 |
Additional Therapy for High-risk Prostate Cancer Treated With Surgery: What is the Evidence?
Scott C Morgan; David P Dearnaley
Authors and Disclosures
Posted: 08/06/2009; Expert Rev Anticancer Ther. 2009;9(7):939-951. © 2009 Expert Reviews Ltd.
Abstract and Introduction
Abstract
Single-modality approaches to the treatment of high-risk prostate cancer, whether radical prostatectomy or external-beam radiotherapy, have yielded disappointing results. Treatment intensification has, thus, been the subject of considerable research activity in recent years. This review will discuss the evidence for neoadjuvant and adjuvant treatment approaches when surgery is chosen as the definitive therapy for high-risk prostate cancer. Particular emphasis will be placed on the randomized trials, both completed and in progress. Trials investigating adjuvant radiotherapy, androgen-deprivation therapy and chemotherapy will each be discussed in turn. Among these, only adjuvant radiotherapy has been shown to prolong survival after surgery, and the recently published evidence for this benefit will be discussed in detail.
Introduction
Prostate cancer is a disease characterized by considerable heterogeneity in its aggressiveness and natural history. In those presenting with low-grade and organ-confined disease, single-modality therapy - radical prostatectomy (RP), external-beam radiotherapy (RT) or interstitial brachytherapy - is curative in a large majority of cases. However, in the presence of adverse clinical or pathologic features connoting high-risk disease, results of local treatment alone are disappointing and there is increasing recognition of the need for a multimodal approach to treatment. The optimal sequence and constituents of such a strategy, however, remain controversial. The use of external-beam RT as primary therapy in conjunction with androgen-deprivation therapy (ADT) is supported by the results of several randomized controlled trials (RCTs).[1-4] Alternatively, surgery may be employed as the definitive initial treatment. While RP has traditionally not been recommended for high-risk disease in the past, its use in this setting is increasing, as evidenced by several large, modern, single-institution series.[5-8] The use of surgery in this context enables complete pathologic evaluation of the prostate gland and, therefore, additional adjuvant therapies may be directed to those at highest risk.
It should be acknowledged that no single definition of high-risk prostate cancer is universally accepted. Several definitions incorporating pretreatment prostate-specific antigen (PSA) levels, biopsy Gleason score, clinical T-stage and other variables have been proposed.[9-11] However, whatever the definition, high-risk disease implies an increased risk of relapse following initial therapy. In a large cohort of men treated with RP, Yossepowitch and colleagues demonstrated that, with the various definitions in current use, high-risk patients have an increased risk of metastatic disease (hazard ratio [HR]: 2.1-6.9; p < 0.05) and death from prostate cancer (HR: 3.2-10.4; p < 0.0005) compared with non-high-risk patients.[12] Attempts at treatment intensification with adjuvant therapies are, therefore, warranted.
This review will begin with a discussion of the factors associated with relapse after RP. Following this, the evidence for the use of additional therapies prior to or following RP in patients with high-risk localized disease will be considered in detail. Adjuvant RT, ADT and cytotoxic chemotherapy will be discussed in turn. Emphasis will be placed on the randomized trials, both completed and in progress, and selected retrospective reports where no RCTs exist. A discussion of adjuvant therapies given before or after primary RT for high-risk disease is beyond the scope of this review.
Risk Factors for Recurrence After Radical Prostatectomy
A number of clinicopathologic factors have been demonstrated to increase the risk of biochemical recurrence after RP for localized prostate cancer. Among these, the three most firmly established factors are clinical T-stage, biopsy Gleason score and pretreatment serum PSA. Each of these is subject to limitations. Clinical T-stage is prone to interobserver variability.[13] Serum PSA can be influenced by many non-malignant conditions. Gleason scoring is also susceptible to inter-reader variation, the scores appear to have shifted over time[14] and there can be considerable heterogeneity in outcome among patients with the same score (in particular, Gleason 7 disease).[15] Hence, when used singly, these factors are of limited usefulness. When used in combination, however, they provide a more robust prediction of treatment outcome. Several investigators have incorporated the three factors into clinical risk-prediction tools. Some of these are models that stratify patients to discrete risk groups,[9] while others are nomograms that have been validated in large populations.[16] The latter enable a more precise individualization of risk.
