Keywords

1 Introduction

Total mesorectal excision (TME) using meticulous dissection along surgical planes remains the mainstay of curative treatment options for rectal carcinoma. In patients treated with surgery alone, prognosis depends primarily on the pathological stage of the tumour. Cuthbert Dukes originally proposed a classification system for colorectal cancer in 1935, which recognised the importance of pathological nodal status. Till now, this remains one of the most important factors in predicting future outcome both in terms of local recurrence and disease-free survival (DFS) and overall survival (OS), although the Dukes’ classification has been superseded by more precise tumour, node, metastasis (TNM) systems, which are used by the American Joint Committee on Cancer (AJCC) and the Union for International Cancer Control(UICC) for rectal cancer [1,2,3] (see Table 65.1).

Table 65.1 TNM classification (version 8, 2017) with sub-classifications
Full size table

TNM staging in rectal cancer is based on the extent of the primary tumour (pT) and its relationship to the muscularis propria, the absence or extent of nodal involvement (pN0), the number of the involved regional lymph nodes (pN1 and pN2) and the presence or absence of distant metastases (M).

Patients with histologically node-positive rectal cancer tend to be those with more advanced T sub-stage [4] and more frequently involved circumferential resection margins. Population data suggest that approximately 18–20% of patients with pT3 tumours have four or more involved lymph nodes (pN2) [5]. Hence, the various combinations of T-stage, N-stage and M-stage have been combined into stage groupings (stage I to IV) that give a more defined risk stratification and determine prognosis.

However, a staging system is only valuable if the results can reliably reproduce an accurate prediction of outcomes (in terms of local and distant metastases). This prediction can then guide postoperative treatment decisions regarding the need for additional adjuvant treatment, and direct the type and intensity of follow-up surveillance.

In this chapter, we will discuss the conflicting findings regarding lymph nodes (LNs) in rectal cancer when surgery alone is performed and neoadjuvant treatments are delivered. We also examine the issues surrounding the importance of detecting and examining as many lymph nodes as possible after radio(chemo)therapy to define ypN status, review some of the challenges that arise in resecting and subsequently identifying pathological lymph nodes, and provide recommendations for the examination and reporting of surgically resected lymph nodes in rectal cancer after preoperative radio(chemo)therapy.

2 Factors That Influence Lymph Node Count in the ‘Non-neoadjuvant Setting’

However, the number of LNs retrieved (and hence the number of nodes that can be examined) depends on many factors including patient’s age [6], gender, obesity [6], histology (signet ring or other), other patient pathology, tumour size and localisation [7], differentiation grade [6], lymphoid reaction [8], RAS/p53 status, (MSI) [9] and the use of preoperative chemoradiation (CRT) [6, 10, 11]. There is also a significant variation in the total lymph node count (LNC) between individual hospitals, which may reflect the surgeon or pathologist [12].

The effectiveness of the host immune system may be relevant. It is recognised that MSI is associated with several characteristic clinicopathological features, including increased number of LNs in the specimen [13], but the number of LNs (if negative) may simply reflect a well-functioning immune system. MSI is rare in rectal cancer [14].

Some series have shown that older patients have fewer lymph nodes retrieved in specimens. This finding could demonstrate either aging of the immune system with LNs undergoing a process of involution [15]. Alternatively, this finding could simply reflect more conservative surgery in older patients due to the presence of more comorbidities compared to younger patients [16, 17]. Fewer lymph nodes in older adults however have not been confirmed in other studies [18].

The number of nodes reported may also relate to the quality of surgery and represent a measure of the quality of surgery [19]. A total mesorectal excision (TME)in the muscularis propria plane (i.e. poor plane of surgery achieved) is defined by having little bulk to the mesorectum with defects down onto muscularis propria, very irregular circumferential resection margin, or both. This finding may mean the surgeon has left lymph nodes within the patient, because more nodes are found when the surgical quality is good.

Most trials, although nominally performing TME, are bedevilled by the lack of good quality mesorectal excisions – only 50% had TME within the mesorectal plane in CR07 [20]. Hence, in the pre-TME era of surgery, involved mesorectal lymph nodes would have been left behind within the patient due to inadequate oncological clearance of pelvic local disease. These nodes would almost certainly have played a role in disease ‘recurrence’ and probably account in part for the substantially higher pelvic recurrence rates documented in pre-TME era studies that have formed the basis for current adjuvant and neoadjuvant treatment policies.

