Terminology

There is a certain degree of confusion surrounding the definitions of PLND in bladder cancer. Historically, a standard PLND included the removal of all lymphatic tissue between the bladder and pelvic sidewall and along the external and internal iliac vessels proximally up to the midportion of the common iliac vessels, including tissue from the fossa of Marcille. Distally the limits were defined by the crossing of the circumflex iliac vein over the external iliac artery and the LN of Cloquet. Today, limited PLND is generally restricted to removal of the lymphatic tissue in the obturator fossa, in some cases with a dissection along the external iliac vessels. In contrast, an extended PLND includes all lymphatic tissue up to the boundaries of the aortic bifurcation. Some centers also include the LNs surrounding the distal aorta and inferior vena cava up to the level of the takeoff of the inferior mesenteric artery (IMA).

To avoid confusion with the terminology, the authors propose subdividing the current terminology into 3 groups ( Fig. 1 ). Limited PLND refers to tissue along the external iliac vein and from the obturator fossa. The limits of an extended PLND are laterally, the genitofemoral nerve; medially, the bladder wall; distally, the inguinal ligament and the pelvic floor; dorsally, the hypogastric vessels, including both the medial (considered presacral by some) and the lateral sides; and proximally, along either side of the common iliac vessels up to where the retracted ureter crosses them (ie, approximately between the proximal and middle third of the common iliac vessels). Also included is the tissue in the fossa of Marcille, dorsolateral to the bifurcation of the common iliac vessels and dorsal to the external iliac vein (considered presciatic by some). For a superextended PLND, the proximal border is extended up to the level of the origin of the IMA. The true presacral nodes, which lie dorsal to the bladder and rectum and are few in number, are, to the authors’ knowledge, hardly ever included in a PLND for bladder cancer.

Depicts the boundaries of a limited (I), extended (II), and superextended (III) PLND.

What Can Be Considered an Adequate Template for PLND?

The lymphatic drainage pattern of a malignant tumor determines the field of LN dissection. The available data on lymphatic tumor spread have been gained by using surgical specimens to identify the location of LN metastases. Leissner and colleagues examined LNs at 12 different anatomic sites in 290 patients who underwent radical cystectomy and a superextended PLND. Sixteen percent of LN metastases were found above the aortic bifurcation but only in patients with 2 or more LN metastases. None of the 29 single LN metastases was located above the aortic bifurcation. In this setting, a limited PLND would have missed 27% and extending the PLND to the iliac bifurcation, 10% of positive LNs in patients with a single positive node, respectively. This is in line with the report of Dangle and colleagues, who demonstrated that a limited PLND fails to identify positive nodes in 25% and a standard PLND (up to the iliac bifurcation) in 11% of patients. Vazina and colleagues evaluated 176 patients undergoing radical cystectomy and superextended PLND. Of those with pT3 or pT4 disease stage, 16% had LN metastases along the common iliac artery and at or above the aortic bifurcation but always combined with other LN metastases inside the small pelvis. In a prospective mapping study of 200 patients, Ghoneim’s group evaluated the probability of disease (node-positive) clearance with increasingly wider fields of LN dissection. After removal of all LNs in the pelvis (extended dissection), 65% of all positive nodes were identified. By extending the dissection to the level of the aortic bifurcation, this proportion increased to 79%. Again, no metastatic lesions located solely outside of the small pelvis (skip lesions) were found. Of the 48 LN-positive patients, 39% presented with bilateral LN involvement, leading to a recommendation of bilateral resection. These findings imply that a solitary LN metastasis outside the pelvis is rare and that exposing patients to the risk of including the nodes up to the IMA may be of limited benefit. LN metastases can be found at the level of the aortic bifurcation and IMA. They are generally associated, however, with multiple affected LNs with a correspondingly poor prognosis.

Certain problems are inherent to all mapping studies. The location to which the tissue is assigned may vary. For example, tissue from the common iliac bifurcation can be considered external, internal, common iliac, or from the obturator fossa, depending on the institutional standards. Another issue is that information is only gained on the nodes that were removed and not on the nodes that were possibly missed. If no nodes were retrieved at the aortic level, then the conclusion is that there are no positive nodes to be found at that location.