Since the creation of these models, numerous additional factors have been found to predict for risk of biochemical failure following surgery. These factors relate to pathologic findings from biopsy or RP specimens, kinetics in PSA levels prior to treatment and host factors. A list is presented in Box 1. Updated nomograms and risk-scoring systems that take account of some of these additional features have been developed.[17,18] Tissue and molecular factors influencing prognosis and response to treatment are being actively studied.[19,20]
Postoperative Radiotherapy
Radiotherapy is the only known curative treatment modality in patients with locally recurrent disease after RP. Its optimal use and timing following surgery, however, is controversial. RT may be delivered as an adjuvant treatment to all patients with adverse pathologic features at surgery or, alternatively, as a salvage treatment to those patients who demonstrate biochemical relapse. An adjuvant treatment policy offers potential advantages in terms of cancer control and overall survival, and respects the oncological principle of early intervention when residual disease is at its lowest volume. An early salvage policy, on the other hand, might offer advantages in terms of reducing the treatment burden and improving quality of life (QoL). Opinion among urological oncologists as to which is the superior approach is divided.[21] In this section, the published evidence for adjuvant and early salvage approaches to RT following RP are presented. A more detailed review of the adjuvant and salvage literature than can be provided here has recently been published.[22] Discussion of technical aspects of postoperative RT, including target volume delineation and dose-fractionation schedules, falls outside of the scope of this review and will not be covered hereafter.
Adjuvant Radiotherapy
Three multicenter, cooperative group Phase III trials of adjuvant RT following RP have been completed, two of which have been published, while the third has been presented in abstract form.[23-25] A meta-analysis of the trials based on summary data has also recently been reported.[26] As more than two decades have elapsed since the launch of the first of the trials and the publication of mature results, practice has naturally evolved in the interval and the trial results are not entirely applicable to contemporary practice. Nonetheless, they represent the only level I evidence that exists concerning postoperative RT and merit thorough examination.
Launched in 1987, Southwest Oncology Group (SWOG) 8794 was the first RCT designed to evaluate the efficacy of postoperative RT in localized prostate cancer with adverse pathologic features.[24] A total of 425 patients with histological evidence at RP of extracapsular extension (ECE), seminal vesicle invasion (SVI) or surgical margin positivity (SM+) were enrolled and randomized to receive either immediate prostatic bed RT to 60-64 Gy in 30-32 fractions or 'usual care and observation alone'. Importantly, in those randomized to observation, the trial protocol did not precisely specify the timing and nature of salvage therapies to be instituted in the event of disease recurrence. Metastasis-free survival (MFS) was chosen as the study's primary end point.
Trial results were initially reported in 2006 at a median follow-up of 10.6 years. At that time, the benefit in terms of MFS with adjuvant RT fell just short of statistical significance (HR: 0.75; 95% CI: 0.55-1.02; p = 0.06).[24] Trial results have recently been updated at a median follow-up of 12.6 years.[27] Adjuvant RT is now shown to significantly improve both MFS (HR: 0.71; 95% CI: 0.54-0.94; p = 0.016) and overall survival (OS; HR: 0.72; 95% CI: 0.55-0.96; p = 0.023). The 10-year OS was 74 and 66%, while median OS was 15.2 and 13.3 years for those randomized to adjuvant RT and observation, respectively. Hence, for the first time, adjuvant RT has been shown to improve survival compared with a policy of observation and usual care. Of note, the magnitude of this benefit is similar to that seen in RCTs of postmastectomy RT in node-positive breast cancer.[28-30]
An analysis of patterns of failure among patients enrolled in SWOG 8794 has been published.[31] In patients randomized to observation, gross local recurrence was found to be a more common event than metastatic relapse. Specifically, the cumulative incidence of macroscopic local failure was 24%, while that of distant failure was 16%. That local recurrence was the predominant mode of failure supports the rationale for and value of additional local therapy in this population with pathologically advanced prostate cancer.
As SWOG 8794 was launched soon after the advent of PSA testing, the trial protocol inevitably did not foresee some of the ways in which PSA data have come to be used in modern practice. First, the significance of attaining an undetectable PSA following surgery was not well understood when the trial was designed, and only 66% of enrolled patients had an undetectable (< 0.2 ng/ml) postoperative PSA. Second, there was no consensus on whether, and how, PSA levels should be used to guide therapy for recurrent disease, and thus no stipulations were made in the trial protocol on the timing of salvage therapy in relation to PSA failure. Of the 112 patients randomized to observation that recurred biochemically, only 70 received salvage pelvic RT, and of these PSA failure was the inciting factor in only 39 cases. In retrospect, it may be argued that salvage RT was unduly delayed or omitted altogether in many cases.
