This study enrolled high-risk PCa patients with estimated LN involvement of 21.6% on average. PS-matching followed by survival analysis using preoperative factors showed a statistically significant value of extended PLND with a promising hazard ratio. Attempts to analyses in limited cohorts, such as in cohort excluding patients with PSA > 0.2 ng/mL on the first checkup and in cohort stratified by timeframe with different LN yield, to reduce potential bias as much as possible also supported the existence of oncological benefits of extended PLND. In addition, Cox regression multivariate analysis using postoperative factors showed its independent predictive value. These results definitely indicate the superiority of extended PLND for BCR-free survival and the optimal risk-range setting for patients enrolled in the study.
After safety and feasibility was established, extended PLND became the gold standard for providing accurate nodal staging [17]. Guidelines from both the American Urological Association and European Association of Urology recommend PLND with an extended template in patients with intermediate- and high-risk PCa as an expert opinion under situations without distinct evidence of the oncological benefits of PLND [18, 19]. Some systematic reviews and meta-analysis referred to oncological outcomes, but most consisted of reports with low levels of evidence, and furthermore, there are considerable variations in the LN yield and the positive rate of metastasis [7–10]. However, recent reports found a modest superiority for PLND regarding the effects of BCR-free survival with hazard risks of approximately 0.6–0.7 [8, 9].
The first attempt of a RCT to determine the oncological benefit of extended PLND unfortunately failed to show a statistically significant difference, despite the study being precisely designed to expect a 15% advantage in BCR-free survival [12]. The report is an important reference in the field of PLND. However as pointed out in correspondence, there was a favorable cancer profile characterized by more than one-third of patients having an ISUP grade group 1, and the estimated risk of LN involvement was 11 to 12%, which might be due to incorrect patient selection that reduced the effect of extended PLND [20]. On the contrary, a recent RCT to determine the role of prophylactic pelvic irradiation for high-risk, locally advanced PCa successfully showed improved disease-free survival [21]. Although one cannot compare the different treatment modalities on the same basis, the crucial difference is that the enrolled patients had a higher risk PCa in the latter RCT; and the estimated risk of LN involvement was more than 20%. Therefore, careful planning prior to a study is essential for the selection of patients and risk-range setting, especially when shedding light on the therapeutic role of PLND.
There is an attractive report regarding molecular LN staging that may alter patient risk stratification and thus affect PLND application or efficient adjuvant therapy in the near future. Molecular LN analysis is determined by prostate-specific expression of Kallikrein 3 (KLK3) using a quantitative polymerase chain reaction that detect LN metastases with a higher sensitivity and the risk of BCR than histopathology [22]. The study group later proposed a combined KLK3/ transmembrane serine protease 2 panel as a diagnostic and prognostic tool for molecular LN analysis [23]. Interestingly, patients with pN0/molecular N1 were observed at the same proportion or slightly more that those with pN1/N1, and this implies that the PLND removes unrecognized micro metastases by histopathology to varying degrees.
The interpretation of this study suggests unique characteristics and limitations. Primarily, the current study enrolls PCa patients whose clinical stage has been determined by a combination of DRE and multiparametric MRI. The use of an MRI finding leads to a significant upstaging of clinical T-stage and risk grouping, and further affects treatment intensification in around 30% of patients. These effects arise as MRI shows higher sensitivity and lower specificity compared to DRE [24]. Recent reports suggest that MRI staging has advantages in the established risk grouping and LN involvement nomograms originally developed with DRE staging in contemporary PCa patients; to determine a locally advanced high-risk group in the European Association of Urology (EAU) risk group and to improve model performance both at Memorial Sloan Kettering Cancer Centre 2018 and Briganti 2012 nomograms [25, 26].
Whereas, in our cohort, the re-classification to high-risk from the intermediate-risk was only 4.7% (12/253 patients) and to VHR from SHR subgroup was 20.0% (25/125 patients). For all, the risk migration remained less than 9.8%. This is probably because the staging did not rely solely on MRI, and the NCCN classification with T3 or higher as one of the high-risk criteria was used in the risk stratification. Actually, the concordance rate between MRI and DRE was high in determining clinical T3 stage in our cohort. Nevertheless, the measured values of BCR-free survival cannot merely be compared to those in other reports since the Will Rogers phenomenon arising from the risk migration has occurred often [27]
Secondly, many previous studies set patients that received preoperative ADT as one of the exclusion criteria, while this study does not. In the real-world practice, considerable preoperative ADT is still performed, and the rate increases at higher disease stages. We took into consideration that multiple biases in patient selection or confounding of background would be unavoidable, whether these patients were included or excluded. Nevertheless, it is well acknowledged that preoperative ADT has no significant effect on BCR-free survival [28], and recent findings include no significant effect even on resection margin positivity [29]. These are the rationale for deciding the patients to be enrolled in this study, and naturally preoperative ADT was adjusted in the PS-matched analysis; the reduction of positive resection margin was not observed, while the independent predictive value for BCR was lost in the Cox regression analysis.
A statistical approach using the rigorous application of a causal inference framework and the concept of causal mediation analysis has attracted attention, which is frequently used analysis in the field of psychology. Although its effectiveness as an analytical method for lymph node dissection in the field of oncology is still immature, this method may clearly quantify the direct effect and the indirect effect of PLND on survival. A recent report indicated that the impact of extended PLND was not restricted to a staging benefit, but also involved a therapeutic benefit of reducing BCR or second-line therapy experience at approximately 30% [30]. When a well-designed RCT is not feasible due to ethical issues, such methodology may resolve the analytical challenges for exploring oncological benefits of PLND.
Despite that data were collected prospectively, and PS-matching adjusted the differences in known background factor, the study design is a retrospective analysis of outcomes by multiple surgeons at a single institution, the comparison with patients ranged from without to standard PLND, the short follow-up periods, and small number of patients would limit the oncological impact of the results. Despite such limitations, this report should serve as a reference indicating the positive therapeutic role of PLND.