Embed Size (px)
Citation preview
606
Para-aortic lymph node metastasis incarcinoma of the head of the pancreasMitsuru Sakai, MD, Akimasa Nakao, MD, Tetsuya Kaneko, MD, Shin Takeda, MD, Soichiro Inoue, MD,Yasuhiro Kodera, MD, Shuji Nomoto, MD, Naohito Kanazumi, MD, and Hiroyuki Sugimoto, MD,Nagoya, Japan
Background. Metastasis to para-aortic lymph nodes often occurs in pancreatic head cancer, but factorsthat predict it are not well known.Methods. Using histopathologic data of 178 patients who underwent extended lymph node dissection forpancreatic head cancer, we analyzed the distribution of metastases to lymph node groups classified indetail and attempted to identify the lymph node groups that have a strong relation with metastasis topara-aortic lymph nodes.Results. A high incidence of lymph node metastasis was found in para-aortic lymph nodes (No. 16,19%) as well as in regional lymph nodes, such as those on the posterior aspect of the pancreas head (No.13, 47%), on the anterior surface of the pancreas head (No. 17, 29%), along the superior mesentericartery (No. 14, 28%), and along the hepatoduodenal ligament (No. 12, 19%). Statistical analysisshowed that metastases to para-aortic lymph nodes had a strong correlation with metastases to Nos. 12,13, 14, and 17 lymph nodes. Para-aortic lymph node metastases were seldom observed among thepatients who had no metastases to Nos.13, 14, and 17 lymph nodes.Conclusions. Examination of Nos. 13, 14, and 17 lymph nodes may be useful to predict the involvementof para-aortic lymph nodes. (Surgery 2005;137:606-11.)
From the Department of Surgery II, Nagoya University School of Medicine
PANCREATIC CANCER remains a disease with a dismaloutcome despite continuous efforts to improve itssurvival rate. Systemic resection of lymph nodeshas been performed in various cancer types in anattempt to improve survival in patients who havenode metastases; long-term survivors among thosewith node metastases may serve as proof of thebenefits achieved from such a procedure. Extend-ing the range of node dissection may, in theory,increase the chance for cure, but the survivalbenefit of extended lymphadenectomy for gastriccarcinoma, for instance, has not been provenclearly in prospective randomized trials.1 The issueof adequate extent of lymphadenectomy for pan-creatic cancer also is currently a matter of contro-versy.2 Para-aortic lymph nodes usually are outsidethe range of resection in the standard surgicaltreatment for pancreatic cancer. Metastases to the
Accepted for publication February 23, 2005.
Reprint requests: Akimasa Nakao, MD, Department of SurgeryII, Nagoya University School of Medicine, Tsurumai-cho 65,Showa-ku, Nagoya 466-8550, Japan. E-mail: [email protected].
0039-6060/$ - see front matter
� 2005 Mosby, Inc. All rights reserved.
doi:10.1016/j.surg.2005.02.009
SURGERY
para-aortic lymph nodes, however, are observedcommonly among patients with carcinoma of thehead of the pancreas.3-6 The true indication forsystematic resection of these lymph nodes is unde-termined currently.
Since 1981 we have performed isolated pancre-atectomy with extended lymphadenectomy thatincludes systematic dissection of the para-aorticlymph nodes, in addition to the standard range oflymph nodes, and coresection of the portal veinusing a catheter bypass procedure.3,7-9 As a result,a relatively large body of cumulative data from asingle institution on the lymph node involvementof cancer of the head of the pancreas has enabledus to conduct a retrospective analysis. On the basisof these data, we analyzed the distribution ofmetastases to lymph node groups classified indetail and attempted to identify the ‘‘junctionallymph nodes’’ to para-aortic lymph nodes. Theconcept of ‘‘junctional lymph nodes’’ has beenadvocated recently in an attempt to choose be-tween limited and more extensive lymphadenec-tomy.10 In this view, cancer cells are considered tospread to distant lymph nodes via the junctionallymph nodes, the examination of which allowsthe surgeon to decide whether to stop the dissec-tion of lymph nodes there or to proceed to moreextensive lymphadenectomy. According to our
SurgeryVolume 137, Number 6
Sakai et al 607
study, patientswithnometastasis to the lymphnodeson the posterior aspect of the pancreas, on theanterior surface of the pancreas head, and along thesuperior mesenteric artery seldom had para-aorticlymph node metastasis. These lymph nodes couldbe considered as ‘‘junctional lymph nodes’’ to thepara-aortic lymph nodes. This approach may estab-lish more adequate criteria for dissection of para-aortic lymph nodes.