Limitations in the reporting of toxicity data from SWOG 8794 have been described previously, including the lack of a validated instrument for grading toxicity.[26] Nonetheless, an excess of both rectal complications (3.3 vs 0%; p = 0.02) and urethral strictures (17.8 vs 9.5%; p = 0.02) was reported with adjuvant RT. The results of a companion study investigating QoL in approximately half (n = 217) of SWOG 8794 trial patients have also been reported.[32] Patients completed QoL questionnaires at intervals over a 5-year period following randomization. While those randomized to adjuvant RT experienced greater compromise in bowel function for the first 2 years of follow-up, little difference was seen between the two arms thereafter. Global-health-related QoL was also initially worse in the adjuvant RT arm. Interestingly, by the end of the 5-year period, a significantly greater proportion of patients in this arm had normal global-health-related QoL than in the observation arm (69 vs 51%).
The European Organisation for the Research and Treatment of Cancer (EORTC) trial 22911 is the largest Phase III trial of post-RP RT.[23] Between 1992 and 2001, 1005 patients found at RP to have at least one of ECE, SVI or SM+ were enrolled. As with SWOG 8794, patients were randomized to either immediate RT (60 Gy in 30 fractions) to the prostate bed or to a 'wait and see' policy. Biochemical progression-free survival (bPFS) was the primary end point. As the trial was designed early in the PSA era, the methodological limitations of SWOG 8794 discussed earlier - namely, the inclusion of patients with detectable postoperative PSA levels and the lack of a clearly defined protocol for managing biochemical failure in the observation arm - also apply to the EORTC trial. Approximately 30% of enrolled patients had postoperative PSA in excess of 0.2 ng/ml.
Trial results were published in 2005 at a median follow-up of just 5.4 years.[23] No difference was seen in OS between the two arms (HR: 1.09; 98% CI: 0.67-1.79), with only 89 deaths and 59 distant failure events observed in the study population. The data are too immature at this time to draw conclusions regarding OS and MFS.
Follow-up is sufficiently lengthy to conclude that adjuvant RT confers a significant improvement in bPFS compared with observation (HR: 0.48; 98% CI: 0.37-0.62). Of those randomized to adjuvant RT, 74.0% (98% CI: 68.7-79.3%) were free of PSA failure at 5 years compared with only 52.6% (98% CI: 46.6-58.5%) of those randomized to observation. However, given the design of the trial, wherein patients in the observation arm were allowed to fail biochemically prior to initiation of salvage treatments, the clinical relevance of bPFS as an outcome is uncertain. A significant benefit in the more clinically important end point of locoregional recurrence was observed with adjuvant RT. At 5 years, adjuvant RT had reduced the incidence of locoregional failure by two-thirds (15.4 vs 5.4%; p < 0.0001).
The Late Radiation Morbidity Scoring Scheme of the Radiation Therapy Oncology Group (RTOG) and EORTC were employed to grade toxicity in EORTC 22911. An excess in the incidence of any-grade toxicity (p = 0.0045) and grade 2-3 toxicity (p = 0.0005) was observed in the adjuvant RT arm compared with the observation arm, but there was not a significantly increased rate of grade 3 toxicity when considered alone. The cumulative incidence of grade 3 toxicity at 5 years of follow-up was 4.2% with adjuvant RT and 2.6% with observation (p = 0.0726). No episodes of late grade 4 toxicity occurred.
A post-hoc subgroup analysis performed by the EORTC 22911 investigators raises the hypothesis that surgical margin status may be predictive of benefit from adjuvant RT.[33] Specifically, following a central pathologic review, margin status was found to have an interaction with treatment effect such that the benefit with respect to bPFS was no longer seen in margin-negative patients (HR: 0.87; 95% CI: 0.53-1.46; p = 0.601). A highly significant benefit was seen in patients with SM+ (HR: 0.38; 95% CI: 0.26-0.54; p < 0.0001). As the analysis was not prespecified, included only approximately half of all trial patients and was at odds with an earlier analysis performed on all patients,[34] its results should not be used to justify withholding RT in patients with pathologic T3, margin-negative disease.
Arbeitsgemeinschaft Radiologische Onkologie (ARO) 96-02/AUO AP 09/95, a third multicenter RCT investigating the use of adjuvant RT, has been carried out by the German Cancer Society. Results have been presented to date only in abstract form.[25] Limited details are available regarding its methodology. Between 1997 and 2004, 385 patients found at RP to have pathologic T3 disease (that is, ECE or SVI), with or without SM+, were randomized immediately after surgery to either adjuvant RT to 64 Gy in 32 fractions or observation, but were only considered eligible if an undetectable postoperative PSA was achieved. With the exclusion of those with detectable PSA after surgery, the intervention arm of ARO 96-02, unlike those of SWOG 8794 and EORTC 22911, may be considered to be a purely adjuvant approach.
