Nodal staging in prostate cancer and patient selection for PLND

LN status is an important prognostic indicator of outcome in patients undergoing radical prostatectomy for localised prostate cancer.8–10 The risk of LN metastases in an individual patient can be estimated using preoperative nomograms, which are based on a number of clinical and pathological parameters such as prostate specific antigen (PSA) level, clinical stage, biopsy Gleason Score and the number of positive cores at biopsy.11–14

The current European Association of Urology guidelines recommend PLND for all high-risk patients, as well as intermediate-risk patients when the nomogram-based predicted risk of LN involvement is >5%.1

For patients at risk of LN metastasis, bilateral PLND remains the gold standard procedure. However, the optimal anatomical extent of PLND has been the subject of debate. Standard PLND generally consists of removal of all nodal tissue anteromedial to the external iliac vein and from within the obturator fossa.15 A number of studies have demonstrated increased LN yield and detection of metastases with an extended PLND template.15–20 As a result, extended PLND is currently the standard of care recommended in all patients for whom nodal staging is indicated.1 This involves the removal of LNs overlying the external iliac artery and vein, within the obturator fossa, as well as medial and lateral to the internal iliac artery.1 ,15 The common iliac and presacral areas may also be resected as part of the extended PLND template, as there is evidence that this improves nodal staging.1 ,21 ,22

In view of the risk of complications and morbidity following an extended PLND there is an interest in alternative, less invasive methods of nodal staging. The concept of sentinel lymph node (SLN) dissection in prostate cancer was first described by Wawroschek et al in 1999.23 Since then a number of studies have investigated the utility of SLN mapping to guide PLND in prostate cancer and potentially reduce the extent of dissection, although the procedure is regarded as experimental at this stage.1 ,24 A recent review of the literature reported that SLNs were identified outside of the dissection template for an extended PLND in 4.1–25% of cases, and of patients with positive LNs 3.5–17% had SLN metastases outside of the extended PLND field.25 These findings highlight the potential for SLN mapping to improve nodal staging. However, the procedure is limited by false-negative results, particularly in high-risk patients, which may occur due to the obstruction of lymphatic channels by tumour.26 It has been proposed to combine SLN guided LN dissection with extended PLND to optimise staging25 although this may represent an excessive dissection for many patients.27 Due to these limitations, SLN sampling is not currently recommended for nodal staging of patients with prostate cancer, however the future potential of this technique is an area of ongoing research.

Imaging studies such as abdominal computed tomography (CT) and multiparametrical MRI may be useful for nodal staging when there is gross LN involvement, however they lack sensitivity for the detection of microscopic LN metastases.1 ,28 Newer imaging techniques such as 11C-choline or 18F-choline positron emission tomography (PET)/CT and prostate-specific membrane antigen-PET CT (PSMA PET/CT) are not currently recommended for LN staging.1

LN yield at PLND

One of the key concerns for both the surgeon and the pathologist is what constitutes an adequate PLND in terms of the LN yield. A number of studies have established a relationship between the number of LNs examined at PLND and the rate of metastatic disease detection18 ,29 ,30 and this partly underpins the rationale for performing an extended PLND in all patients. As a result, the LN yield is often used as a surrogate marker for the completeness of surgical dissection and, by analogy, it's adequacy for accurate staging. Despite this, there are currently no guidelines as to the median number of LNs that should be examined following a PLND for prostate cancer. This is in contrast to many other organ sites where such recommendations have been established.

Individual studies have attempted to address this question, and a variety of LN counts have been proposed in the literature (table 1). In a cadaver-based study, Weingärtner et al31 sought to identify the normal anatomical LN yield that should be expected from a standard PLND, which comprised all of the nodal tissue from between the internal and external iliac arteries and the obturator fossa. The study included 30 cadavers, where death was not cancer related, as well as 59 patients with prostate cancer. Based on their findings, they proposed 20 LNs as a guideline estimate for a standard PLND, however they reported marked interindividual variation, with the number of LNs retrieved in an individual case ranging from 8 to 56. Similarly, another cadaveric study designed to examine the LN yield at PLND performed for urothelial carcinoma, also found a significant variation in LN counts between individuals.32

Others have correlated the rate of LN metastasis with the number of nodes retrieved in an attempt to establish the minimum yield required for accurate staging. Barth et al recommended examination of at least 13 LNs from a standard PLND, based on their study of 283 patients where the rate of LN metastasis more than doubled when >12 LNs were examined.30 In a larger series where the majority of patients underwent an extended PLND, Briganti et al reported that the ability to detect LN involvement was close to zero when <10 LNs were retrieved and >90% when >28 LNs were examined.19 Similarly, a population-based study of 20 789 cases identified using the Surveillance, Epidemiology, and End Results database found that examination of 20 LNs should allow accurate staging in >90% of cases.30