PATIENTS AND METHODS
Between July 1981 and June 2002, 178 patientswith invasive ductal carcinoma of the head of thepancreas underwent extended radical surgery withsystematic lymph node resection, including para-aortic lymph nodes, at the Department of SurgeryII, Nagoya University. Patients with intraductalpapillary mucinous neoplasms were excluded.These 178 patients consisted of 120 men and 58women with a mean age of 62 years (range, 38-83years). Ninety-eight pancreatoduodenectomies, 26pylorus-preserving pancreatoduodenectomies, and54 total pancreatectomies were performed. Portalvein resection was performed in 144 patients. Themean number of dissected lymph nodes was 45(range, 2-139), and the mean number of dissectedpara-aortic lymph nodes was 7 (range, 0-33). Allresected specimens were examined histopatholog-ically after being fixed and stained with hematoxy-lin and eosin. Pathologic findings were evaluatedin accordance with the second English edition ofthe Classification of Pancreatic Carcinoma proposedby the Japan Pancreas Society.11 This classificationscheme is more detailed than the TNM classifica-tion by Union Internationale Contre le Cancer.12
Lymph nodes are classified into several lymphnode stations named according to the anatomiclocation and numbered (Fig 1). These stations arefurther classified into several groups according tothe distance from the pancreas. Lymph node Nos.13 and 17 belong to group 1, Nos. 6, 8, 12, and 14to group 2, and other lymph nodes to group 3.Para-aortic lymph nodes (No. 16) in this study referto those that are surrounded by the celiac trunk,the inferior mesenteric artery, the right margin ofthe inferior vena cava, and the left margin of theabdominal aorta. Lymph nodes belonging to group1, group 2, and para-aortic lymph nodes were alldissected completely at operation. Absence of alymph node in resected specimens to be examinedin a given station was treated as no metastasis tothe station. The number of lymph nodes foundin each station varied significantly among thepatients.
Statistical analyses were performed with the useof StatView statistical software (SAS Institute Inc,Cary, NC).13 The correlations between the inci-dence of metastases to various lymph node stationsand those between para-aortic lymph node involve-ment and the other pathologic parameters werestudied with the Fisher exact test. Survival rates,including postoperative death, were calculated bythe Kaplan-Meier method. Differences in survivalrates among the subjects were analyzed by the log-rank test, and a P value < .05 was consideredsignificant.
RESULTS
Incidence of metastasis to each lymph nodestation and survival of patients with metastasis to agiven station. Lymph node metastases occurred in118 (66%) of 178 patients with carcinoma of thehead of the pancreas. The incidence of metastasisto each of the lymph node stations defined by thesecond English edition of the Classification of Pan-creatic Carcinoma11 is given in Table I. The inci-dence was particularly high for lymph nodes at theposterior aspect of the head of the pancreas (No.13, 47%), on the anterior surface of the head ofthe pancreas (No. 17, 29%), and along the supe-rior mesenteric artery (No. 14, 28%). For para-aortic lymph nodes, metastases were detected in asmany as 34 of 178 patients (19%); the incidencewas the same as that of the lymph nodes alongthe hepatoduodenal ligament (No. 12, 19%) thatbelong to the group 2 lymph nodes.
