Management of Lymph Node–Positive Prostate Cancer: The Role of Surgery and Radiation Therapy
Management of Lymph Node–Positive Prostate Cancer: The Role of Surgery and Radiation Therapy
There is no clear consensus on how to manage a subset of patients with prostate cancer (PCa) who present with involved lymph nodes (LN+). Although outcomes for these patients are uniformly worse than those for patients with localized PCa, they are better than outcomes for patients with bone metastases, with more than 60% of patients alive at 10 years after the initial diagnosis. This article reviews the existing data on outcomes for patients treated with various combinations of systemic and local therapies. Current evidence suggests both a disease-control benefit and a survival benefit to multimodality therapy, which combines systemic androgen deprivation therapy (ADT) with local therapies, such as surgery and radiation, without evidence of excessive treatment-related toxicities.
Prostate cancer (PCa) is the most common malignancy in men. In the United States, one out of six men will be diagnosed with prostate cancer during their lifetime.[1,2] Because of stage migration during the prostate-specific antigen (PSA) era, the vast majority of patients with newly diagnosed PCa have clinically localized disease, defined by the absence of nodal or distant metastases. In the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE) registry, out of 12,000 patients diagnosed between 1990 and 2004, only 2.4% presented with bony metastatic disease (M+) at their initial diagnosis. There are no recent data on initial presentation with radiographically evident lymph node–positive (LN+) disease in the absence of bony metastases. The current (7th) edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual groups M+ and LN+ disease together as stage IV. Many physicians tend to view these two categories of patients as having similar outcomes, and approach their treatment with the same algorithm.
The goal of this article is to review the data on management of PCa with lymphatic involvement (LN+) and focus on the role of multimodality therapy, which combines systemic and local therapies in managing these patients. Patients with LN+ PCa can be broadly separated into two groups, largely based on the bulk of nodal disease. Patients who undergo staging scans prior to surgery and have no radiological evidence of pelvic lymphadenopathy but are then found to have involved pelvic lymph nodes at the time of pelvic lymph node dissection (PLND) have pathologically node-positive (pN+) disease, and they are considered to be in the favorable category. This is in contrast to clinically node-positive (cN+) patients, who have pathologically enlarged lymph nodes on staging scans and in whom surgery is rarely even attempted. According to the 2013 National Comprehensive Cancer Network (NCCN) guidelines, patients with metastases to the lymph nodes by imaging studies (cN+) should be treated systemically with androgen deprivation therapy (ADT) or radiation therapy (RT) with long-term ADT, and patients with involved lymph nodes detected at the time of radical prostatectomy (pN+) are offered observation, ADT, or RT with long-term ADT (Figure 1). While ADT is the common denominator, there is no consensus on the role of local therapies in management of patients with LN+ PCa.
Management of Microscopic LN+ PCa, as Determined at PLND (pN+)
The incidence of nodal metastases in patients with presumed clinically localized PCa decreased from rates of 20% to 40% in the 1980s[6,7] to rates of 4% to 6% in the most recently published reports,[8-10] again likely due to stage migration in PCa during the PSA era. Indeed, the Mayo series revealed a marked decline in the incidence of positive lymph nodes at the time of radical prostatectomy (RP), from 9.1% between 1988 and 1993 to 1.8% between 1998 and 2001. Traditionally, involvement of regional lymph nodes in PCa was regarded as a poor prognostic factor, indicating systemic disease with limited long-term survival regardless of treatment. In the European Organisation for the Research and Treatment of Cancer (EORTC) Genito-Urinary Group protocol 30846, launched in 1986, RP was aborted if lymph node involvement was found during PLND, and men were randomized to either immediate or delayed ADT, which consisted of a 3.6-mg 1-month depot of the luteinizing hormone-releasing hormone (LHRH) agonist goserelin (Zoladex) combined with 50 mg of the antiandrogen cyproterone acetate three times daily during the first 4 weeks, for flare prevention. Orchiectomy was also allowed in lieu of medical castration. In the delayed-treatment arm, treatment was started at disease progression, defined as development of bone metastases, new lymph node or soft tissue metastases, or a 30 cc or greater increase in primary tumor volume. At the time of progression under endocrine treatment in each arm, further treatment was left to the discretion of the investigator but the continuation of endocrine treatment was suggested. At a median follow-up of 9.6 years, the time to death in patients with LN+ PCa (median 6.2 and 7.8 years with delayed vs immediate treatment, respectively) was greater than that in patients with M+ disease (median 2.5 years, based on a contemporary meta-analysis by the Prostate Cancer Trialists’ Collaborative Group). The update of the EORTC 30846 trial with a 13-year follow-up revealed that approximately 60% of deaths on both arms were due to PCa. This trial provides a good benchmark for the natural history of pN+ PCa in the absence of local therapy.
