The lymphatic drainage of the prostate has been well-described and confirmed by classic-era scintigraphic studies. More recently Cellini et al. demonstrated ascending, lateral, and posterior ducts which drain into external iliac nodes, internal iliac and obturator nodes, and sub-aortic sacral nodes, respectively [16]. Mattei at al. demonstrated that one third of the landing sites would be resected by a limited PLND and two thirds would be resected by an extended PLND [17]. Despite this detailed understanding, quite a bit of research has gone into determining which of these nodal groups, if any, represents the primary landing site for metastatic prostate cancer. In a study of metastatic cancer deposits from a cohort of 88 relatively high-risk men with pN+ disease after RP and extended pelvic lymphadenectomy, the most common site for metastasis (60 %) was the obturator fossa [18]. Overall however, 58 % had deposits in the internal iliac (hypogastric) and 36 % in the external iliac nodal areas, while 19 % had positive nodes in the hypogastric distribution alone. This broad spread of lymphatic metastases mirrors the broad drainage of lymph from the prostate gland, as nuclear medicine mapping studies have confirmed [17, 19].

In order to assess for nodal disease and perhaps achieve prostate cancer cure, several templates for PLND have been described. There is no value in performing PLND unilaterally for prostate cancer as drainage from the prostate is not limited to one side. A limited bilateral PLND can be considered the removal of the node-bearing tissue below the external iliac vein, out to the pelvic sidewall, bounded inferiorly by the obturator nerve, and medially by the bladder, collectively known as the external iliac and obturator nodes. A more standard PLND would consist of this dissection and in addition, all node-bearing tissue in the obturator fossa, deep to the obturator nerve and as proximal as feasible under the external iliac vein (the obturator nodes), although this template still consists of a limited sampling of nodes that drain the prostate. Finally, an extended PLND includes both of the above plus all fibrofatty tissue surrounding the internal iliac vessels up to their insertion at the common iliac bifurcations (the hypogastric or internal iliac nodes); some surgeons even include the additional removal of sub-aortic and pre-sacral nodes in their definitions of ePLND. While the latter description is perhaps the most extensive in the literature, ePLND is most commonly considered to be a node dissection that involves the removal of all node-bearing tissue from an area bounded by the external iliac vein anteriorly, the pelvic sidewall laterally, the bladder wall medially, the floor of the pelvis posteriorly, Cooper’s ligament distally, and the internal iliac artery proximally. There appears to be little indication and potential morbidity to resecting lymph nodes lateral to the external iliac vessels for prostate cancer.

There is likely a diagnostic and prognostic benefit to PLND in terms of disease staging. It has become clear that as more nodes that are harvested, more positive nodes are found [14, 20–23]. In a cadaveric study, Weingartner et al. compared nodal counts in cadavers without prostate cancer subjected to PLND to actual counts resected during radical prostatectomy (RP) + PLND [1]. Their data support the notion that while there exists great heterogeneity in terms of LN number between patients, there exist, on average, some 20–25 nodes in the standard/limited template area if all nodes are carefully counted. Estimates of LN counts necessary for optimal staging accuracy have ranged from 20 to as high as 28 [1, 14]. The current literature supports what makes intuitive sense—that more extensive dissections result in an increasing LN yield, and that an increasing nodal yield detects higher rates LNI. Caveats regarding nodal yield are that (1) unless pathologists are focused on reporting lymph node counts, accuracy of number may not be a priority, (2) greater nodal yields have been noted when PLND are sent as separate packets (external iliac, obturator, etc.) vs. en bloc in the bladder cancer literature [24, 25], and (3) there is significant inter-patient variability of nodal count irrespective of cancer status [1].

There may also be therapeutic benefits to PLND in prostate cancer. Briganti et al. showed that patients with two or fewer positive LN on final pathology had significantly better outcomes at 15 years compared to those with more than two positive LN [26]. Daneshmand et al. similarly found that a positive node density of <20 % improved disease progression rates and survival [27]. Palapattu et al. found that 52 % of men with a positive node density of <15 %, Gleason score 7 or less, and negative seminal vesicle invasion remained free of BCF at 5 years [28]. Finally, several large series of patients with positive LN at RP as well as the well-known Messing study cohort have shown biochemical recurrence-free survivals of 14–20 % over the long-term in the absence of any adjuvant therapy [10, 22, 28, 29, 38, 39]. Keeping in mind the landing zone studies, it is likely that only those whose positive nodes all lie within the surgical dissection region truly benefit from a PLND.

