Ratio of Metastatic Lymph Nodes to Total Number of
Nodes Resected is Prognostic for Survival in Esophageal
Carcinoma
Clive J. Kelty, MB, ChB, PhD, FRCSEd, FRCS(Gen Surg),* Catherine W. Kennedy, RMRA,†
and Gregory L. Falk, MBBS, FRACS, FACS*†
Introduction: The role of the number of metastatic nodes in esophageal cancer surgery is of interest. We assess predictors of survival after oesophagectomy for esophageal and gastroesophageal junction malignancy.
Methods: Prospective data of consecutive patients undergoing oesophagectomy and systematic lymphadenectomy between 1991 and 2007.
Results: Of 224 patients, 148 patients (66%) had adenocarcinoma, 70 (31%) squamous cell carcinoma, and 6 (2.6%) were other tumor types. Five-year survival was 43% with hospital mortality of 3.5%. Locoregional recurrence occurred in 14%. The total number of affected nodes significantly reduced survival (four or more meta-static nodes). Further analysis of the ratio of nodes affected to the total number resected showed a significant decrease in survival as the percentage of positive nodes increased (p⬍ 0.001).
Conclusions: Patients undergoing surgery for esophageal cancer should be staged according to a minimum total number of metastatic lymph nodes and ratios because this more accurately predicts sur-vival than current staging systems.
Key Words: Esophageal cancer, Esophagectomy, Lymphadenec-tomy, Staging.
(J Thorac Oncol. 2010;5: 1467–1471)
A
denocarcinoma of the esophagus is increasing in inci-dence more rapidly than any other cancer in the Western world.1The mainstay of treatment for this and squamous cellcarcinoma remains surgical, namely esophagectomy, despite advances in other therapies. However, the overall prognosis
for these patients is dismal, with survival in most countries around 10%. Outcomes after surgery are marginally better with reported 5-year survival rates averaging 25%.2– 4 It is
known that these tumors spread radially via the lymphatic system, thus resection of the primary tumor and lymphade-nectomy are advocated.5The rationale for lymphadenectomy
is primarily based on the Japanese experience in gastric cancer. However, the evidence for this in esophageal carci-noma is less clear and remains a controversial issue.6
Intu-itively, the more radical the resection, the more lymph nodes (LNs) will be removed, and this will improve staging of the disease and possibly affect prognosis, either by reducing or delaying locoregional recurrence, or increasing the disease free survival. However, the extent of lymphadenectomy is not standardized among surgeons, and there is a lack of random-ized trials assessing this issue.
The number of LNs removed during lymphadenectomy is of value, in particular the presence of LN metastases, because it seems that there is prognostic significance in the number of metastases This may have treatment implications. According to the Sixth International Union Against Cancer tumor node metastasis (TNM) classification (which is also uniform with the sixth edition of the American Joint Com-mittee on Cancer staging), any regional LNs with metastatic involvement were staged as “N1.”7,8 However, this binary
grading was controversial because there is increasing evi-dence to suggest that the number of metastatic nodes, rather than positive or negative metastatic nodes, may affect prog-nosis. It has been shown in several studies that increasing numbers of resected nodes with metastatic deposits is one of the most important prognostic factors; there is a significant reduction in survival after curative oesophagectomy for both squamous cell carcinoma and adenocarcinoma.9 –20This was
addressed by the seventh edition of the American Joint Committee on Cancer Staging Manual that was published in January 2010.21 The new classification now recognizes
dif-ferent nodal classification (N1-3) based on coarse groupings of the number of nodes involved (0, 1–2, 3– 6, and ⬎7, respectively).
It has also been suggested that it is not simply the number of nodes that is important but that the LN yield needs to be adequate for staging17–19; the numbers suggested range
from 10 to more than 18. This is in stark contrast to the sixth
*Department of Upper GI Surgery, Concord Repatriation General Hospital; and †Department of Surgery, University of Sydney, Strathfield Private Hospital Campus, Sydney, New South Wales, Australia.
