Curative surgery is the treatment of choice for gastric cancer, but the extent of this surgical procedure is still highly controversial.[1] Extended lymph node dissection (D2 or more) in the surgical treatment of gastric cancer has been introduced and routinely performed over the last two decades in Japan.[2,3] This technique achieves to remove a large number of lymph nodes around the celiac axis and from the ligament hepatoduodenal (level II: N2 disease according to the Japanese Classification[4]) which is left behind after a limited (D1) node dissection. The probability that these nodes contain metastatic disease is rather substantial and consequently D2 dissection, as compared with a D1 dissection, is associated with a lower risk of residual disease in these level II lymph nodes. This potential of D2 dissection has been well recognized and established

However, a possible beneficial effect of D2 dissection, if it is really exist, on long-term survival has not yet been documented according to the criteria of evidence-based western medicine.[1] Two are the major claims against clinical use of D2 dissection: (1) It increases morbidity and mortality. (2) It does not improve survival.

Indeed, in the two European randomized trials[5,6] available, D2 dissection increased significantly the risk of morbidity and mortality. However, as it has been repeatedly emphasized, surgeon's experience with the technique of D2 dissection is the predominant factor for the safety of this procedure.[7,8] In both trials the participated surgeons performed one or two only D2 dissection per year and were not familiar with this technique before they start to perform D2 dissection. This lack of experience and distal pancreatico-splenectomy routinely performed in these trials are the causes for the increased morbidity and mortality assessed.[8] Currently, credible evidence for the safety of D2 dissection, when it is performed with a systematic and standardized pancreas-preserving technique by experienced surgeons, provides a latest Japanese randomized trial[9] confirming previous results from specialized institutions.[2,3,7,10] Taken all these data together the message is clear. D2 dissection should not be performed by surgeons inexperienced with this technique. In this case a D1 dissection is preferable. However, after a learning curve and with a standardized pancreas-preserving technique, D2 dissection can be performed with safety.

While the effect of D2 dissection on post-operative morbidity and mortality has been elucidated, assessment of this effect on long-term survival is complex, sophisticated and still unclear. This issue remains the challenge of current clinical research in the surgical treatment of gastric cancer. Many retrospective Japanese studies have showed a survival benefit after D2 dissection, but there is still lack of scientific evidence that D2 dissection improves survival. Several key point questions should be addressed for the understanding of this complicated problem:
(1) What is exactly the quantity of additional (level II) lymph nodes removed by a D2 dissection and left behind by a D1 dissection?
(2) What is the magnitude of the risk of metastatic disease existence in these level II nodes and how can it predicted?
(3) Does the elimination or reduction of this risk due to D2 dissection reflected also by a survival benefit?


It is clear that a substantial amount of fatty connective tissue is removed from around the celiac axis (superior border of the pancreas) and hepatoduodenal ligament by a D2 dissection and this tissue with all containing nodes is left behind by a D1 dissection.(Figures 1 to 6) According to the Japanese classification, which is based on the anatomical site of lymph nodes, the level II includes the extraperigastric lymph nodes. These are the nodes around the celiac axis, i.e., along the celiac artery (station 9), left along the splenic artery (station 11) toward splenic hilum (station 10) and right the nodes along the hepatic artery (stations 8) toward liver hilum including the nodes in the ligament hepatoduodenal (station 12). Little attention has received the calculation of this nodal yield. In a recent prospective study[11] this nodal quantity was calculated on the basis of references value from an experimental study in cadavers[12] and the findings of the Dutch trial.[13] The authors estimated that in mean 11,4 to 13 lymph nodes[11] can be retrieved from the fatty tissue which left behind after a D1 dissection. The results indicated also that there is a biologic intrinsic variation in the number of lymph nodes that varies considerably in each individual patient. Although this nodal quantity is considerable, clinically important is to be assessed how often these nodes contain metastasis in order to satisfy the need for dissection.

Currently the magnitude of this risk could be calculated. On the basis of D1 dissection-attributable risk of residual disease, well-documented histological data of Japanese[2,3] and western prospective studies[5-7,10,13] were reviewed. Among patients who had undergone a curative D2 dissection, the analysis focused on the study of those who had nodal metastasis beyond the level I, i.e., in level II (N2 disease according to the Japanese classification). The results consistently indicated that at least 30% of patients had metastatic disease in the level II nodes at the time of surgery.[14] Consequently; D1 dissection results in residual disease approximately one in three patients suitable for curative surgery.