If a low burden of nodal disease portends a good prognosis after PLND, then PLND might improve prostate cancer outcomes even in patients that were pN0 on routine pathologic analysis. If we are missing a significant number of occult nodes as suggested by Pagliarulo et al. from USC then the number of nodes removed and the extent of lymphadenectomy might improve outcomes in node negative patients [30]. This group carefully re-examined 3,914 “negative” nodes by immunohistochemistry in 274 pT3 patients, and found that 13.3 % of 180 patients originally defined as N0 harbored occult LN metastasis. These patients had significantly poorer survival rates than patients who were truly LN negative, and had outcomes comparable to men who had LNI on initial pathology. Overall however, it remains controversial whether the removal of nodes deemed negative actually reduces PSA recurrence. Studies of increasing lymph node yield in pN0 patients, a surrogate for extensiveness of PLND, remain conflicting, with reports indicating both an improvement as well as no effect on prostate cancer oncologic outcomes [22, 31]. Recent data from Murphy and colleagues appears to corroborate that LN yield is not a predictor of BCF in pN0 patients, even when the patients were stratified into high- and low-risk groups [32].

The benefits of performing a PLND must be weighed against the potential risks of developing a complication. Major complications of PLND include lymphocele formation, deep venous thrombosis (DVT), vascular, neurologic and ureteral injury, and lower extremity edema. Among these, lymphocele formation is the most common, occurring symptomatically in up to 10 % of cases [54]. Patients who go on to develop a lymphocele typically present within the first month of surgery and experience a wide range of symptoms including pain, abdominal fullness, constipation, urinary urgency and frequency, symptoms of a DVT or related pulmonary embolism, fever, and leukocytosis. Commonly these patients lack associated physical exam findings and the diagnosis is only made with pelvic ultrasound or cross-sectional imaging. Treatment for a lymphocele depends on the clinical context but often can be managed expectantly or with percutaneous drainage.

Risk factors for the development of a lymphocele include the extent of PLND [55–57] and subcutaneous heparin administration for DVT prophylaxis [58, 59]. Of note, classically it was felt that the risk of lymphocele was lower among patients undergoing a transabdominal versus extraperitoneal operation, as it was felt that the transabdominal approach allowed for a larger surface area for the reabsorption of lymphatic fluid. This, however, appears to not be the case as a not insignificant number of patients undergoing laparoscopic or robotic surgery still experience this complication [60]. Regardless of surgical approach, this risk of a lymphocele can minimized through meticulous use of surgical clips and the avoidance of subcutaneous heparin. When heparin administration is recommended by current guidelines, injection in the upper extremities is advised [61, 62].

One complication of a PLND directly related to the development of a lymphocele is venous thromboembolism. This complication is felt to arise due to the compressive effects of a lymphocele on the deep veins of the pelvis and lower extremities, thus resulting in a DVT. In a meta-analysis by Eifler et al. [63], the authors found that undergoing a PLND at the time of RP was associated with an approximately twofold increased risk of developing venous thromboembolism. Consistent with this finding, one prospective study in which patients underwent an extended PLND on one side of the pelvis and a limited PLND on the contralateral side, found an increased rate of lymphocele and DVTs on the side of the extended template dissection [56].

Other less common complications of a PLND include vascular, neurologic and ureteral injury as well as the development of lower extremity edema. In terms of neurologic injury, structures at risk during a PLND include the obturator, gentiofemoral and femoral nerves. Of these, injury to the obturator nerve is most common and can be avoided by intraoperative visualization and careful clip placement. Most nerve injuries are recognized postoperatively with the development of pain, weakness or paresthesia. Fortunately serious injuries are rare and most can be managed with physical therapy and neuroleptic pain medications. Like neurologic injuries, vascular and ureteral injuries are also rare. These, however, are commonly identified intraoperatively and repaired at the time of recognition without any postoperative sequelae. One last uncommon complication of PLND is lower extremity lymphedema. Risk factors for this complication include extent of PLND and history of radiation therapy. Management options include the use of compressive stockings and exercise to promote fluid return from the lower extremities.