Recently, Kluth et al33 ,34 developed predictive models to assess the minimum number of LNs required for accurate staging in an individual patient based on tumour characteristics in the preoperative or postoperative setting. The authors propose that patients at a higher risk of LN metastases require examination of a larger number of LNs to accurately predict LN-negative status and this can be used to tailor the extent of PLND. For preoperative risk assessment, they developed their model on a cohort of 4770 patients whose PLND extent was at the discretion of the surgeon and subsequently validated it on 3595 patients who underwent an extended PLND.33 Using their model, high-risk patients in the validation cohort required examination of at least 10 LNs to achieve 80% probability of being truly LN-negative.31

One of the key issues when interpreting the literature on this topic is the lack of uniformity between studies, both in terms of the extent of PLND and the method of pathological evaluation. In fact, even the definition of what constitutes an extended PLND varies between studies. Importantly, the LN count is not solely a reflection of the anatomical extent of dissection, but is also influenced by pathological processing methods and individual anatomical variation. The latter variable is clearly illustrated by the marked heterogeneity between individuals seen in anatomical autopsy studies,31 ,32 as well as clinical studies where the number of LNs retrieved in an individual case ranged from 1 to 40.18 ,29 ,30 A survey conducted by the International Society of Urological Pathology (ISUP) also found significant variability among genitourinary pathologists in the number of LNs they usually identify.35 The majority of pathologists (86%) reported finding <10 LNs on average, with only 8% and 3% recording an average LN yield of 11–15 LNs and >15 LNs, respectively.35 Although the extent of PLND among respondents was not known, these findings suggest a lower average LN yield in routine clinical practice when compared with LN counts proposed in the literature.35

Based on the heterogeneity of studies in terms of surgical and pathological approaches, in addition to anatomical variation between individuals, it is difficult to define the adequacy of staging using LN yield. Additionally, in some studies, the mean/median yield across the study cohort was significantly lower than their proposed minimum LN yield, highlighting the current difficulty with applying these thresholds to an individual patient in routine practice (table 1). Importantly, a consistent approach to the pathological handling and evaluation of PLND specimens is required in order to evaluate LN yield as a quality indicator across institutions. Furthermore, collaboration between surgeons and pathologists is essential to optimise staging and maximise benefit for the patient. For example, submission of PLND zones in separate containers can aid gross pathological assessment,1 ,36 while a low nodal yield <10 LNs should, in our opinion, always trigger further evaluation of the specimen by the pathologist.

Pathological processing of PLND specimens

The pathological examination of surgical specimens plays an important role in accurate staging and can influence the final LN count at PLND. Currently, however, there are no established guidelines as to the optimum method of examining PLND specimens, and the approach may vary considerably between individual pathologists and institutions. This is further compounded by a number of practical issues related to the identification and enumeration of individual LNs.

The standard approach to isolating LNs from any surgical specimen is to manually palpate and dissect them from the surrounding fat. Compared with other anatomical sites, however, pelvic LNs are often extensively infiltrated by fat and may have an elongated tortuous configuration, requiring meticulous dissection for accurate counting of LNs. Alternatively, some may prefer to serially slice the intact specimen for ease of tissue submission. This approach, however, makes enumeration of individual LNs very difficult, and is prone to inflation of the LN count due to ‘double-counting’ of multiple slices from a single LN as separate individual LNs (figure 1). This is particularly true given the fatty and tortuous nature of LNs at this location. Consequently, careful gross dissection of individual LNs is preferable for accurate LN retrieval. LN revealing solutions have been used to aid the identification of LNs, however experience with these techniques in PLND specimens is limited.37 ,38

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Grossly identified fatty lymph node was bisected and counted as one lymph node (A), however with serial slicing this could be interpreted as multiple (up to four) separate lymph nodes (B).

Even following careful gross delineation of individual nodes, microscopic counting of LNs is not always straightforward and is dependent on subjective interpretation in certain scenarios. The presence of an architectural framework such as a capsule and subcapsular sinus should be identified to differentiate small LNs from lymphoid aggregates. However, microscopic sections of a single fat-infiltrated LN can yield the appearance of separate LNs, while truly separate small LNs which are not grossly annotated may exist in the fat adjacent to a larger LN. In these cases, the LN number is open to interpretation and pathologists must exercise their own judgement based on the combination of gross and microscopic features.