Disclosure: The authors declare no conflicts of interest.
Address for correspondence: Clive J. Kelty, MB, ChB, PhD, FRCSEd, FRCS(Gen Surg), Department of Surgery, Doncaster Royal Infirmary, Armthorpe Road, Doncaster DN2 5LT, United Kingdom. E-mail: clive.kelty@bigfoot.com
Presented in part to Association of Upper Gastrointestinal Surgeons, Cardiff, September 2007 and Society of American Gastrointestinal and Endo-scopic Surgeons, Philadelphia, April 2008.
Copyright © 2010 by the International Association for the Study of Lung Cancer
TNM classification that only required six nodes.7 This was
not specifically addressed in the seventh edition; however, it states that an adequate lymphadenectomy requires between 12 and 22 nodes and advocates as extensive a lymphadenec-tomy as possible.
What remains less clear, however, is whether it is simply the absolute number of metastatic nodes that is im-portant or whether the ratio of affected nodes to nodes resected may be an independent prognostic factor.
We currently perform a systematic two-field lymphad-enectomy that has been reported previously.22Therefore, we
examined survival in a contemporary cohort of patients di-agnosed with esophageal carcinoma who underwent an en-bloc lymphadenectomy, and whether the extent of LN me-tastases, in particular the ratio of nodes affected compared with the total yield of resected nodes, had a prognostic effect on survival.
METHODS
A prospective database of patients who had carcinoma of the esophagus and gastroesophageal junction (excluding proximal lesions and Siewert type III adenocarcinomas) and who underwent subtotal oesophagectomy and systematic lymphadenectomy in a tertiary referral center between 1991 and 2007 was collected. Routine preoperative evaluation included upper gastrointestinal endoscopy with biopsy, com-puterized tomography scan, and selective endoscopic ultra-sound. Staging laparoscopy was also performed on a selective basis for junctional (Siewert II) tumors. Operative fitness was determined by clinical assessment, pulmonary function test-ing, arterial blood gas, and cardiac imagtest-ing, as clinically indicated. All patients fit for surgery and assessed as feasible for R0 resection despite evidence of nodal disease on imaging were offered surgical resection on an all-comers policy. Latterly in the series, patients were offered neoadjuvant therapy if the tumor was considered borderline resectable (borderline attachment to pleura, pericardium or aorta or bulky tumors and adjacent lymphadenopathy) on preopera-tive staging. This constituted three cycles of ECX (epirubicin, cisplatin, and capecitabine) in adenocarcinomas and two cycles of cisplatin and 5-fluouracil in squamous tumors. The surgical procedure comprised Ivor-Lewis oesophagogastrec-tomy and systematic two-field en-bloc lymphadenecoesophagogastrec-tomy. All patients had given prior consent for data to be collected on the database. We assessed demography, in-hospital mortality, the extent of resection, the number of nodes resected, those affected by metastatic deposits and survival. The follow-up protocol consisted of 3-monthly consultant review for 2 years, followed by 6-monthly review thereafter to 5 years and then annually. When it was not possible for face-to-face review, information was obtained from the patient’s primary physician. Further imaging was based on clinical grounds.
Data were analyzed with the method of Kaplan-Meier,23log-rank test, and Cox’s proportional hazards
regres-sion for survival analysis.24 RESULTS
Two hundred twenty-four consecutive patients were identified from the database.