One-year, 2-year, and 3-year survival rates ofpatients with metastatic involvement of each ofthe lymph node stations also are given in Table I.Survival of patients with lymph node metastasis,and those with metastases to the para-aortic nodesin particular, was significantly worse compared withthose without node metastasis (Fig 2). One-year,2-year, and 3-year overall survival rates for thepatients with nodal metastases were 42%, 19%,and 12%, respectively. In comparison, 1-, 2-, and 3-year survival rates of patients with metastasis to thepara-aortic nodes were inferior at 30%, 7%, and3%, respectively. But, metastases to the para-aorticlymph nodes added little effect to the already poorsurvival rate observed among a subset with lymphnode involvement. A histopathologic confirmationof the cancer-free margin at the dissected peri-pancreatic tissue has been reported to be a signif-icant prognostic factor.7 Also in this study, 1-, 2-,and 3-year survival rates of patients with a negativeresection margin were relatively high at 61%, 33%,and 20%, respectively. In contrast, the prognosis of
SurgeryJune 2005
608 Sakai et al
Fig 1. According to the second English edition of the Classification of Pancreatic Carcinoma proposed by theJapan Pancreas Society, the nomenclature of the major lymph node stations is defined as follows: No. 5,suprapyloric lymph nodes; No. 6, infrapyloric lymph nodes; No. 7, lymph nodes along the left gastric ar-tery; No. 8a, lymph nodes in the anterosuperior group along the common hepatic artery; No. 8p, lymphnodes in the posterior group along the common hepatic artery; No. 9, lymph nodes around the celiac ar-tery; No. 10, lymph nodes at the splenic hilum; No. 11p, lymph nodes along the proximal splenic artery;No. 11d, lymph nodes along the distal splenic artery; No. 12a, lymph nodes along the hepatic artery; No.12p, lymph nodes along the portal vein; No. 12b, lymph nodes along the bile duct; No. 13a, lymph nodeson the posterior aspect of the superior portion of the head of the pancreas; No. 13b, lymph nodes on theposterior aspect of the inferior portion of the head of the pancreas; No. 14p, lymph nodes along the prox-imal superior mesenteric artery; No. 14d, lymph nodes along the distal superior mesenteric artery; No. 15,lymph nodes along the middle colic artery; No. 16, lymph nodes around the abdominal aorta; No. 17a,lymph nodes on the anterior surface of the superior portion of the head of the pancreas; No. 17b, lymphnodes on the anterior surface of the inferior portion of the head of the pancreas; and No. 18, lymph nodesalong the inferior margin of the pancreas.
patients with a positive margin was extremely poor,with 1-, 2-, and 3-year survival rates of 13%, 2%,and 2%, respectively. Rather surprisingly, 1-, 2-, and3-year survival rates of 16 patients with metastasesto the para-aortic lymph nodes among a subset ofpatients with negative dissected margin were rela-tively high at 39%, 16%, and 8%, respectively.Nevertheless, the prognosis of 18 patients withpara-aortic lymph nodes metastasis and positivedissected margin was poor, with 1-year and 2-yearsurvival rates of 22% and 0%, respectively.
Correlation between metastases to lymph nodestations and those to para-aortic nodes. We deter-mined the correlation between metastases to eachlymph node station as defined by the Classificationof Pancreatic Carcinoma.11 Strong correlations
(P values < .0004) were observed between metastasisto lymph node Nos. 13 and 17, Nos. 12 and 16,Nos. 16 and 17, Nos. 13 and 16, and Nos. 12 and13. Metastasis to the para-aortic lymph node hadstrong correlations with metastases to lymph nodesNos. 12, 13, 14, and 17 (P values < .001).
Distribution of metastases to other stationsamong patients with para-aortic lymph node in-volvement. We determined distribution patterns ofmetastatic lymph nodes among the subset of 34patients with metastases to the para-aortic lymphnodes. Of these, 7 had involvement with only1 other lymph node station. Of the 7 patients,3 had metastasis to station No. 13, 3 patients toNo. 14 lymph node, and the remaining 1 to No. 17lymph node. Only 4 of the 34 patients had
SurgeryVolume 137, Number 6
Sakai et al 609
neither No. 13 nor No. 14 node involvement,indicating the significance of these nodes as pre-dictive factors of metastases to the para-aorticlymph nodes. Metastases to No. 12 lymph nodestations also had a strong correlation with involve-ment of the para-aortic lymph node, but allpatients with involvement of stations Nos. 12 and16 also had metastasis to either No. 13 or No. 17.Only 2 patients had metastasis to the para-aorticregion with no other lymph nodes involved.