All grossly identified LNs should be submitted in their entirety for histological examination, except for those with clearly identifiable metastatic deposits where sampling will suffice. However, following isolation of all palpable LNs, there is currently no consensus regarding the routine evaluation of the remaining fatty tissue, which may contain microscopic, impalpable LNs.35 This may be important as a poor correlation between LN size and the presence of metastatic disease has been demonstrated.39 Total submission of PLND tissue has been shown to increase the LN yield39 ,40 although the clinical impact and cost-effectiveness of this strategy remains uncertain. In one study, submission of all the remaining tissue following isolation of palpable LNs increased the mean LN yield from 3.8 to 10.8 while requiring an additional 3–5 blocks.40 ,41 Metastatic prostate carcinoma was identified in eight cases, including one with a solitary metastasis in an impalpable LN. This suggests that all embedding the tissue may impact the final LN stage in a small number of cases, although the cost: benefit of routinely using this approach has not been established. Montironi et al42 proposed the use of large format histology, or mega cassettes, to facilitate complete embedding of PLND specimens to increase nodal yield and detection of metastases. However, specimens in the study were serially sliced and submitted, a technique which can be prone to overestimation of the LN count. The authors identified separate LNs by correlating the microscopic size and contour of individual sections, although this may not be reliable given the often irregular and meandering configuration of LNs at this site. Following isolation of palpable LNs, an ancillary technique such as differential inking could be applied to enable the use of mega cassettes for the examination of multiple LNs. Issues related to the processing, storage and cost of using mega cassettes in this setting may pose limitations and would need to be considered prior to adopting such an approach.

Although there is evidence that total submission of PLND tissue may alter nodal staging in a small number of cases, the data are limited and further study is required to determine if this approach is justified in all cases. Current data sets from the Royal College of Pathologists43 and the College of American Pathologists44 do not specifically recommend all embedding of PLND specimens. However, the pathologist does have a clear responsibility to ensure a thorough examination of the tissue for accurate staging. All palpable LNs should be carefully isolated and submitted, along with any firm areas which may potentially represent nodal tissue. Serial sectioning of larger nodes along their short axis will increase the surface area examined histologically, while smaller LNs can be bisected. Individual judgement may be used regarding submission of the remaining fat and will be influenced by the amount of fatty tissue present, the number of LNs identified in the specimen and the experience of the prosector examining the specimen. If the tissue is not entirely submitted, a low nodal yield following microscopic examination should always prompt the pathologist to examine the gross specimen and consider submission of additional tissue.

Frozen section (FS) analysis of pelvic LNs was previously advocated for intraoperative staging to guide surgical management.45 However, evidence that radical prostatectomy confers a significant survival benefit in patients with LN metastases46 ,47 has made this procedure redundant, and intraoperative FS analysis of pelvic LNs is no longer recommended.1

Metastatic tumour burden and prognostic histopathological parameters

The current (2009) tumour, node, metastases (TNM) cancer staging manual groups all node-positive patients into a single category (pN1).48 However, there is evidence that the nodal burden of disease can further stratify patients in terms of prognosis and outcome. Several studies have demonstrated superior long-term outcomes in patients with limited nodal involvement as compared with those with more extensive nodal disease.8 ,10 ,49–52 As such, LN-positive patients may be further stratified on the basis of the number of positive LNs, with studies suggesting one or two positive LNs as the cut-off for risk grouping.10 ,49 ,50 This makes careful enumeration of LNs by the pathologist critical, as any sampling method prone to ‘double-counting’ may falsely elevate the number of metastatic nodal deposits reported.

Nodal cancer volume, or the size of the largest metastatic tumour deposit, has also been shown to have prognostic significance.53–57 In fact, the size of the largest metastatic deposit may be more important than the number of positive LNs.53 ,58 Fleischmann et al57 used a cut-off of 10 mm for the size of the largest metastasis and found this to be the strongest independent predictor of recurrence-free survival, disease-specific survival and overall survival in their series of 102 patients. As such, there was consensus among genitourinary pathologists at the 2009 ISUP conference that the diameter of the largest metastatic deposit should be routinely evaluated and included in the final pathology report.35 This will often comprise a microscopic measurement, however the gross measurement will be more accurate for large tumour deposits.