The third and clinically most important question is whether eliminating this residual disease-risk in the level II nodes due to their removal by a D2 dissection we could produce an improved recurrence-free survival and overall survival. This issue has been still highly debated. Previous nonrandomized studies have showed a stage-specific survival benefit after a D2 dissection. Only patients with nodal metastasis confined to the level I (N1), mainly with stages II and IIIA, according to the International Union against Cancer and American Joint Committee on Cancer (UICC/AJCC)15 classification, had significantly better survival after D2 dissection than those who had a D1 dissection.[7] In the two European randomized trials no significant difference in the 5 year survival rates between D1 dissection (45%) and D2 dissection (47%) groups was found. However, the authors of the Dutch trial in an update of the follow-up data, revised the initial results supporting now that patients with stage II or IIIA significantly benefited from a D2 dissection.[16]

Thus, a current question emerges: Do patients with stage II or IIIA really benefited from a D2 dissection? The findings of these studies are not consistent and can not provide credible evidence for the effectiveness of D2 dissection for the following reasons:

(1) D2 dissection is essential for accurate nodal staging. In D1 dissection groups the nodal stage of disease is substantially underestimated because it is well known that not enlarged nodes may contain metastasis detectable only by histopathological examination. Thus, even in the randomized trials patients were classified to have N1 disease although they had metastasis in the level II nodes undetectable with D1 dissection. This stage migration confounds the comparison of survival results because patients with a more advanced stage of disease (D1 group) are compared with those with an earlier stage (D2 group). Stage migration would be eliminated if the survival benefit demonstrated in stages II/IIIA subgroups would also be resulted in an overall survival advantage in the total population studied (all stages). However, in any of these studies the overall survival rate in D2 group was significantly higher than that in D1 group.

(2) The description D2 dissection has been emerged from the nodal staging according to the Japanese classification (anatomical site of lymph nodes). Therefore, a comparison of D1 and D2 groups according to the UICC/AJCC stages of the disease is incorrect and leads to false conclusions. Stage II or IIIA includes patients with various nodal staging (N0, N1, N2) that confounds comparison. Japanese studies recognizing the significance of this comparison have showed better survival with D2 dissection according to the nodal stage but unfortunately they used historic control groups. Such a comparison is of little evidence to make treatment decision.

(3) It is hard to explain why patients without nodal metastasis (N0 disease) or metastasis confined to the first only level (N1 disease) benefit from a D2 dissection. The level I nodes are dissected as completely as with a D1 dissection and logically no survival benefit would be expected by a D2 dissection. The hypothesis that patients with N0 or N1 disease benefited from D2 dissection because they have micrometastasis in the level II nodes has not been confirmed.[1]

Alternatively, the assessment of the effect of D2 dissection on long-term survival could be evaluated on the basis of a recently described concept.[10] The benefit of D2 vs. D1 dissection, from surgical-oncological point of view, is clearly restricted to the additional dissection of fatty-connective tissue from the superior border of pancreas and hepatoduodenal ligament. Therefore, the estimation of survival in the subgroup of patients with metastatic disease in these level II nodes (N2 disease), would reveal whether D2 dissection is of therapeutic value. A review of survival data among patients with N2 disease who had undergone a D2 dissection reveals a 5-year survival rate ranging between 20% and 39%.[10,17] Based on the hypothesis that none of the patients with N2 disease would survived 5 years after D1 dissection because of residual disease and fatal relapse, we can conclude that the survival benefit of D2 dissection is substantial. D2 dissection seems to offer a long-term survival chance approximately in one of four patients with N2 disease undergoing curative surgery. If it is true, why in any of the prospective studies available an overall survival benefit could not be demonstrated? The most likely explanation is that since the proportion of patients with N2 disease accounts for 30% of all patients and only about 25% of these survive more than 5 years after curative D2 dissection, it is calculated that only 7.5% (25X30%) of all patients benefit from a D2 dissection. The results indicate that in the population received potentially curative surgery, one additional life would be saved for every 13 patients treated with D2 dissection. Therefore, a much larger sample size is required to reach statistical significance than that reported in all studies available.

The lack of credible evidence for the survival benefit of D2 dissection can be attributable either to the inappropriate design and conduction of randomized trials available or very simply D2 dissection does not results in improved survival. Some investigators argue the effectiveness of D2 dissection supporting that patients with N2 disease have at the time of surgery additional clinically undetectable metastatic disease in distant organs or in the lymph nodes beyond the level II. Thus, even D2 dissection is associated with residual disease and fatal relapse. Indeed, this is true for most patients with N2 disease, but in a minority of these patients, probably those with a few only positive level II nodes, D2 dissection offers a chance of long-term survival. Furthermore, D2 dissection may prolong survival due to removal of micrometastatic disease in the level II nodes. Irrespective of mechanisms by which D2 dissection may improve survival further prospective evaluation is needed.

Because D2 dissection requires surgeon's experience one could claim that an adjuvant treatment after limited (D1) surgery could results in similar or better survival than D2 dissection. However, up until now there was no credible evidence for the effectiveness of adjuvant treatment after gastric resection.[1,18] Promising are the findings of a latest randomized USA trial suggested a survival benefit with adjuvant chemoradiotherapy after surgery.[19] However, this multimodality treatment was associated with frequent major toxicity and although it is hard to be compared, the long-term survival results seem worse than that reported with D2 dissection alone from specialized centers.[2,3,7,10,17] or by the Dutch trial.[5]

Conclusions
Extended, as compared with limited, lymph node dissection achieves to remove a substantially larger number of additional lymph nodes. Approximately 30% of patients receiving "curative" D1 dissection are at high-risk of residual metastatic disease in the level II nodes. D2 dissection eliminates this risk resulting in a better local/nodal control. Overall local and nodal recurrence rate of 12% currently reported after D2 dissection 20 is superior to rates of previous reports with over 40% after D1 dissection. Whether this better local/nodal control by D2 dissection is translated into a better disease-free survival and overall survival still remains non-evident. Despite this lack of evidence, D2 dissection should be considered the treatment of choice because curative surgery in a substantial proportion of patients can be achieved only due D2 dissection and because no adjuvant treatment at present is effective to eliminate D1-attributable residual disease.