A French study assessing quality indicators found the examination of ≥12 LNs was associated with five separate factors: ECOG score, two tumour characteristics (location and stage) and two hospital characteristics (administrative district and volume of colorectal cancer (CRC) procedures) [21].

So, ultimately, the role of the surgeons and the pathologists is probably the predominant factor in determining lymph node yield [22, 23].

Intriguingly, the number of nodes examined after surgical resection with no neoadjuvant therapy is significantly associated with relapse and survival – among those patients where nodes are not involved [24]. However, this study did not show the results improved in patients with involved nodes, when more nodes were examined. So the surgical procedure itself is unlikely to be the major contributor to this difference. Using data from 1792 patients with T3/T4 or node-positive rectal cancer within the intergroup 0114 trial, Tepper recommended examination of 14 or more nodes as an appropriate number for adequate assessment of the nodal status [24].

Recommendations in the era before preoperative SCPRT or CRT was routinely given often stated that a minimum of 12 [25, 26] nodes needed to be studied to define nodal status accurately [24]. Some recommended a smaller but safe minimum number [27,28,29]. Others have suggested that the number of patients classified as node-positive will continue to increase until 17–20 nodes are found [30].

The optimum lymph node yield for tumour staging following surgery for rectal cancer remains controversial. Using data from 63,381 patients in the surveillance, epidemiology and end resulted (SEER) database, a recent study showed evaluation of 18 lymph nodes were required for those treated without neoadjuvant radiotherapy and 16 nodes for those treated with RT to prevent stage migration in rectal cancer [31].

2.1 Prognostic Impact of Lymph Node Metastases in the ‘Non-neoadjuvant Setting’

In patients treated with surgery alone, there is a very clear relationship between nodal status and outcome in rectal cancer [32]. In general, rectal cancers tend to have fewer nodes than colonic tumours. The number of lymph nodes detected and examined in the mesorectum varies (6–60) between individuals [33], but some studies suggest the more lymph nodes to be examined the better the prognosis. The finding of positive lymph nodes has been considered as a biomarker for the risk of both local recurrence [34, 35] and distant metastases in rectal cancer – although the risk of local recurrence may be overestimated if a good quality TME is performed and all LNs are actually removed rather than being left in the patient [20, 36].

The overall number of resected/reported nodes (whether involved or not) and proportion of involved lymph nodes are also independent prognostic factors for survival [25, 37,38,39,40].

Several studies have shown that the precise number of lymph nodes involved, i.e. 1, 2–3, 4–6 or >7 positive nodes [5, 41] after surgery alone determine prognosis. The 5-year survival rates range according to the number from 81% to 26% [41]. Prognosis is much worse when the validated cut-off of four or more lymph nodes is involved (pN2). So clearly it is important that the number of involved nodes is accurately documented.

3 Lymph Node Ratio (LNR)

Individuals have a wide range of lymph nodes detected, which can impact on the validity of the lymph node status if insufficient nodes are examined. To overcome the uncertainties associated with simple number-based LN classification, a lymph node ratio (LNR), i.e. the ratio of the number of lymph nodes showing malignant cells compared to the overall number of examined lymph nodes) – was potentially viewed as a prognostic factor by Berger to circumvent this problem in colon cancer [42]. A higher LNR in colon cancer is associated with a worse DFS [43, 44]. LNR offers additional evidence on prognosis in addition to the prognostic information derived from the 7th edition of the TNM Classification of Malignant Tumours, and is an independent risk factor for OS [45]. A review of 16 studies with a total of 33,984 patients with stage III colon or rectal cancer [46] showed the prognostic value from the LNR outperformed the simple number of positive nodes. The total number of positive nodes dropped out as an independent prognostic indicator when LNR was included in the regression model in seven of the studies, and in a further four studies the number of positive nodes did predict outcomes but with a lower significance than the LNR.

3.1 Location of Lymph Nodes and Metastases in Relation to Radiation Fields

There are numerous potential sites of involved nodes in rectal cancer, i.e. mesorectal nodes, inferior mesenteric nodes up to the origin of the left colic artery, internal iliac and obturator nodes, and less frequently, inguinal and external iliac nodes. It is unusual for the pathologists to specify the origin of the positive node.