A more favourable prognosis for patients with micrometastatic disease has been reported but the definition of micrometastases is not uniform among studies and the term has been applied to tumour deposits of varying sizes.55 ,57–59 In organs such as the breast, the definition and significance of small metastatic deposits including micrometastases and isolated tumour cells is well established and incorporated into the TNM staging system.48 However, in relation to prostate cancer no such definitions exist, and risk grouping based on size is not part of the current staging system.

Extracapsular extension (ENE) of tumour beyond the LN was reported to have prognostic significance in one early study,60 however subsequent studies have not found ENE to be an independent prognosticator in patients with prostate cancer with LN metastases.53–57 ,61 ,62 Similarly, Boormans et al56 found a Gleason Score >7 within the LN metastasis to be a predictor of survival. Other studies could not corroborate the independent prognostic value of nodal Gleason Score when controlled for other factors.62 ,63 Consequently, it is not necessary to routinely report these parameters.

Histological step-sectioning and the use of immunohistochemistry

As previously discussed, gross sectioning of LNs at short intervals increases the surface area available for histological evaluation to maximise the detection of metastatic deposits. A number of studies have also examined the utility of additional H&E stained step sections and cytokeratin immunohistochemistry (IHC) as an aid to the identification of small metastatic deposits in pelvic LNs.64–70 Engvad et al64 performed extensive step-sectioning and cytokeratin IHC of all LNs in 169 patients initially staged as pN0 following routine pathological examination of the PLND specimens. This resulted in upstaging of five patients (2.38%), with metastatic deposits ranging in size from 0.3 mm to 2.2 mm (mean 2.2 mm) identified on step sections. IHC did not provide additional information. Based on the authors' estimations, the cost of their extended LN evaluation was more than three times the cost of standard processing. In another series of 180 patients initially staged as pN0, occult tumour cells were identified in 13.3% following the application of IHC, and were associated with decreased survival.65 However, most studies have generally reported a low yield in terms of patient upstaging following the application of these ancillary studies.66–69 Recently, Kehr et al70 did not find any benefit from the use of IHC in patients who had received neoadjuvant hormonal therapy. Overall, given their cost implications, these ancillary techniques do not yield sufficient benefit to justify their routine application in the examination of pelvic LN specimens.

A further area under investigation is the utility of additional molecular techniques to improve the detection of LN metastases. For example, expression of genes such as PSA, PSMA and prostate cancer gene 3 using quantitative real-time reverse-transcriptase PCR analysis has been proposed as a useful adjunct in the identification of LN metastases.71–75 At this time, these techniques remain experimental and their utility in clinical practice remains to be established.

LNs within the anterior fat pad and periprostatic tissue

The anterior fat pad (AFP) comprises a mass of fibrofatty tissue that is often excised during prostatectomy to facilitate visualisation of the surgical anatomy, in particular for dissection of the apex during robotic assisted radical prostatectomy.76 A number of studies have demonstrated that this fatty tissue may contain LNs, and these can occasionally harbour metastatic disease.77–82 Published series have identified LNs in the AFP tissue in 5.5–17% of cases, with metastatic tumour in 0.94–2.5% of cases and consequent upstaging in 0.39–2.5% of cases. These findings were based on entire submission of the AFP tissue in the majority of studies.77–79 In most cases, tumours that metastasised to the AFP LNs had intermediate-risk or high-risk features at prostatectomy and were often anteriorly located in the prostate.77–79

Currently there are no guidelines regarding the pathological handling of these specimens. Given that LNs within the AFP may occasionally be the only site of metastatic disease, if this tissue is dissected at surgery, it should be routinely submitted for pathological examination (figure 2). Aning et al77 recommended complete submission of the AFP tissue, as the majority of their specimens could be submitted in three blocks. However, although the median maximum diameter of the AFP specimens in their study was 40 mm, they also had cases measuring up to 115 mm, which would require a larger number of blocks for total embedding. Conversely, other studies have suggested limiting the pathological examination of AFP tissue to patients with intermediate-risk or high-risk features.80 ,81 We propose examination of AFP tissue similar to PLND specimens, with isolation and submission of all palpable LNs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Metastatic tumour in a lymph node identified within the anterior fat pad tissue. This represented the only site of metastasis in the case.

Infrequently, LNs are also found in periprostatic fat, particularly around the base and seminal vesicles, and they are mostly detected incidentally when examining histological sections of the prostate.83–85 For example, Deng et al83 found periprostatic LNs in 0.8% of 2663 radical prostatectomies, with 0.3% of cases harbouring metastatic tumour. Given the current single category staging system, patients with metastases to AFP or periprostatic LNs are staged as pN1, however their clinical significance in isolation has not been clearly established.