The Role of Radical Prostatectomy in Management of pN+ PCa
Although direct comparison of different trials is a flawed approach, their collective results do provide interesting insights. At the same time as the EORTC was enrolling patients on protocol 30846, the Eastern Cooperative Oncology Group (ECOG) launched a randomized trial of immediate vs deferred ADT in patients who were found to have pathologically involved lymph nodes at the time of RP, but in contrast to the EORTC trial, these patients underwent RP and were randomly assigned to receive immediate, continuous ADT (3.6 mg of goserelin monthly or bilateral orchiectomy, by patient choice), or to be followed up and receive ADT when clinical recurrence was detected (not counting events with only detectable or rising PSA concentrations). With a median follow-up of 11.9 years, median overall survival (OS) was 13.9 vs 11.3 years in the immediate-ADT and delayed-ADT groups, respectively. In the immediate-ADT group, 41% of patients died of PCa compared with 89% in the delayed-ADT group.
Other surgical series have been published, and outcomes from selected studies are summarized here as well as in the Table. Frohmuller et al published their series of 139 patients treated with PLND at the University of Wrzburg between 1969 and 1992. In this group, 87 patients received ADT alone and 52 received RP with ADT. The great majority of these patients were not staged with pelvic imaging prior to surgery. With a short follow-up of 4.5 years, the actuarial OS rates, prostate cancer–specific survival (PCSS), and progression-free survival (PFS) at 10 years were in favor of radical prostatectomy. The groups were heavily unbalanced by nodal disease bulk. Among patients in the RP + ADT group, 75% were staged as pN1 and 25% as pN2, whereas in the ADT-alone group 25% were staged as pN1 and 75% as pN2. However, the importance of this publication is in a detailed analysis of local tumor progression in the two groups of patients. Local disease progression occurred in 60 of the 87 (69%) patients who had not undergone radical prostatectomy. Because of clinically significant symptoms associated with local disease progression, such as urinary retention or bleeding, surgical intervention was necessary in 29 of these 87 patients (33%). Of these 29 patients, 18 required a single transurethral resection of the prostate (TURP) and 11 required multiple TURPs. In contrast, in the RP+ADT group, 4 patients out of 52 (8%) developed a local recurrence, but none required surgery; one patient after RP required urethral dilatation due to a stricture at the level of the vesicourethral anastomosis. Due to group imbalances and lack of robust statistical adjustment for these imbalances, comparison of outcomes in the two groups is challenging, but knowledge of the natural history of primary disease and the importance of local control, despite systemic progression, are emphasized by this publication.
