs) (primarily phase III) published at the time of writing of this article that address clinically localized prostate cancer, we argue that this evidence may have relevance not only to the issue of AS but also to screening. Our rationale for including all of these trials comes from the third question posed by the United States Preventive Services Task Force (USPSTF), which asks, “What are the benefits of treatment of early-stage or screening-detected prostate cancer?” On the surface, this question seems simple enough; however, on deeper inspection, many uncertainties arise. First, the question does not define “early-stage” prostate cancer (ie, localized vs low-risk vs high-risk), how much of an improvement in survival might be expected, and when these benefits should be expected. Second, the question does not address whether some subsets of screened patients may benefit from treatment. To answer this question, the USPSTF chose to include only two RCTs for evaluating the impact of treatment on localized disease. For a more comprehensive review, we choose to include 50 trials (2) in our analysis. We argue that the body of literature included in this review may provide important information as to who, when, and how much patients can be expected to benefit from various types of treatment and early detection. Although many of the trials summarized below included patients who are not relevant to the issue of screening and AS, we argue that these studies may provide useful insight into this issue nonetheless.
Table 1 summarizes trials involving radical prostatectomy (RP) with or without the addition of androgen deprivation therapy (ADT) or RP compared with “watchful waiting” (WW). Although some say that WW is not to be compared with AS, there is no clear consensus as to what constitutes WW and AS; thus, useful information may still be gleaned from their collective evaluation. The only major trial comparing WW with RP showed that with a median of 12.8 years, 166 men in the RP group and 201 in the WW group died of any cause (P = .007) (3). Of note, 55 and 81 of the deaths were attributed to prostate cancer to the RP and WW groups, respectively. The survival benefit observed appeared to be confined to men younger than age 65 years. The number needed to treat to avert one death was 15 overall and 7 for men younger than age 65 years. The remaining studies demonstrated that there is no advantage to adding ADT except as an adjuvant therapy in men with positive lymph nodes (11).
Table 2 summarizes the three major trials addressing adjuvant postoperative radiotherapy (RT). The relevance of these studies to screening and AS may seem questionable because, for example, only 302 of the 431 patients on the Thompson trial had a preoperative prostate-specific antigen (PSA) and approximately 50% of these men had a PSA < 10ng/mL. However, despite screening, 25%–30% of men undergoing a radical prostatectomy will still have evidence of extracapsular extension or positive margins and thus will be candidates for adjuvant RT (15). The data from the Thompson trial provide information regarding both the sample size and timeline required to document a survival benefit with adjuvant therapy. It took more than 10 years to show the benefits of postoperative RT; thus, it is likely that it would take an even longer follow-up period to demonstrate the impact of treatment in the remaining men with organ-confined disease. Furthermore, nonscreened men who are diagnosed with more advanced disease might be expected to have a prolonged survival due to early aggressive postoperative interventions further complicating the short-term analysis of men offered deferred management.
Table 3 addresses the role of primary external beam radiotherapy (EBRT) compared with other modalities (eg, radical prostatectomy or cryoablation with higher doses). Because of their small size and short follow-up at first review, it may appear that these studies would be of little relevance to the screening or AS controversies. However, their shortcomings tell us “what not to do” and explain why we have not resolved the uncertainty concerning the relative effectiveness of these treatment options. The point here is that we should avoid conducting underpowered studies when assessing interventions that are likely to have relatively small differences in effectiveness.
Table 4 addresses the impact of different types and doses of radiation used in the treatment of prostate cancer. Several studies did not provide sufficient details to allow an accurate assessment of the number of low-risk patients included (23, 24, 29, 32). Some studies specifically excluded low-risk patients (26, 28, 30, 31, 34). In several studies, approximately 20% of the patients could be considered low-risk patients (20–22). In other studies, up to 40%–50% or more would generally be considered low-risk patients (25, 27, 33). Ultimately, although the Table 4 studies consistently show a reduction in the PSA detected recurrence rates with higher doses of radiation, there was no evidence that survival was improved. These findings should discourage investigators from expecting to detect improvements in survival between patients with early prostate cancer who were treated with relatively modest doses of radiation.
Table 5 summarizes phase III trials performed by the Radiation Therapy Oncology Group (RTOG) using radiation with or without ADT, and Table 6 lists major non-RTOG phase III prostate cancer radiation trials with or without ADT. These studies when taken together provide a large body of level I evidence for the addition of ADT to EBRT in selected patients with intermediate- to high-risk prostate cancer. Their relevance to early-stage screening is that they provide data from which relative risk estimates, sample size estimates, and timelines can be made.
Table 7 addresses the role of primary ADT with or without EBRT. Again, although the relevance of these data to AS or screening is not immediately obvious, there are lessons learned from these trials. The first two and the last two trials included only patients with locally advanced disease. The third trial listed included a subset of 1627 men with early-localized disease, who were randomized to placebo or 150mg of bicalutamide. When taken together, these studies demonstrated that treatment with ADT was beneficial in men with locally advanced disease, but such treatment resulted in a lower survival in men on WW (54). This study highlights that even though an intervention is beneficial to men with locally advanced disease, it does not necessarily mean that it will be beneficial in men with early disease. In addition, these data reinforce the notion that PSA alone is not an adequate endpoint and that it is important that studies be adequately powered.
