Determinants of Long-term Survival of Patients With Locally Advanced Prostate Cancer

The Role of Extensive Pelvic Lymph Node Dissection

M Moschini; N Fossati; F Abdollah; G Gandaglia; V Cucchiara; P Dell'Oglio; S Luzzago; SF Shariat; F Dehò; A Salonia; F Montorsi; A Briganti


Prostate Cancer Prostatic Dis. 2016;19(1):63-67. 

In This Article


Our group recently proposed for, we believe, the first time the impact of ePLND in pN1 PCa patients during RP.[11] In this context, high-risk PCa population are at major risk to suffer from evident[9] or occult[10] node metastases after RP, and a multimodal approach in this category has been proposed with the use of aRT that is related to an improvement in survival.[6–8] However, no one at the time has ever tested whether ePLND could guarantee any improvement in cancer survival in these patients. To verify this hypothesis, we tested the relationship between the number of RLNs and CSM risk after RP in patients with locally advanced disease after RP categorized as pT3-pT4.

We report several findings. First, at univariable analyses, RLNs was directly and positively related to CSM rate. Specifically, a higher number of RLNs was associated with a higher CSM rate (HR: 1.02; P=0.01). However, after adjusting for all available confounders, the relationship between the number of RLNs and CSM rate flipped. In fact, removing more LNs was associated with lower CSM rate (HR: 0.97; P=0.03). As explained in our previous report,[11] the controversy between the results of univariable and multivariable analyses might be explained: in fact, patients with more aggressive tumors are offered a more ePLND. For this reason, should be considered the possibility that exist a selection bias where patients with the higher number of RLNs experienced the more aggressive tumors. Second, factors other than the number of RLNs appear to have an impact on CSM rate in pT3-pT4 patients, specifically, patients with positive nodes, older, with higher grade and those not receiving aRT appear to have less favorable CSM rate.

In previous report, Schiavina et al.[14] analyzed data of 872 patients, and 470 were found with intermediate- and high-risk PCa and were patients treated with RP. Authors found that removing ≥10 LNs was associated with improvement in biochemical recurrence-free survival. However, these results were limited by the fact that only 567 patients were submitted to PLND and the mean number of LNs removed was only of 10.9 and the follow-up was only 55.8 months. Other reports focalized on pN+ or on pN0 population, but these were strongly limited by the fact that some pN0 patients according to Pagliarulo et al.[10] experienced occult node metastases, especially in locally advanced PCa with a rate of 13.3% for pT3 population. Another limitation to consider pN0 or pN1 population is the well-described Will Rogers phenomenon[15] where, pN0 patients with higher number of RLNs are better staged and thus are more likely to be really free from node metastases. Controversially, pN0 patients with lower number of RLNs are less likely to be accurately staged and might actually harbor an overlooked nodal metastases. We addressed these problems using a pT3-pT4 N0-1 PCa population in our analyses and homogeneously treating this population with ePLND.

Some reports for pN+ and pN− described an improvement in survival outcomes: Joslyn et al.[16] analyzed data of 13 020 PCa patients treated with RP. Authors found that removing ≥4 LNs was associated with a more favorable survival rate in both node-negative and node-positive patients. In another report, Abdollah et al.[11] described in 351 pN1 PCa patients treated with RP that a removal of a higher number of LNs during RP was associated with an improvement in CSM. In pN0 population, Murphy et al.[17] examined the data of 964 patients PCa patients treated with RP, assessing that the number of RLNs pT2-pN4-N0M0 does not impact on survival, although Masterson et al.[18] found in 4611 PCa patients treated with RP that the number of RLNs was associated with a lower biochemical recurrence risk when pN0 patients were selected. Finally, Bivalacqua et al.[19] have shown that patients undergoing ePLND had better oncological outcomes at 10-year follow-up compared with their counterparts receiving a limited PLND. Conversely, other authors have found that the number of RLNs was not associated with biochemical recurrence and/or survival.[20–22] In context, our study was the first to demonstrate a relation between RLNs and survival in patients with pT3-T4 disease. We can hypothesize that, given the relatively high risk of occult nodal invasion in this patient category, maximizing local control by an anatomically defined ePLND might have a beneficial impact on CSM.

Our study has several clinical implications. Our findings showed that a more ePLND offers better cancer control outcomes in patients with pT3–pT4 after RP. Given that, and considering the improved risk in facing evident or occult LN metastases in patients with advanced local stage, an ePLND should be offered to patients with preoperatively risk factors in order to improve staging procedure and survival. However, our results cannot answer the clinically relevant question regarding the optimal anatomical extent of ePLND in pT3-T4 patients. Certainly, all these patients should receive an anatomical dissection of all lymphatic tissue in the obturator fossa, as well as along the external and internal iliac vessels. In addition, our results seem to support not only a meticulous and careful dissection of all these areas but also of the presacral and common iliac areas in patients with adverse PCa characteristics, as previously recommended,[23,24] However, such an approach can be only suggested, but not fully supported, by our results. In this context, it should be highlighted that our analyses included only patients with pathologically defined locally advanced disease. Future studies are needed to validate our results in patients with cT3-4 disease as defined by preoperative magnetic resonance imaging.[25] Of note, Muralindhar et al.[26] recently showed that patients with occult T3a disease had less than half the risk of positive margins as compared with those with clinical T3 PCa. However, men with occult T3b disease or high-grade T3a disease had similar risk of positive surgical margins as compared with those with clinical T3 disease. Therefore, identification of these patients with novel imaging modalities, including preoperative magnetic resonance imaging, might have important clinical implications. Finally, the impact of aRT already described in other series was confirmed,[6,7] additionally for the first time our report assesses the impact of ePLND on survival, suggesting that not only local control is important in these patients but also LN invasion could impact on their survival regardless of local control. Our results therefore confirmed previous findings,[8,27] supporting the beneficial impact of aRT on locally advanced PCa.

Our study is not devoid of limitations. First, our results were derived from retrospective, observational data. Therefore, our findings should be considered in the context of retrospective, observational evidence and warrant prospective, randomized validation. Despite an anatomically defined ePLND was routinely offered when clinically indicated to all the patients undergoing RP in our institution, a fluctuation in the number of RLNs was observed. This might be derived from individual variability related to patient characteristics, previous infections or the technique used by pathologists. Moreover, we were not able to identify patients who received a PLND that included the presacral stations. Therefore, we did not adjust our analyses for this variable. Second, the use of aRT and/or androgen-deprivation therapy was based on the clinical judgment of each treating physician according to patient and cancer characteristics.