Because experience with the technique of D2 dissection is the predominant factor for the safety of D2 dissection, it should be performed either in specialized institutions where D2 dissection is routinely and safely performed or in the other settings only after an essential learning curve. D1 dissection remains the correct treatment decision when the surgeon is not familiar with the technique of D2 dissection.

References
1. Roukos DH, Fatouros M, Xeropotamos N, Kappas AM. Treatment of gastric cancer: early-stage, advanced-stage cancer, adjuvant treatment. Gastric Breast Cancer 2002; 1(1): 12-22.
2. Marujama K, Okabayashi K, Kinoshita T. Progress in gastric cancer surgery in Japan and its limits of radicality. World J Surg 1987; 11:418-25.
3. Fujii M, Sasaki J, Nakajima T. State of the art in the treatment of gastric cancer: from the 71st Japanese gastric cancer congress. Gastric Cancer 1999; 2:151-7.
4. Nishi M, Omori Y, Miwa K, editors. Japanese classification of gastric carcinoma. Japanese Research Society for Gastric Cancer (JRSGC). 1st English edition. Tokyo: Kanehara & Co., 1995.
5. Bonnenkamp JJ, Hermans J, Sasako M, van de Velde CJH, et al. Extended lymph-node dissection for gastric cancer. N Engl J Med 1999; 340: 908-14.
6. Cuschieri A, Weeden S, Fielding J, et al. Patient survival after D1 and D2 resection for gastric cancer: long-term results of the MRC randomised surgical trial. Surgical co-operation group. Br J Cancer 1999; 79: 1522-30.
7. Siewert JR, Boettcher K, Stein HJ, et al. Relevant prognostic factors in gastric cancer. Ten-year results of the German Gastric cancer Study. Ann Surg 1998; 228: 449-461.
8. Brennan MF. Lymph-node dissection for gastric cancer. New Engl J Med 1999;340:956-8 (editorial).
9. Sano T, Sasako M. Randomized controlled trial to evaluate para-aortic lymphadenectomy for gastric cancer (JCOG 9501): An update. 4th International Gastric Cancer, New York April 29-May 2, 2002, Abstracts, S45, p.663.
10. Roukos DH, Lorenz M, Encke A. Evidence of survival benefit of extended lymphadenectomy in Western gastric cancer patients based on a new concept. A prospective long-term follow-up study. Surgery 1998; 123: 573-578.
11. Roukos DH, Paraschou P, Lorenz M. Distal gastric cancer and extensive surgery: A new evaluation method based on the study of the status of residual lymph nodes after limited surgery. Ann Surg Oncol 2000 Dec;7(10):719-26.
12. Wagner PK, Ramaswamy A, Ruschoff J, Schmitz-Moormann P, Rothmund M. Lymph node counts in the upper abdomen: Anatomical basis for lymphadenectomy in gastric cancer. Br J Surg 1991;78:825-7.
13. Bonnenkamp JJ, Hermans J, Sasako M, Van de Velde CJH. Quality control of lymph node dissection in the Dutch randomized trial of D1 and D2 lymph node dissection for gastric cancer. Gastric Cancer 1998; 1: 152-159.
14. Roukos DH. Extended (D2) lymph node dissection for gastric cancer: do patients benefit? Ann Surg Oncol. 2000 May; 7(4):253-5. Editorial.
15. Bearhs OH, Henson DE, Hutter RVP, Kennedy BJ, editors. Manual for staging. American Joint Committee on Cancer. 4th edition. Philadellphia: J. B. Lippincott, 1992.
16. Hartgrink HH, van de Velden CJH, Dutch Gastric Cancer Group. Update of the Dutch D1 vs D2 Gastric Cancer Trial. 4th International Gastric Cancer Congress, New York, April 29-May 2 2001, Abstracts, Abstract # S47, p.665.
17. Katai H, Marujam K, Sasako M, Sano T. Incidence of nodal metastasis around the superior border of the pancreas based on number of metastatic perigastric nodes. Gastric Cancer 1998;1:115-7.
18. Mari E, Florani I, Tinazzi A, et al. Efficacy of adjuvant chemotherapy after curative resection for gastric cancer; a meta-analysis of published randomized trials. A study of the GISCAD. Ann Oncol 2000; 11: 837-43.
19. Macdonald JS, Smalley SR, Benedetti J, et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001; 345: 725-30.
20. Lorenz M, Roukos DH, Karakostas K, Hottenrott C, Encke A Accurate prediction of site-specific risk of recurrence after curative surgery for gastric cancer. Gastric Breast Cancer 2002; 1(2): 23-32.