There were 162 men (72%) with a median age of 65 years (range, 29 – 84 years). One hundred forty-eight patients (66%) had adenocarcinoma, 70 (31%) squamous cell carcinoma, and 6 (3%) were other tumor types (2 neuroendocrine tumors). Neoadjuvant therapy was given to 67 patients (30%). Of these, 13 with extensive squamous carcinoma were given concurrent 45 Gy radiotherapy. The median length of follow-up was 20 months (range, 1–174). One hundred eighty-four patients (82%) had been reviewed within the last 12 months (up to August 2008). The in-hospital mortality was 3.5%. A R0 resection was achieved in 88%, R1 (defined as tumor extending to the resec-tion margin, including circumferential margin) in 10%, and R2 in 2%. The AJCC stage distribution was as follows: stage 0, n⫽ 11 (5%); stage 1, n⫽ 38 (17%); stage 2A, n ⫽ 46 (20%); stage 2B, n⫽ 37 (16%); stage 3, n ⫽ 78 (35%); and stage 4, n ⫽ 14 (6%). Nodal disease was detected in 122 patients (54.4%). Overall 5-year survival as plotted by Kaplan-Meier analysis was 43%, with a median survival of 33.6 months (Figure 1). Patients with node-positive disease had a 5-year survival of 23% (me-dian, 31 months). Locoregional recurrence occurred in 14% after a median of 11.5 months. There was also a significant relation-ship in survival between both the TNM stage of disease (p⬍ 0.001; Figure 1) and increasing depth of tumor invasion (p ⬍ 0.001). However, there was no relationship between either histologic type of tumor (adenocarcinoma versus squamous cell;
p⫽ 0.66; Figure 2) and tumor site (mid, lower esophagus, and
gastroesophageal junction; p⫽ 0.96). There was no significant difference between patients who had undergone neoadjuvant therapy or surgery alone (p⫽ 0.98; Figure 3). With multivariate analysis, independent predictors for diminished survival were male sex, age older than 75 years, and residual tumor.
The median number of nodes resected in the specimens was 17 (range, 12–26). Analysis of the number of nodes affected by metastatic disease showed that there was a sig-nificant relationship between those having four or more nodes involved. Patients with less than four nodes involved had a median survival of 17.7 months, but four or more nodes
FIGURE 1. Patient survival according to AJCC (6th Edition) TNM staging.
involved gave a median survival of 9 months (p ⬍ 0.001; Figure 4). Further analysis of the ratio of nodes affected to the total number resected showed a significant decrease in sur-vival as the percentage of positive nodes increased (p ⬍ 0.001; Figure 5). In patients who had less than 20% of nodes involved, median survival was 27.4 months, falling to 22 months for 20 to 40% of nodes affected, 12.8 months for 40 to 60% of nodes involved, and when more than 60% of resected nodes were affected, survival was a dismal 6.5 months.
DISCUSSION
It is well recognized that there is an adverse effect on survival when there are metastatic nodes after esophageal resection. However, the extent of LN resection remains con-troversial because some series have reported similar out-comes after minimal resection compared with those advocat-ing dissections that are more extensive.4However, it remains
a justifiable principle of cancer surgery, as demonstrated in
other tumors, that more radical resections with LN clearance can confer a survival benefit.25It is based on this principle
that en-bloc resection of the esophagus is more likely to achieve a real R0 resection by not leaving positive LNs, and so results are likely to be better after surgery. The level of local recurrence will also be improved (as in this series). Extensive resection with lymphadenectomy has shown im-proved survival in reported series with 5-year survival rates of 40 to 50%,4,22 compared with a contemporary
meta-analysis of 25%.2
The sixth TNM staging system only took account of whether regional LN metastases are present or not, and this was translated into a stage grouping. This created a problem, because there was no provision for the absolute number of nodes affected, which has been shown in several series to predict survival more accurately.9 –20It also failed to measure
the completeness of the lymphadenectomy. The general con-sensus of these studies is that four or more metastatic nodes is prognostic, and this is borne out in our series, because there
FIGURE 2. Overall survival in patients according to histo-logic subtype.
FIGURE 3. Survival according to the administration of neo-adjuvant therapy versus surgery alone.
FIGURE 4. Survival according to absolute number of lymph node metastases resected.