3.1.1 Mesorectal Nodes

In a series of 121 patients with locally advanced rectal cancer, who underwent chemoradiotherapy, the majority of lymph nodes were found along the major supplying vessels, but 77% of the involved lymph nodes were found in the mesorectum [47]. The majority of mesorectal nodes are located in the upper-posterior part of the mesorectum [48, 49], and a high proportion in the outer layer of the mesorectum [50]. In one study, 68% of mesorectal lymph nodes were above the peritoneal reflection [51]. The distribution in relation to the tumour was 53% above, 36% adjacent to and only 11% below the tumour itself [51].

It is easy for the surgeon to come out of the correct plane posteriorly and laterally, particularly before reaching the level of the tumour. An incomplete specimen will therefore leave lymph nodes behind within the patient.

For low cancers within 5 cm of the anal verge, the finding of involved mesorectal nodes also confers a risk of involved lateral pelvic lymph nodes in the region of 10–15%. Since these nodes are not routinely resected in the West, a higher risk of systemic disease may be observed in the case of low tumours with ypN+ status.

3.1.2 Lateral Pelvic Sidewall Nodes

There is a discussion whether external and common iliac LN should be defined as M1. Involved nodes outside the MRF (which are not usually resected in the West) appear uncommon in the mid and upper rectum, and are found mainly in patients with distal rectal cancers close to the anal verge with visible nodal metastases within the mesorectum.

3.1.3 Apical Lymph Nodes

The significance of metastatic apical nodes around or at the root of the inferior mesenteric artery remains controversial and may relate to whether a surgeon performs a ‘high-tie’ or not. Apical lymph node status in the past was considered an important prognostic factor and included as a marker for the pN3 category in the 1987 version (TNM 4) of the AJCC staging system, but this concern has largely been replaced by defining the total number of positive lymph nodes or the lymph node ratio since publication of the 5th edition of TNM staging system.

The prognostic effect of apical lymph node metastasis may have been diluted by the number of positive nodes evaluated. Huh et al. studied the prognostic value of lymph node distribution in 1205 patients with sigmoid and rectal cancer, none of whom received preoperative treatment. Lymph node metastasis at the origin of the inferior mesenteric artery was found to be an independent prognostic factor for 5-year DFS and OS. Subgroup analysis further confirmed the prognostic value in predicting 5-year DFS and OS in node-positive sigmoid colon and rectal cancer [52].

Apical nodes are usually located above the L5/S1 junction and hence lie outside the conventional pelvic radiation field. For this reason, apical lymph node metastasis in patients receiving neoadjuvant CRT may be similar to those not receiving it. If the superior border is lowered to S2/S3 as currently recommended even more nodes may be left unirradiated – but as this omission may not increase the risk of local recurrence, it will be difficult to quantify its impact on subsequent outcomes.

4 Size of Lymph Nodes

Lymph nodes may be very small and still contain metastatic disease. Early pathological studies suggested that 30–50% of the positive nodes and the majority of negative ones are less than 5 mm in diameter [53, 54]. More recent larger studies continue to confirm these findings [55], but others report one-third of involved rectal cancer nodes are often <5 mm and many may not even be visualised on preoperative imaging. In colon cancer, there is limited clinical relevance for lymph nodes smaller than 3 mm in colon cancer as only 2.8% are likely to be involved, but 45% of metastatic nodes were 3–5 mm in size [56]. In a recent meticulous study in rectal cancer, 95/334 (28%) of positive nodes were ≤3 mm in diameter [51]. Others have reported that involved LNs had significantly greater long and short diameters (P < 0.001) than negative LNs, although this is not a reliable feature as the majority of involved nodes were ≤5 mm [57]. Positive predictive values of the size criteria are even less accurate for internal iliac and obturator lymph nodes lower than that for mesorectal lymph nodes [58].

5 Impact of Neoadjuvant Treatment on Lymph Node Counts

Neoadjuvant CRT and SCPRT produce major changes both in the cancer and in surrounding structures, i.e. fibrosis, deep stroma alteration, muscular wall thickness, muscle disarrangement, tumour necrosis, calcification, and inflammatory infiltration, which can be identified in the histopathological specimen. Therefore, the initial preoperative cTNM stage may be altered/modified by any preoperative therapeutic intervention (whether chemotherapy, SCPRT or CRT) – in which case histopathological pTNM is given a ‘y’ prefix [59].