To date, only outcome data for the primary end point of bPFS have been reported at a median follow-up of 4.5 years.[25] As in the other RCTs, adjuvant RT significantly prolonged bPFS (HR: 0.53; p = 0.0015). At 5 years, 72% of patients randomized to adjuvant RT were free of biochemical failure, compared with 54% of those randomized to observation. Toxicity data were only reported for the intervention arm. Publication of the complete trial report is awaited.
To summarize and place in context the magnitude of effect conferred by adjuvant RT, a 'number-needed-to-treat' analysis of the results of SWOG 8794 is helpful.[35] Compared with the salvage approach employed in the trial (wherein 33% of patients randomized to observation ultimately received salvage RT), approximately two additional courses of RT, given adjuvantly, are needed to prevent a PSA failure and seven additional courses are needed to prevent a death from any cause. These figures compare favorably with the findings of Scandinavian Prostate Cancer Group (SPCG)-4,[36] a trial comparing watchful waiting with RP, in which 20 RPs were required to avoid one prostate cancer death; and are similar to those of SPCG-7,[4] in which eight courses of primary RT, when given in addition to ADT for locally advanced prostate cancer, were required to avoid one death from prostate cancer.
Salvage Radiotherapy
In the context of prostate cancer, salvage RT refers principally to treatment instituted in one of two scenarios: a persistently elevated PSA following RP or a rising PSA following initial undetectability. As noted earlier, a proportion of patients included in both the SWOG 8794 and EORTC 22911 trials had detectable postoperative PSA levels and detailed guidance in the protocols on how and when to institute salvage therapy was lacking. The trials thus compared immediate (but not purely adjuvant) RT with a range of deferred therapies instituted for a range of early and late biochemical and clinical failures. While a survival benefit for immediate RT has been established in this context, it does not follow that a similar benefit would necessarily remain if comparisons were made between adjuvant RT and a rigorous policy of PSA monitoring and early salvage RT in the event of progression. Of interest, in the updated report of SWOG 8794 a subgroup analysis suggests that the magnitude of the benefit seen in MFS does not differ between patients with undetectable and detectable postoperative PSA (HRs presented graphically but not quantified), lending support to the hypothesis that early salvage RT may yield outcomes equivalent to that of adjuvant RT.[27]
The salvage RT literature is entirely retrospective, observational and nonrandomized, consisting principally of small, single-institution series. As such, the data are prone to numerous sources of bias, not least patient selection. Accepting these limitations, the largest retrospective study merits discussion. Stephenson and colleagues performed an analysis of 501 patients pooled from five academic centers, all of whom had received salvage RT for a rising PSA after RP.[37] The patients were judged by their treating physician to have isolated local relapse. RT was delivered to the prostate bed by a variety of techniques and the median prescribed dose was 64.8 Gy. Only 17% of patients received neoadjuvant ADT in addition to RT. A complete response to salvage RT, defined as the attainment of nadir PSA 0.1 ng/ml or less, was observed in 67% of patients. That such a large proportion responded initially to RT suggests that biochemical failure following RP is often due to isolated local recurrence, in keeping with the analysis of patterns of failure in SWOG 8794 described earlier.[31]
Progression-free probability (PFP) at 4 years of follow-up following salvage RT was 45% (95% CI: 40-50%) for the cohort as a whole. Given its large size, the cohort provided the necessary statistical power to identify prognostic factors among treated patients. Factors found on multivariate analysis to be significantly (p < 0.05) associated with PSA progression included a Gleason score of at least 8 (HR: 2.6), pre-RT PSA over 2.0 ng/ml (HR: 2.3), negative surgical margins (HR: 1.9) and PSA doubling time of 10 months or less (HR: 1.7). Outcomes were then stratified by these factors. Patients possessing none of these factors had a 4-year PFP of 77%. Yet even in subsets of patients possessing multiple features known to increase the risk of distant relapse - in whom salvage RT would not have been traditionally recommended - a sizeable proportion were shown to have durable PSA PFS after salvage RT. For example, in patients with completely resected, high-grade disease and a short PSA doubling time, early salvage RT (commenced when PSA = 2.0 ng/ml) conferred a 4-year PFP of 37%.
The analysis was later repeated using a larger cohort (n = 1540) with longer follow-up. The 6-year PFP following RT was 32% for the entire cohort. Stephenson specifically analyzed the association between PSA level prior to salvage RT and outcome, and found it to be a highly significant predictor of PFP. When patients were stratified by pre-RT PSA of less than or equal to 0.50, 0.51-1.00, 1.01-1.50 and more than 1.50 ng/ml, the PFP at 6 years was 48, 40, 28 and 18%, respectively. A nomogram taking account of pretreatment characteristics was developed to provide patients with an individualized probability of remaining free from progression 6 years after salvage RT.[38]
1 2 3 4
|