FIGURE 5. Overall survival in patients stratified with in-creasing metastatic lymph node ratios.
was a significant reduction in survival after surgery in pa-tients who had four or more nodes affected. Based on this, the seventh revision of the TNM system to include subgroups of the N stage hopes to allow more accurate prediction of survival.
The sixth revision of the TNM classification states that a minimum number of six nodes should be sampled to allow classification of the nodal status.7 This is low, particularly
when it takes into account other sites, such as colon and stomach, which require 12 and 15 nodes, respectively. It should also be noted that these cancers also have different node groupings (N1-3). The seventh edition is less specific in the absolute number of nodes resected; however, it states that the degree of lymphadenectomy required is dependant on the T stage. In reality, this is unrealistic, because the staging is based purely on pathologic stage. It is difficult to justify the extent of surgery performed purely on clinical staging, because more extensive lymphadenec-tomy has been shown to provide a survival benefit,22 and
clinical staging is less accurate and so difficult to establish the “right” number and extent of nodal removal. It is simply necessary to be able to adequately remove LNs without excess morbidity and mortality.
A recent study by Bogoevski et al.18 has examined
whether the number of nodes resected should be higher. They suggested that a minimum number of 18 should be examined to allow accurate staging. In 255 patients who underwent oesophagectomy with curative intent, the numbers of nodes resected were stratified into three groups, less than 6, 6 to 18, and 19 and higher. This approach demonstrated that patients with either N0 or N1 disease with less than 19 nodes had similar survival rates. They concluded that it is probable that this could be explained by a degree of missed metastatic disease when fewer LNs are analyzed, as with low numbers there is a risk of sampling error (i.e., not finding metastatic nodes and grouping a patient as N0 when in fact the patient should be grouped as N1).
Bogoevski’s group further reviewed the ratio of LNs with metastatic involvement. Groups were identified with ratios of less than 11%, 11 to 33%, and more than 33%. This demonstrated significantly improved survival in patients with lower ratios and was further borne out by the observation that there was little difference in patients staged as N0 with 18 or fewer nodes analyzed (median survival 36 months) and in patients with N1 disease but more than 18 nodes resected and a low ratio of affected nodes (median survival 28 months). They concluded that the current nodal staging classification was insufficient, and that LN ratios should be considered in substaging within the pN1 classification. In our series, the median LN yield was 17. Further analysis of node yield during two separate time periods, namely the first (1991– 1998) and second (1999 –2006) half of the series showed average yields of 13 and 20 nodes, respectively. However, this was not associated with any difference in survival be-tween the two groups. There was no change of technique during the series, so this finding implies that the increased detection of nodes was due to improved pathologic reporting
rather than being due to more radical surgery or experience of the surgeons.
Other series have examined the prognostic effect of involved node to total node ratios,9,12,13,17,18 –20but they have
only examined an arbitrary-defined cutoff point (between 0.1 and 0.3). Two recent studies have tried to stratify this further by using two cutoff points to stratify patients to three risk groups and have shown that there are further prognostic differences.18,20 We have tried to be less arbitrary and to
substratify the ratios more fully in our series. This has shown significant differences between four different groups with cutoff points of 0.2, 0.2 to 0.4, 0.4 to 0.6, and more than 0.6. We would suggest that this is a more sensitive analysis, particularly when there are significant differences between these groups. However, it is important to recognize that a systematic lymphadenectomy will remove more nodes and allow more accurate staging using this system because when the total number of nodes removed increases, it will likely drive down the ratios to a better prognostic group.
At the time of analyzing the data and constructing this article, the seventh edition was not in print, and hence this discussion was based on the previous classification (as is all the current evidence). The new classification now recognizes different nodal classification (N1–3) based on coarse group-ings of the number of nodes involved. Although there is now recognition that increasing nodal involvement (and so tumor burden) is prognostic, there is still no requirement for assess-ment of the ratio of nodal metastases to total lymphadenec-tomy. Indeed, the manual states that the data do not support LN ratio; paradoxically, this article and others however do support this.9 –20 The main issue with LN ratio is that the
denominator (i.e., the number of nodes sampled) is variable and that this can dilute the prognostic value. In this study, the number of nodes resected was higher (range, 12–26) than previously required, and the recommendation from the AJCC is that ideally 12 to 22 nodes should be sampled.