Finally, Table II shows the incidences of para-aortic lymph node involvement among thepatients with or without metastasis to lymph nodestation Nos. 12, 13, 14, or 17. Patients with metas-tases to any of station Nos. 12, 13, 14, and 17 had ahigh risk of metastases to the para-aortic lymphnodes. Conversely, the incidence was very low at5% for a subset of patients who had no metastasisto station Nos. 13 and 14, and 3% for those whohad no metastasis to station Nos. 13, 14, and 17.
Correlation between para-aortic lymph nodeinvolvement and other macroscopic pathologicparameters. We next analyzed the correlation be-tween para-aortic lymph node involvement andother macroscopic findings as defined by theClassification of Pancreatic Carcinoma of theJapan Pancreas Society.11 These findings includedtumor diameter, serosal invasion, retropancreatictissue invasion, distal bile duct invasion, duodenalinvasion, and portal venous system invasion. Mac-roscopic invasion of the distal bile duct and tumordiameter had a significant correlation with para-aortic lymph node involvement (P value < .05).
Table I. Frequency of each lymph node metastasisand survival rates
Lymph nodestation
Frequency ofmetastasis(n = 178)
Survival rates of patientswith lymph nodeinvolvement (%)
1-year 2-year 3-year
6 21 (12%) 20 0 07 0 — — —8 17 (10%) 29 6 09 2 (1%) 50 0 0
10 0 — — —11 14 (8%) 29 7 712 33 (19%) 39 0 013 83 (47%) 34 14 714 50 (28%) 34 5 215 2 (1%) 50 50 016 34 (19%) 30 7 317 51 (29%) 29 12 618 3 (2%) 33 0 0No involvement 60 (34%) 57 32 23
DISCUSSION
Since Fortner14 advocated an extended radicalsurgical procedure for pancreatic head cancer in1973, we have performed extended radical surgerywith wide dissection of lymph nodes, including thepara-aortic lymph nodes, to improve the prognosisof patients with this disease.3,7-9 In the currentstudy, 118 of 178 patients had lymph node involve-ment, and 34 patients had para-aortic node metas-tasis. The mean number of lymph nodes retrievedwas 45, including 7 para-aortic nodes for eachpatient. Despite these efforts, the prognosis of ourpatients did not surpass that of those treated withmore limited lymphadenectomy,2 pointing to thefact that extended lymphadenectomy has little
Fig 2. Survival rates of the patients with or without No.16 lymph node metastasis. Prognosis of the patientswith any lymph node metastasis was significantly inferior.No significant difference in the survival rates of the pa-tients with or without para-aortic lymph node involve-ment was found among patients who had lymph nodeinvolvement.
Table II. Rates of para-aortic lymph nodeinvolvement
nRates of metastasis to
para-aortic lymph node (%)
LN12metastasis(+) 33 46metastasis(�) 145 13
LN13metastasis(+) 83 31metastasis(�) 95 8
LN14metastasis(+) 50 36metastasis(�) 128 13
LN17metastasis(+) 51 41metastasis(�) 127 10
LN13 and LN14metastasis(�) 79 5
LN13, LN14, and LN17metastasis(�) 70 3
SurgeryJune 2005
610 Sakai et al
survival benefit for patients with pancreatic headcancer. These findings and experience with othercancer types15,16 suggest that the indication forextended lymphadenectomy should be reconsid-ered seriously.
In this study, as indicated by several otherreporters from Japanese institutions,4-6 the inci-dences of metastasis exceeded 10% in lymph nodestation Nos. 12, 13, 14, 16, and 17. These lymphnodes belong either to group 1 or 2 according tothe Classification of Pancreatic Carcinoma of theJapan Pancreas Society,11 with the exception ofpara-aortic nodes (No. 16). Resection of the lymphnode stations belonging to groups 1 and 2 hasbeen recommended as a standard procedure inJapan, whereas the indication of extensive resec-tion of No. 16 has been a matter of controversy. Wetherefore examined in detail various patterns oflymphatic spread in a group of patients with cancerof the pancreatic head to readjust the indicationfor extended lymphadenectomy that includes thepara-aortic nodes.