Based on these data, the following conclusions concerning the impact of various treatments on the survival of men with clinically localized disease can be drawn:
Despite the considerable progress made due to the trials summarized above, there are still many unanswered questions. A partial list of completed, ongoing, or closed trials addressing some of the remaining questions among men treated for clinically localized disease is provided below:
Although many of the studies included in the review contain participants who would not have been candidates for AS, there still appear to be lessons that can be learned from these trials. In order to understand the risk associated with AS, it is important to understand the potential benefits of treatment. In order to understand the benefits of treatment, it is critical to understand the magnitude and timeline in which benefits of treatment might be expected for men with low-risk disease. There are, however, very limited data available on which to make such estimates. In addition, some men who appear to have limited disease when AS is initiated will, in fact, have higher-risk disease later, which will require treatment. Neither the trials reviewed nor the cohort studies chosen by the USPSTF are robust enough to provide such data. For example, in the manuscript published by the USPSTF, they also included eight cohort studies including as few as 316 men with or without prostatectomy and five cohort studies including as few as 334 men treated with or without RT, with follow-up as short as 4 years (2). From these cohort studies, there appeared to be a decrease in all-cause mortality with both treatment approaches compared with WW. However, the RCTs (which they did not include) suggest that the magnitude of the benefits for treatment might have been greater had postoperative RT been routinely added to patients with adverse pathological features prostatectomy (National Cancer Institute National Clinical Trial Network’s SWOG) and ADT been added to RT in subsets of patients with T1-2 disease diagnosed in the PSA-era (eg, RTOG 9408). These sorts of details might have helped inform the discussion concerning Question #3 from their report asking, “What are the benefits of treatment of early-stage or screening-detected prostate cancer?”
This perspective is not meant as a criticism of the USPSTF’s position on prostate cancer screening nor is it meant to argue against the potential value of AS, but rather to set realistic expectations using existing data from phase III treatment trials. Our premise is that understanding data provided by the numerous randomized clinical trials that have demonstrated the benefits of therapy in men with localized prostate cancer, we can rationally find support for determining in which patients AS might likely be safest. Similarly, the treatment data can help us model the distribution of advanced disease in the population that would need to be exceeded in order to maximize the benefits of treatment after screening. Some might argue that we already know that we need huge sample sizes and very long timelines to assess the benefits of screening and AS and that we already know that AS entails some risk of missing the “window of curability.” We argue that the data from the randomized control treatment trials can actually allow us to answer these questions concerning which subsets of patients are actually likely to benefit the most from which type of treatment and who might be best served by AS. For example, it is well known that roughly 10%–15% of men with newly diagnosed prostate cancer have high-grade disease (Gleason scores of 8–10), and they represent the patients for whom the benefits of treatment have most consistently been shown in a number of randomized trials (eg, RTOG 8531, RTOG 9202, EORTC) (summarized in Tables 5 and and6).6). Given this fact, such patients should probably be excluded from studies assessing the merits of AS. Thus, we provide a cautionary point to the USPSTF that by limiting the assessment of the benefits of treatment to only two trials, they limit their ability to accurately address the complicated issues related to the timeline of treatment delivery and thus the ability to assess the merits of delayed treatment and screening.
Despite the studies completed and pending completion, it is also clear that important gaps in our scientific and medical knowledge will remain for years to come. Perhaps the two most promising areas involve advances in imaging that will help us assess the true extent and distribution of disease and the identification of biomarkers that hold promise for helping select the most appropriate level of therapy for an individual patient (59,60).
Unfortunately, because of our reimbursement structure and nature of the criteria required to secure Food and Drug Administration clearance, the prospects for imaging advances appear dire. Despite a large body of literature published in Europe and elsewhere demonstrating the value of positron emission tomography using agents such as acetate and choline as well as the promising results with magnetic resonance imaging using dextran-coated nanoparticles, neither of these agents/modalities are currently available for routine reimbursement in the United States (61–70).
Although the prospects for the development of prognostic biomarkers appear to be somewhat less challenging than the development of imaging agents, the results to date have been less promising. Despite many studies completed to date, none appear “ready for prime-time” (60). If these biomarkers could be proven to predict disease outcome and guide treatment, they would hold great promise for selecting how, when, and if patients may best be treated. Acquiring this understanding could consequently lower the cost and morbidity of treatment for clinically localized prostate cancer.
The data found in these prospective randomized trials provide clear evidence that certain types of patients benefit from certain types of treatment. The data also suggest that the follow-up time and the sample size required to show the benefits from treatment are inversely proportional to the risk group. In other words, low-risk patients require very long follow-up, whereas high-risk patients require a shorter follow-up and smaller sample size. Based on this body of treatment literature, we should expect to be able to better identify those subsets of patients for whom less might be more.