Previous editions of the TNM classification ignored histologic type, but the seventh edition specifically separates adenocarcinoma and squamous cell carcinoma and stages them differently. The data indicate that squamous cell carci-noma has a poorer prognosis, but this was not borne out in this series (and paradoxically this trend was reversed, albeit not significantly).
One of the weaknesses of the new staging classification is that it is based purely on pathologic data of patients who have undergone surgery alone. It is increasingly common for patients to receive neoadjuvant therapy because data show that this is of benefit in patients with operable esophageal cancer.26Because many studies report varying definitions of
responses, the specific reason for this benefit is less clear; it may be due to downstaging of the primary tumor or from treatment of tumor cells that have already spread. Despite this reported benefit, our analysis did not show a significant difference between these groups.
Both the findings of no difference in overall survival based on the histologic type, or use of neoadjuvant therapy, adds further weight to the argument that it is the degree of
lymphadenectomy performed rather than the tumor site per se that is important.
We note that this cohort of patients were a heteroge-neous group, with different tumor types resected. However, the same conclusions can be drawn, because the pattern of LN metastases for the tumor sites is similar. LN involvement is a significant prognostic indicator of overall survival and disease-free survival, and as the number of involved nodes increases, it may suggest a greater tumor burden or more aggressive tumor biology, and so the likelihood of locore-gional or systemic recurrence could be assumed to be higher. However, this series shows that prognosis after surgery can be significantly affected by the ratio of affected nodes to the number of nodes resected after radical resection, and there-fore stage grouping of the LN is not simply positive or negative. In an age when it is increasingly important to stage patients correctly to make decisions on adjuvant therapy, inclusion into clinical trials, and inform patients, this type of prognostic information is vital to both physicians and patients alike. This study demonstrates that the accuracy of the current TNM classification is suboptimal and that there should be revision of the nodal reporting, taking into account the ratio of involved nodes to the total number resected. Therefore, we propose oesophagectomy with radical en-bloc lymphadenec-tomy should become the standard of care, because it is apparent that survival is likely to be improved after surgery. It also provides optimal staging and is therefore a better prognostic indicator of survival than current conventional pathologic reporting.
REFERENCES
1. Devesa SS, Blot WJ, Fraumeni JF Jr. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 1998;83:2049 –2053.
2. Jamieson GG, Mathew G, Ludemann R, et al. Postoperative mortality following oesophagectomy and problems in reporting its rate. Br J Surg 2004;91:943–947.
3. Rouvelas I, Zeng W, Lindblad M, et al. Survival after surgery for oesophageal cancer: a population-based study. Lancet Oncol 2005;6: 864 – 870.
4. Stein HJ, von Rahden BHA, Siewert JR. Survival after oesophagectomy for oesophageal cancer. Langenbecks Arch Surg 2005;4:280 –285. 5. Stein HJ, Feith M, Bruecher BL, et al. Early esophageal cancer: pattern
of lymphatic spread and prognostic factors for long-term survival after surgical resection. Ann Surg 2005;242:566 –573.
6. Tachibana M, Kinugasa S, Hirahara N, et al. Lymph node classification of esophageal squamous cell carcinoma and adenocarcinoma. Eur J
Car-diothorac Surg 2008;34:427– 431.
7. Sobin LH, Wittekind C (Eds.). TNM Classification of Malignant Tu-mours (UICC S.), 6th Ed. Hoboken, NJ: John Wiley & Sons Inc., 2002.