A number of reports5,17,18 referred to thelymphatic flow to the para-aortic lymph node.Nagakawa et al,17 for example, injected activatedcarbon particles or 111In colloid in their patientswith pancreatic cancer to investigate the lymphaticspread from the head of the pancreas to the para-aortic lymph node and concluded that the mainlymphatic route to the para-aortic lymph node wasthrough the nodes in the posterior part of thehead of the pancreas (No. 13) and around thesuperior mesenteric artery (No. 14). Nagai18 foundthat dye injected into the posterior region of thepancreas head drained toward the right or poste-rior side of the superior mesenteric artery andfinally to the para-aortic lymph nodes. In addition,Kayahara et al5 indicated that the main lymphaticpathway from the head of the pancreas to the para-aortic lymph nodes was via the No. 14 lymphnodes, which harbored metastasis in all 7 patientswith para-aortic lymph node involvement in theirseries. Together with these reports, our study alsosupports the hypothesis that the lymphatic path-ways from the primary to the para-aortic lymphnodes were via station No. 13 or 14. So-called skipmetastasis, which in this case denotes metastases tothe para-aortic nodes without detectable metasta-ses to either station No. 13, 14, or 17, was observedin only 2 of 178 patients, amounting to less than3% of the current series. We therefore advocatethat lymph nodes belonging to station Nos. 13, 14,and 17 are ‘‘junctional lymph nodes.’’ This indi-cates that if no metastases to these lymph nodesare detected by surgical exploration or intraoper-
ative pathologic examination, metastasis to para-aortic lymph nodes can be considered unlikely.
It has not been possible to define the metastaticpathway for the 2 patients with skip metastases tothe para-aortic nodes within the scope of this study,but we cannot deny the possibility that microscopicmetastases that are undetectable with routinehistopathologic examination may have existed inthe nodes that belong to station Nos. 13, 14, or17 from which the metastases to para-aortic nodesmay have occurred. This brings us to anotherunresolved problem of micrometastasis. It is wellknown that highly sensitive detection methodssuch as immunostaining or polymerase chainreaction-based assays have detected micrometastasesin the lymph nodes that had been diagnosed ascancer negative by routine pathologic examination.Such phenomena have been reported in pancreaticcancer19-21 as well as in several other cancer types.Niedergethmannet al21 found throughmultivariateanalysis that the presence of micrometastases to thepara-aortic lymph nodes diagnosed through detec-tion of K-ras mutation was an independent prog-nostic factor. This finding suggests that para-aorticlymphadenectomy might be indicated even in theabsence of definite metastasis to the para-aorticnodes. However, other reports22 argue against theprognostic impact of micrometastases, and the truebenefit of dissecting micrometastases through ex-tended lymphadenectomy remains an issue forfuture investigation.
According to our analysis of the correlationbetween para-aortic lymph node involvement andseveral macroscopic parameters, no patient with-out distal bile duct invasion or with a tumordiameter of less than 2 cm had para-aortic lymphnode involvement. Since these parameters can beevaluated before or during operation, it may alsobe useful to incorporate these factors into thedecision making with regard to dissection of thepara-aortic lymph nodes. The correlation betweentumor diameter and incidence of metastases to thepara-aortic lymph nodes remains controversial,however, given the contradictory report fromNagakawa et al.17
CONCLUSION
The lymph nodes that predict metastases to thepara-aortic lymph node in pancreatic head cancerwere considered to be those on the posterioraspect of the head of the pancreas, on the anteriorsurface of the head of the pancreas, and alongthe superior mesenteric artery. In patients withno metastasis to these lymph nodes, the risk of
SurgeryVolume 137, Number 6
Sakai et al 611
para-aortic lymph node metastasis is consideredhighly unlikely.
REFERENCES
1. Kodera Y, Schwarz RE, Nakao A. Extended lymph nodedissection in gastric carcinoma: Where do we stand after theDutch and British randomized trials? J Am Coll Surg 2002;195:855-64.
2. Yeo CJ, Cameron JL, Lillemoe KD, Sohn TA, Campbell KA,Sauter PK, et al. Pancreaticoduodenectomy with or withoutdistal gastrectomy and extended retroperitoneal lymphad-enectomy for periampullary adenocarcinoma, part 2: ran-domized controlled trial evaluating survival, morbidity, andmortality. Ann Surg 2002;236:355-66.