A recently published study identifying the true lymphatic drainage pattern of the bladder, however, is less limited by these problems. By applying new imaging techniques to accurately define the bladder’s primary lymphatic landing sites, for the first time, further insight into the lymphatic drainage pattern of the bladder was gained and the clinical observations (discussed previously) substantiated. Technetium Tc 99m–nanocolloid was injected into 1 of 6 specified non–tumor-bearing sites of the bladder to minimize the risk of artifacts by blocked lymph vessels or nodes in the case of lymphangiosis. The primary lymphatic landing sites were located by preoperative single-photon emission CT (SPECT) combined with CT. They were verified intraoperatively with a gamma probe and double-checked by a backup extended PLND. If the identified primary landing sites from all 6 injection sites in the bladder were added, a median of 28 nodes per patient was found. Ninety-two percent of all LNs were found distal and caudal to where the ureter crosses the common iliac arteries and 8% proximal thereof. No proximal nodes were detected without simultaneous detection of additional radioactive LNs within the endopelvic region. In a similar study on 40 patients with unilateral bladder cancer, technetium Tc 99m–nanocolloid was injected into the non–tumor-bearing lateral bladder wall. All patients had at least one radioactive LN on the ipsilateral side and 40% of patients had at least one additional radioactive LN on the contralateral side. This study, as others, relies on the operative procedure (PLND) for verification and a few LNs may have been missed despite preoperative SPECT/CT, intraoperative gamma probe verification, and extended backup PLND. The risk of error, however, is much lower than in conventional mapping studies, making this the most accurate description of the bladder’s lymphatic drainage pattern to date.

What Can Be Learned From This Technetium-Based Mapping Study?

First the idea of sentinel node, as observed in cancers of other organs, may not be valid for the bladder. Because a median of 4 primary lymphatic landing sites was found in 1 of 6 bladder sites, the median number of primary lymphatic landing sites per bladder is approximately 28. This implies a more complex lymphatic drainage pattern in the bladder than in other organs, such as the breast, a finding also described by other investigators. The removal of one sentinel node to determine if pelvic lymphadenenctomy is necessary or not does not seem applicable for the bladder. Instead, removal of multiple nodes, potentially on both sides of the pelvis, is required. In this setting, removal of nodes/lymphatic tissue from a well-defined template may be more reliable and cost effective. This was supported by Liedberg and colleagues’ observations that, when applying intraoperative sentinel node detection in invasive bladder cancer, 19% of LN-positive patients proved false negative and would have been missed. Technetium seems reliable for the detection of normal but not metastatic nodes.

Second, a basis to discuss the template, which can be considered appropriate, was given. In 1962, Whitmore and Marshall described a standard template for bilateral PLND, which included all tissue from the pelvis up to the proximal third of the common iliac artery, at the level of the ureteric crossing. This standard template corresponds to what was defined as an extended PLND. Based on the mapping study (discussed previously), the percentage of primary landing sites left in situ, between where the ureter crosses the common iliac artery and the IMA, would be 8%. Is it worthwhile to resect the remaining 8% to 10% of potential primary lymphatic landing sites between the uretero-iliac junction and the IMA?

Cure is achieved only in patients with a limited number of metastases in normal-sized nodes, a situation not commonly found limited to locations outside of the pelvis. The potential benefit of extending PLND up to the IMA can be calculated. The prevalence of histologically positive LNs in patients with clinically negative LNs is approximately 25%. Of patients with positive LNs, 35% will survive, which corresponds to approximately 8 of 100 cystectomy patients. Because at most 8% to 10% of patients can be expected to have positive LNs above the uretero-iliac junction, at most 1 to 2 of 100 cystectomy patients would benefit from a superextended PLND up to the IMA. It is questionable whether overtreatment in 99% of cystectomy patients is justifiable. In patients receiving an ileal orthotopic bladder substitute, there is also a potential risk of impaired continence through damage to the hypogastric nerves after higher resection as well as a potentially increased complication rate.

Before resecting the para-aortic/paracaval LNs, the LNs of the fossa of Marcille should be removed. Four percent of all technetium Tc 99m–positive LNs or 12% of all technetium Tc 99m–positive LNs along the external iliac vessels were found in the fossa of Marcille (ie, the presciatic nodes), which are dorsal to the external iliac vein and dorsolateral to the bifurcation of the common iliac vessels. This approach is easier and faster and may be of greater benefit to patients. Meticulous dissection of the internal iliac vessels is also important; 10% of all technetium Tc 99m–positive LNs are found medial to the internal iliac artery (some term these presacral), an area often neglected.