A larger Munich Cancer Registry series of 938 patients accounted for such imbalances in their statistical methods. Although cases were not matched, random samples were generated. In this series, 688 patients were similarly treated with RP after PLND revealed LN+ disease, whereas 250 were treated without RP. About 75% of all patients received adjuvant ADT and 25% received adjuvant RT, with adjuvant therapies equally balanced between the two groups. With a median follow-up of only 5.6 years, OS was significantly better in the RP group, and after adjustment for age, clinical T stage, number of positive lymph nodes, World Health Organization (WHO) grade, and PSA level, patients who underwent RP were more likely to be alive, with a hazard ratio (HR) of 2 (95% confidence interval [CI], 1.6–2.6). A large Mayo series of RP in 507 LN+ patients at the time of PLND, with the longest follow-up being 10.3 years, revealed a 10-year PCSS of 86%, with 56% of men free from biochemical recurrence at the time of their last follow-up visit. In this series, 90% of the patients received adjuvant ADT and 9% received adjuvant RT. On multivariate analysis, Gleason score 8–10, positive surgical margins, and two or more positive lymph nodes were adverse predictors of PCSS, whereas adjuvant ADT use was not associated with systemic progression or cancer-specific survival. A meticulous PLND with no adjuvant therapies produced intriguing results in a publication by Bader et al. A total of 88 patients treated at the University of Bern between 1989 and 1999 were followed over a median of 3.7 years. Authors described dissection of tissues along the external iliac vein, in the obturator fossa, and along the internal iliac artery, with removal of a median of 21 lymph nodes. The projected probabilities of OS and of PCSS for these 88 patients were both 74% at 5 years. In a multivariate Cox proportional hazards model, which included the number of lymph node metastases, tumor stage, and Gleason score, the only variable affecting progression and cancer-specific death was the number of positive lymph nodes. In this study, only 8% of patients with one positive lymph node died of PCa, compared with a 25% mortality rate from PCa among patients with two positive lymph nodes and a 36% rate among patients with more than two positive lymph nodes.
Bulk of Lymphatic Involvement as a Predictor of PCa Outcome
There are no radiological data on bulk of lymphatic involvement and PCa outcome, most likely due to the rare presentation of isolated lymphatic metastases in the absence of distant metastatic disease (LN+M0). However, the available pathological data provide a meaningful insight into the relationship between the size and number of involved lymph nodes and disease prognosis. Schiavina et al confirmed the association between number of positive LNs and disease outcome. A total of 98 consecutive patients treated at the University of Bologna with RP between 1995 and 2011 were divided into three groups based on number of positive LNs and Gleason score. Of note, 80% of these patients received adjuvant ADT and 50% received adjuvant RT, with equal rates of adjuvant therapies between the groups. With a median follow-up of 5.7 years, patients with one to three positive LNs had PCSS at 5 and 10 years of 91% and 84%, compared with patients with more than three positive LNs, whose PCSS rates were 72% and 44%, respectively.
After finding a strong association between the number of positive LNs and the rate of PCSS, the researchers at the University of Bern looked further into the histopathological characteristics of the LNs in 102 patients treated between 1989 and 2002 with the same meticulous PLND, as described earlier. The median diameter of the largest LN metastasis per patient was 6 mm (range, 0.3–35 mm). Initially, the authors hypothesized that the presence of extranodal extension (ENE) would independently predict survival; however, on multivariate analysis—which included cancer volume in the prostate, T stage, Gleason score, ENE, number of positive LNs, and diameter of the largest LN metastasis per patient—only the diameter of the largest metastasis per patient (< 6 mm vs ≥ 6 mm) and the Gleason score predicted survival. In a subsequent analysis, the diameter of the largest metastasis (with a cut-off of 10 mm) per patient was the strongest prognosticator for all endpoints. The 5-year overall, disease-specific, and relapse-free survival rates for patients with largest nodal metastases of 10 mm or less were 85%, 87%, and 33%, compared with 54%, 57%, and 18%, respectively, for patients with largest nodal metastases greater than 10 mm. The HR of 4.2 for disease-specific survival (DSS) implied a more than quadrupled risk of dying from cancer for patients with larger LN metastases. One might hypothesize from this analysis that patients who are found to have LN+ disease by imaging are likely to have outcomes similar to those of patients with larger LN metastases; that is, patients with cN+ PCa might have roughly a quadrupled risk of death from PCa compared with patients with pN+ disease determined at the time of PLND.