3. Nakao A, Harada A, Nonami T, Kaneko T, Murakami H,Inoue S, et al. Lymph node metastasis in carcinoma of thehead of the pancreas region. Br J Surg 1995;82:399-402.
4. Kayahara M, Nagakawa T, Ohta T, Kitagawa H, Ueno K,Tajima H, et al. Analysis of paraaortic lymph node involve-ment in pancreatic carcinoma. Cancer 1999;85:583-90.
5. Kayahara M, Nagakawa T, Kobayashi H, Mori K, Nakano T,Kadoya N, et al. Lymphatic flow in carcinoma of the head ofthe pancreas. Cancer 1992;70:2061-6.
6. Ishikawa O, Ohigashi H, Sasaki Y, Kabuto T, Furukawa H,Nakamori S, et al. Practical grouping of positive lymphnodes in pancreatic head cancer treated by an extendedpancreatectomy. Surgery 1997;121:244-9.
7. Nakao A, Kaneko T, Takeda S, Inoue S, Harada A, NomotoS, et al. The role of extended radical operation for pancre-atic cancer. Hepatogastroenterology 2001;48:949-52.
8. Nakao A, Nonami T, Harada A, et al. Portal vein resectionwith a new antithrombogenic catheter. Surgery 1990;108:913.
9. Nakao A, Harada A, Nonami T, Kaneko T, Inoue S, TakagiH. Clinical significance of portal invasion by pancreatichead carcinoma. Surgery 1995;117:60-5.
10. Whiting JL, Hallissey MT, Rowlands DC, Fielding JW.Redefining surgery for gastric cancer. Gastric Cancer1999;2:226-9.
11. Japan Pancreatic Society. Classification of PancreaticCarcinoma. 2nd English edition. Tokyo: Kanehara & CoLtd; 2003.
12. Sobin LH, Wittekind CH. TNM classification of malignanttumours. 6th ed. New York: Wiley-Liss; 2002.
13. StatView [computer program]. Version 5. SAS Institute Inc;Cary, NC, 1998.
14. Fortner JG. Regional resection of cancer of the pancreas.Surgery 1973;73:307-20.
15. Bonenkamp JJ, Hermans J, Sasako M, van de Velde CJ.Extended lymph-node dissection for gastric cancer. DutchGastric Cancer Group. N Engl J Med 1999;340:908-14.
16. Cuschieri A, Weeden S, Fielding J, Bancewicz J, Craven J,Joypaul V, et al. Patient survival after D1 and D2 resectionsfor gastric cancer: long-term results of the MRC random-ized surgical trial. Br J Cancer 1999;79:1522-30.
17. Nagakawa T, Kobayashi H, Ueno K, Ohta T, Kayahara M,Miyazaki I. Clinical study of lymphatic flow to the paraaorticlymph nodes in carcinoma of the head of the pancreas.Cancer 1994;73:1155-62.
18. Nagai H. An anatomic and pathologic study of autopsymaterial on metastasis of pancreatic cancer to para-aorticlymph nodes. Jpn J Surg 1987;88:308-17.
19. Demeure MJ, Doffek KM, Komorowski RA, Wilson SD.Adenocarcinoma of the pancreas: detection of occultmetastasis in regional lymph nodes by a polymerase chainreaction-based assay. Cancer 1998;83:1328-34.
20. Ando N, Nakao A, Nomoto S, Takeda S, Kaneko T,Kurokawa T, et al. Detection of mutant K-ras in dissectedparaaortic lymph nodes of patients with pancreatic adeno-carcinoma. Pancreas 1997;15:374-8.
21. Niedergethmann M, Rexin M, Hildenbrand R, Knob S,Sturm JW, Richter A, et al. Prognostic implications ofroutine, immunohistochemical, and molecular staging inrespectable pancreatic adenocarcinoma. Am J Surg Pathol2002;26:1578-87.
22. Fukagawa T, Sasako M, Mann GB, Sano T, Katai H,Maruyama K, Nakanishi Y, Shimoda T. Immunohistochemi-cally detected micrometastases of the lymph nodes inpatients with gastric carcinoma. Cancer 2001;15:753-60.