In summary, an adequate or extended PLND is what Whitmore and Marshall already defined as a standard PLND in 1962.

Importance of the Number of Nodes Removed

The number of nodes resected has been proposed as a surrogate for the quality of PLND. This is a difficult proposition because the number of nodes detected not only depends on the extent and meticulousness of dissection but also on the pathologist’s dedication to finding and carefully examining the LNs. Subdividing the tissue from a defined template into more separate packages results in a larger number of LNs detected as does requiring a certain number to be submitted by institutional policy. There also is a wide variability in the number of nodes per individual patient. These limitations are underlined by the observations made by Fleischmann and colleagues, who showed that in a single-center series of patients with LN metastases, the number of nodes removed ranged from 10 to 43 after removal of lymphatic tissue from a standardized template. Independent of the number of nodes removed (few or many), recurrence-free survival and overall survival remained unchanged.

These observations all make using the number of nodes as a surrogate for the quality of PLND and consequently for the chance of survival questionable. On the contrary, a standardized template based on a universally accepted nomenclature where the tissue between the defined boundaries is meticulously removed should set the standard for PLND.

Impact of PLND on Prognosis and Outcome

PLND at the time of cystectomy increases the probability of completely removing all cancerous cells. The more complete the PLND, the better the chance of cure. Prognosis further depends on the stage of the primary tumor, the number of positive nodes, LN density, and, most importantly, according to a multivariate analysis, extranodal growth.

In 1950, Kerr and Colby first suggested a benefit for PLND. They reported a 2-year survival in 2 patients after cystectomy and PLND despite nodal disease. Whitmore and Marshall in 1962 stated that for patients with only a few pelvic nodal metastases radical cystectomy with PLND has provided some successful 5-year results. Smith and Whitmore and Pagano and colleagues several years later questioned the benefit of PLND for survival. Smith and Whitmore postulated that distant metastases occur at or approximately the same time as does regional nodal spread in the majority of patients, the conclusion being that extrapelvic disease is a significant determinant of survival and interval to recurrence, regardless of the extent of nodal metastases. In contrast, Skinner in 1982 stated that a meticulous PLND could provide cure and control of pelvic disease in some patients with regional LN metastases without increasing the morbidity ( Table 1 ). Since then, more contemporary data originating from centers, including the University of Southern California in Los Angeles, the Memorial Sloan-Kettering Cancer Center in New York, and the University of Bern, Bern, Switzerland have confirmed that with extended PLND, approximately a third (31%–35%) of bladder cancer patients with LN metastases were alive at 5 years. In the largest single-center cystectomy series to date, published by Stein and colleagues, using superextended PLND, 31 of the 246 (23%) patients who were node positive were alive at 5 years.

Survival benefit is dependent on the extent of the template applied for lymphadenectomy. This was illustrated in 2 recent interinstitutional comparative studies assessing the survival benefit associated with the extent of the PLND performed. Dhar and colleagues compared the outcome of well-defined patient groups (pT2/3, clinically N0M0, no neoadjuvant chemotherapy) at 2 high volume centers performing either a limited or an extended PLND. LN-positive patients in the limited LND group demonstrated significantly decreased 5-year recurrence-free survival when compared with the extended LND group (7% vs 35%). There was also a 2-fold increase in the rate of LN-positive disease in the extended LND group, suggesting significant understaging in patients undergoing limited LND. Zehnder and colleagues compared 2 well-defined patient groups (pT2/3, clinically N0M0, no neoadjuvant therapy). One group was from the same institution as the study (discussed previously) of extended PLND. The other group was comprised of patients from another high-volume institution routinely performing a superextended PLND. Despite more nodes being removed and more positive nodes detected in the patients who underwent superextended PLND, no significant differences in survival outcome were seen. Five-year recurrence-free survival and overall survival rates in LN-positive patients after superextended versus extended PLND were 40% versus 42% and 34% and 38%, respectively. The lack of difference was consistent irrespective of nodal status or receipt of adjuvant chemotherapy. A partial explanation may be that solitary or few metastases directly to these extrapelvic nodes is rare and prognosis in general is poor for patients with nodal disease outside the pelvis. Therefore, the removal of these nodes hardly affects outcome.