There is a widespread belief that it is more difficult to harvest a sufficient number of lymph nodes in patients who undergo preoperative neoadjuvant SCPRT or CRT [60,61,62,63,64]. On imaging after CRT, most lymph nodes become smaller, and many disappear completely [65], which may explain some of the differences. Reduction in the size of initially enlarged lymph nodes after neoadjuvant therapy may be a useful prognostic factor [66]. Also pathologically after CRT, lymph nodes are reported to be smaller with depletion of lymphocytes and replacement with stromal fibrosis, and hence more difficult for the pathologist to detect [10, 67,68,69]. A recent retrospective study examining total lymph node count in surgically resected cT3-4 rectal cancer showed that patients who had preoperative CRT had significantly fewer lymph nodes than those who had surgery alone (12.4 vs. 21.1, P < 0.0001) [70]. Others have shown that significantly fewer tumour-infiltrated lymph nodes are detected in patients who received preoperative radiochemotherapy [71, 72].

A recent meta-analysis of 34 articles (including 37 datasets) concluded that patients undergoing CRT had, on average, 2.1 less lymph nodes resected compared to those who received no neoadjuvant treatment [11]. Approximately one-third of patients with neoadjuvant treatment had less than 12 LNC [7].

There is wide variation in the timing of surgery (4–12 weeks) after SCPRT and CRT due to patient/surgeon choice, recovery from treatment and/or waiting list issues. Population studies [73] and a meta-analysis [74] suggest that longer intervals up to a maximum of 12–15 weeks are associated with an increased chance of achieving a pCR at surgical resection. Prospective trials have been performed randomising between 6 and 12 weeks [75], and 7 and 11 weeks after CRT, respectively [76]. Results published report the number of nodes harvested in the latter trial were not statistically different whether the interval was 7 or 11 weeks median 15.7 versus 15.4, respectively (p = 0.74), nor was the rate of involved nodes statistically different [76]. Hence, the timing of surgery, once at least 6 weeks have elapsed, may not influence the number of nodes identified or the number of involved nodes.

6 Lymph Nodes After Neoadjuvant Chemoradiation Therapy: Nodal Status and Downstaging

Both radiotherapy and preoperative CRT may alter the surgical lymph node yield as well as its prognostic significance. The NSABP R-03 study of preoperative versus postoperative chemoradiation suggested a reduction of nodal status in patients receiving preoperative treatment (25% vs. 34%), although no information is provided on the number of lymph nodes found [77]. In the German CAO/ARO/AIO-94 study, histopathological examination after surgery revealed a lower percentage of tumours with positive lymph nodes in the preoperative chemoradiation arm compared to those who proceeded straight to surgery – 29% versus 44%, respectively [78].

Correlation of the degree of tumour regression with lymph node status has also been studied [79, 80]. Using the Dworak system, response to CRT was significantly associated with nodal stage. Patients without nodal involvement (N0) had 66% of positive tumour responses (TRG 4), while individuals with nodal metastases had an inferior response to CRT (TRG 0, 1 and 2) 35% N1 and 14% for N2 (P = 0.007). In this study, nodal status was independently associated with a poor response to preoperative CRT, even after adjusting for tumour stage, age and gender (OR 0.02, 95% CI 0.0009–0.67) [79]. Hence some investigators categorised poor response according to Rodel and based on TRG 0 and 1 on the Dworak scale have examined lymph node negative subsets [81] and found that poor TRG was the strongest prognostic factor in patients with ypN0.

However, data regarding the role of LNR in locally advanced rectal cancer after CRT remain sparse but as above appears an independent prognostic factor [82,83,84,85,86], one recent study suggests LNR may provide a more effective prognostic marker than either the ypN stage, or circumferential resection margin following CRT [87]. The clinical application of LNR is limited - due to variability in cut-off values; ratios between 0.01 and 0.16 have been proposed.

7 Lymph Nodes After SCPRT (5 × 5Gy) and Immediate Surgery or a Delay to Surgery

The timing of surgery and hence pathological assessment is important after SCPRT. The Polish trial and the TROG-0104 trial compared SCPRT and CRT and showed the interval to pathological assessment influences the rate of pCR, node positivity and CRM positivity [88,89,90]. No reduction in tumour stage or nodal classification is expected if the interval to surgery is less than 10 days, yet some downstaging is observed if surgery is delayed for more than 10 days [91, 92].

In a further Polish study, 154 patients were randomised to SCPRT with surgery either 7–10 days or 4–5 weeks after the end of RT [93]. More downstaging was seen after the longer interval. The Stockholm III trial randomised between SCPRT with immediate surgery (SCPRT), short-course RT with surgery delayed 4–8 weeks (SCPRT-delay) or long-course conventionally fractionated radiotherapy with surgery delayed 4–8 weeks. In an interim analysis, ypN+ status was recorded in 37/118 (31.4%) of SCPRT patients treated with SCPRT, 25/120 (20.8%) with SCPRT-delay and 18/65 (28%) of patients treated with long-course RT [94].