8. Greene FL, Page DL, Fleming ID (Eds.). AJCC Cancer Staging Manual, 6th Ed. New York, NY: Springer, 2002.
9. Roder JD, Busch R, Stein HJ, et al. Ratio of invaded to removed lymph nodes as a predictor of survival in squamous cell carcinoma of the oesophagus. Br J Surg 1994;81:410 – 413.
10. Ellis FH Jr, Heatley GJ, Krasna MJ, et al. Esophagogastrectomy for carcinoma of the esophagus and cardia: a comparison of findings and results after standard resection in three consecutive eight-year intervals with improved staging criteria. J Thorac Cardiovasc Surg 1997;113: 836 – 846.
11. Korst RJ, Rusch VW, Venkatraman E, et al. Proposed revision for the staging classification of esophageal cancer. J Thorac Cardiovasc Surg 1998;115:660 – 670.
12. Nigro JJ, DeMeester SR, Hagen JA, et al. Node status in transmural esophageal adenocarcinoma and outcome after en bloc esophagectomy.
J Thorac Cardiovasc Surg 1999;117:960 –968.
13. Hagen JA, DeMeester SR, Peters JH, et al. Curative resection for esophageal adenocarcinoma: analysis of 100 en bloc esophagectomies.
Ann Surg 2001;234:520 –530.
14. Rice TW, Blackstone EH, Rybicki LA, et al. Refining esophageal cancer staging. J Thorac Cardiovasc Surg 2003;125:1103–1113.
15. Bollschweiler E, Baldus SE, Schro¨der W, et al. Staging of esophageal carcinoma: length of tumor and number of involved regional lymph nodes. Are these independent prognostic factors? J Surg Oncol 2006; 94:355–363.
16. Rizk N, Venkatraman E, Park B, et al; American Joint Committee on Cancer staging system. The prognostic importance of the number of involved lymph nodes in esophageal cancer: implications for revisions of the American Joint Committee on Cancer staging system. J Thorac
Cardiovasc Surg 2006;132:1374 –1381.
17. Barbour AP, Rizk NP, Gonen M, et al. Lymphadenectomy for adeno-carcinoma of the gastroesophageal junction (GEJ): impact of adequate staging on outcome. Ann Surg Oncol 2007;14:306 –316.
18. Bogoevski D, Noken F, Koenig A, et al. Is it time for a new TNM classification in esophageal carcinoma? Ann Surg 2008;247:633– 641. 19. Mariette C, Piessen G, Briez N, et al. The number of metastatic lymph
nodes and the ratio between metastatic and examined lymph nodes are independent prognostic factors in esophageal cancer regardless of neo-adjuvant chemoradiation or lymphadenectomy extent. Ann Surg 2008; 247:365–367.
20. Greenstein AJ, Litle VR, Swanson SJ, et al. Prognostic significance of the number of lymph node metastases in esophageal cancer. J Am Coll
Surg 2008;206:239 –246.
21. Edge SB, Byrd DR, Compton CC, et al. (Eds.). AJCC Cancer Staging Manual, 7th Ed. New York, NY: Springer, 2010.
22. Martin DJ, Church NG, Kennedy CW, et al. Does systematic 2-field lymphadenectomy for esophageal malignancy offer a survival advan-tage? Results from 178 consecutive patients. Dis Esophagus 2008;21: 612– 618.
23. Kaplan EL, Meier P. Non parametric estimation from incomplete ob-servations. J Am Stat Assoc 1958;53:457– 481.
24. Cox D. Regression models and life tables (with discussion). J R Stat Soc 1972;187–220.
25. Pramesh CS, Kulkarni NN, Mistry RC, et al. Lymphadenectomy in cancer: time for a paradigm shift? Lancet Oncol 2006;7:450 – 451. 26. Gebski V, Burmeister B, Smithers BM, et al. Survival benefits from
neoadjuvant chemoradiotherapy or chemotherapy in esophageal carci-noma: a meta-analysis. Lancet Oncol 2007;8:226 –234.