So, at a first glance it seems that approximately 35% of patients with positive LNs could benefit from an extended PLND. But which patients will ultimately benefit?

It seems that patients with LN metastases as well as patients without any identifiable sign of LN metastases may profit from PLND. Herr reported that survival after cystectomy for both patients with LN-negative and node-positive disease was improved (reduced local recurrence rate) when a higher number of nodes was removed. Similarly, a study by Leissner and colleagues demonstrated that of patients without or with 5 or fewer positive nodes who had more than 16 LNs removed, 65% were disease-free and alive at 5 years. In contrast, 51% of the patients who had fewer than 16 LNs resected were alive and disease-free at 5 years. A large collaborative study of the Southwestern Oncology Group evaluating patients treated at 4 high-volume centers found that removal of more than 10 nodes increased 5-year overall survival from 44% to 61% in comparison with removal of fewer than 10 nodes. Recently, Vazina and colleagues reported that enforcing an institutional policy requiring removal of at least 16 nodes improved survival from 40% to 52% in patients with a greater than 2-year follow-up. These data suggest that a greater number of nodes removed (implicating a more extended template) at the time of cystectomy is beneficial for some patients. Using an extended field for PLND increases the number of nodes removed and the probability of removing positive nodes. The removal of nodes found pathologically negative seems equally important, however. Data based on reverse transcription–polymerase chain reaction studies have shown that 20% to 30% of pathologically negative nodes may harbor micrometastatic disease. In this sense, removal of histologically negative nodes harbouring occult metastatic disease may have a significant impact on outcome.

Lerner and colleagues was one of the first to report an increased risk of cancer-related progression and death associated with 6 or more LN metastases identified at the time of cystectomy. Similarly, Mills and colleagues from Bern, Switzerland demonstrated that patients with fewer than 5 involved nodes had a statistically significant survival advantage compared with those with greater than or equal to 5 involved nodes. Herr and Stein and colleagues almost simultaneously proposed the number of positive–to–total number of LNs (LN density) removed ratio as a significant prognostic variable. The hypothesis was based on LN density incorporating both the extent of disease based on the number of positive LNs as well as the extent of LN dissection based on the number of nodes removed. Stein and colleagues from the University of Southern California found that patients with less than or equal to 20% LN metastases density had a 43% 10-year recurrence-free survival compared with a 17% 10-year survival for those with greater than 20% density. Furthermore, recent studies support the observation that LN density may be an independent predictor of survival in patients with muscle invasive bladder cancer. Density reflects the extent of nodal dissection, however. If 1 of 5 nodes is positive, the ratio is 20%. If 1 of 50 nodes is positive, then the ratio is 2%. This patient may well have a better chance to survive because it is likely that there really are no further positive nodes. Patients with 1 positive node of 5 nodes removed have a high probability of having additional occult positive nodes left behind. Another factor having an impact on outcome is the tumor volume in affected nodes. Of patients with grossly node-positive bladder cancer treated with surgery alone, 24% survived and 76% died of disease. The median survival time for all patients was 19 months and 10 years for surviving patients. Skinner reported a 36% improvement in 5-year survival in bladder cancer patients with limited nodal burden. The conclusion was, “patients who probably benefit most from a meticulous dissection are those with undetectable micrometastases to a few nodes.” The authors’ group found that patients with maximal nodal metastasis up to 0.5 cm had a median survival of 84 months compared with only 16 months in those with metastases greater than 0.5 cm.

More recently, the authors demonstrated that although the number of positive nodes and volume of metastatic disease are important prognosticators for survival, the strongest predictive factor in multivariate analysis is extracapsular growth. These more aggressive tumors, which present with extracapsular growth in the LNs, are also the ones that involve multiple nodes and develop metastases more rapidly.

One remaining point of discussion is the role of chemotherapy. In a study from the University of Southern California and the University of Bern comparing superextended and extended PLND, no difference in outcome was observed despite removal of more nodes and more patients undergoing chemotherapy in the superextended PLND group. In multivariate analysis, adjuvant chemotherapy did not affect recurrence-free survival. In an initial report from a randomized prospective multicenter trial to determine the effect of neoadjuvant chemotherapy in patients with stage T2-T4 urothelial bladder cancer, neoadjuvant chemotherapy was associated with a significant survival benefit. On a repeated evaluation, however, the extent of surgery was found a more highly significant predictor of survival.