In a later evaluation of 462 of 545 randomised patients, where specimens were available for reassessment, there was an increase in downstaging in those patients where the interval to surgery was extended to 8 weeks compared immediate surgery within 7–10 days [95], the pCR increased from 2% to 13% (p = 0.001) see Table. A small Dutch retrospective study supports this [96].

The presence of residual ypN+ status after CRT indicates lack of response to CRT, but ypN+ status after 5 × 5 Gy RT and immediate surgery will not have sufficient time to express response. Hence, the histopathology after SCPRT will include both radiosensitive and radio-resistant tumours [97].

Some recent reviews have suggested there is an uncoupling of the relationship between LNC and outcomes after CRT such that the LNC no longer confers any prognostic information [63, 98].

Fewer lymph nodes may be associated with more downstaging as a result of better response to CRT with more favourable pathologic features and better oncologic outcomes rather than inferior quality of surgery and pathology. Habr-Gama et al. reported a study of 281 patients with relatively small early tumours showing that patients with ypNx after CRT patients had better 5-year disease-free survival (DFS) than patients with ypN0 and ypN+ (74% vs. 59% vs. 30%, respectively, P  <  0.001) [99]. However, with larger more advanced tumours, one of the present authors found the number of nodes after CRT was important and if the patient had node-negative histology, and at least three nodes recovered, they had better long-term survival than patients in whom two or less nodes were recovered or with positive nodes [100]. Both of these studies suggest the more nodes found the better.

The likelihood of finding node-positive disease increases with LNC, while after SCPRT and CRT, no increase was noted once 12 and 18 examined nodes were examined, respectively [98]. In contrast, examination of large numbers as in the 2006–2011 National Cancer Data Base, which included 65,271 patients with clinical stage I–III low rectal cancer who underwent a proctectomy provided different results. The 5-year OS for patients with a LNC <12 and >12 in the entire cohort were 77.2% and 81.2%, respectively (p < 0.0001), and for patients undergoing neoadjuvant CRT 79.2% and 83.5%, respectively (p < 0.0001) [12]. Also a large national cohort study in Denmark also suggested that LNC was associated with outcomes whether the patient received CRT or not [101].

8 The Relevance of Nodal Status to Decisions on Adjuvant Treatment

Stage III rectal cancer patients with persistent ypN+ after preoperative CRT (i.e. poor responders to CRT) represent a high-risk group [102], with higher mortality than those patients with pN+ treated with postoperative CRT (median overall survival = 62 vs. 87 months, P = 0.0002) [103]. Hence, a pathological report that confirms the surgical specimen contains involved metastatic lymph nodes (UICC stage III) has therefore usually led clinicians to offer the patient postoperative adjuvant chemotherapy (if performance status, frailty and co-morbidity allow) despite the fact that chemotherapy is toxic and costly, with little firm evidence of benefit after CRT.

In contrast, ypN0 status provides much more variable outcomes. NCCN guidelines for colon cancer recommend the use of postoperative chemotherapy for patients with node-negative disease and with less than 12 nodes identified in the surgical specimen [104]. The rectal guidelines recommend chemotherapy after CRT for all patients irrespective of histopathology [104].

Higher lymph node yields are associated with an increased stage III rate and have been shown to predict better survival in both node-positive and node-negative disease [40, 101]. Others dispute this finding [105] particularly if a major histopathological response is reported [106].

Conclusion

We recommend the identification and examination of as many lymph nodes as possible after resection of rectal carcinoma to have an accurate assessment of the number of involved lymph nodes and achieve precise staging. Very few nodes <3 is clearly insufficient, but after SCPRT and CRT no increase in positive nodes is noted for more than 12 and 18 examined nodes, respectively. For the most accurate nodal staging in rectal cancer, surgical resection and subsequent pathological evaluation should concentrate on the mesorectum in close proximity to the tumour, particularly posteriorly and superiorly and along the track of the superior rectal artery. Many lymph nodes will be small (<3 mm in diameter) particularly after CRT, but should not be overlooked. A minimum of 12 nodes should be aspired to even after CRT. The precise number of LN and the LNR should be balanced against the quality of the mesorectal specimen in making decisions regarding further treatment.