- Email: [email protected]
Gastric cancer: which patients benefit from systematic lymphadenectomy?
European Journal of Surgical Oncology 2000; 26: 498–505 doi:10.1053/ejso.1999.0930, available online at http://www.idealibrary.com on
Gastric cancer: which patients benefit from systematic lymphadenectomy? N. M. Bo¨sing, P. E. Goretzki and H. D. Ro¨her Department of General and Trauma Surgery, Heinrich Heine University, Moorenstraße 5, 40225 Du¨sseldorf, Federal Republic Germany
Aims: The purpose of this study was to evaluate the value of systematic lymphadenectomy (SLA) in curative resected gastric cancer patients with respect to long-term survival, peri-operative morbidity and mortality. Methods: We reviewed our prospectively gathered database of 309 resected gastric cancer patients and analysed the outcome of 185 R0-resected patients (60%) with respect to peri-operative morbidity, mortality and long-term survival by comparing 81 patients resected with SLA (D2-group) versus 104 patients resected without SLA (D1-group). Results: Overall 5-year survival rates of R0-resected patients (n=173; exclusion of peri-operative mortality) amounted to 49% and did not differ significantly between D2- and D1-groups (53% vs 47%; P=0.344). Nevertheless, subgroups of patients taking a benefit from SLA could be defined. Gastric cancer patients without LN metastases (pTx pN0; n=78) and patients with LN metastases only in perigastric lymph nodes (pTx pN1; n=34) showed a significantly better long-term prognosis when SLA was performed (84% vs 51%; P=0.001). Regarding peri-operative morbidity (38% vs 39%) and mortality (6% in each case) we could not find any differences between the D2- and D1-groups. Conclusions: We conclude that SLA is able to improve long-term survival for some tumour stages. Therefore SLA should be recommended as a standard procedure in all gastric cancer patients resected with curative intention. 2000 Harcourt Publishers Ltd Key words: gastric cancer; curative resection; systematic lymphadenectomy (sla); long-term survival.
Introduction There is no doubt that complete tumour resection (according to R0-resection of UICC)1 is the strongest prognostic parameter influencing long-term prognosis. In Japan prognosis of gastric cancer has clearly improved in the last 30 years and overall 5-year survival rates of more than 50% are reported today.2,3 However, in western industrial countries long-term prognosis of gastric cancer has barely changed over the same time period. In Europe and the USA fewer than 30% of all gastric cancer patients survive 5 years or more after gastrectomy with curative intention.4,5 The superior results in Japan may partially be explained by a higher proportion of early tumour stages, but also separate analyses for different tumour stages in eastern countries demonstrated a significantly improved long-term prognosis of gastric cancer patients during the last three decades.6 In the past Japanese authors have reported good experiences with SLA, whereas communications from authors in Europe and the USA regarding SLA are still inconsistent, with respect to long-term survival, as well as peri-operative morbidity and mortality.7,8 Moreover, recently reported Correspondence to: Prof. Dr. H.-D. Ro¨her, Department of General and Trauma Surgery, Heinrich Heine University, Moorenstraße 5, 40225 Du¨sseldorf, Federal Republic Germany. Tel: 0211-81-17351; Fax: 0211-81-17359. 0748–7983/00/050498+08 $35.00/0
results from two randomized prospective European multicentre trials demonstrated no survival benefit for patients resected with SLA but an increased peri-operative morbidity and mortality and consequently the value of SLA is questioned again.9,10 However, on the background of frequently existing micrometastases in clinically inconspicuous lymph nodes, which are classified in pathologists’ routine sections as pN0,11,12 discussion about the impact of SLA appears in a quite new light. Therefore we reviewed our gastric cancer patients after curative resection in order to evaluate whether there are patient subgroups that benefit from SLA.
Patients and methods In this analysis of a time period of 10 years (4/86–12/95) the significance of systematic lymphadenectomy (SLA) in 185 curatively resected gastric cancer patients was evaluated. Long-term survival, frequency of peri-operative complications and peri-operative mortality were the main criteria for outcome evaluation. We defined curative resections (R0-resections) as total or subtotal gastrectomies including the resection of the omentum majus and omentum minus with histological tumour free margins on the one hand and with no intraoperative evidence for macroscopic tumour residue on the other hand. Patient and tumour 2000 Harcourt Publishers Ltd
499
Gastric cancer Table 1. Patient related data of all curatively (R0-) resected gastric cancer patients All patients (n=185) LN-dissection
D2
D1
Number of patients (n)
81
104
Sex (%) Male Female
70 30
64 36
Diseases (%) Present Absent
44 56
61 39
Nutrition (%) Malnutrition Normal (±) Obesity
8 82 10
8 77 15
Karnofsky (%) <80% [80%
5 95
12 88
Status (%) Living Deceased
49 51
39 61
Age (years) Median Mean
62 60
66 66
pTx pN0/1 (n=112) P
D2
D1
39
73
0.325 72 28
59 41
44 56
66 34
11 84 5
9 76 16
D1
27
33
70 30
73 27
44 56
51 49
8 85 8
6 75 19
5 95
10 90
4 96
6 94
80 20
47 53
93 7
52 48
60 58
67 67
57 52
57 56
0.589
0.325
0.058
0.183
0.097
0.001
0.008
P
0.842
0.393
0.042
Between April 1986 and December 1995, 309 total or subtotal gastrectomies were performed in gastric cancer
D2
0.024
0.358
Results
P
0.179
0.045
related data were collected prospectively. During the entire time period total or subtotal gastrectomy including SLA of the first and second LN-compartment (D2-dissection) was the basic principle of surgical radicality in our clinic. In consideration of a supposed elevated peri-operative morbidity and maybe even mortality after SLA deviation from this principle was allowed in patients with actual or at least alleged clearly elevated peri-operative risks. Two comparable large groups of patients—either resected without SLA (D1-group: dissection of at least the perigastric lymph nodes (LN 1–6); n=104) or resected with SLA (D2-group: dissection of at least the first and second LN compartment (LN 1–11); n=81)—resulted from this strategy and were analysed. Patient related data (age, sex, frequency of concomitant diseases, Karnofsky index), as well as tumour specific data (tumour localization, Lauren- and TNM-classification, distribution of tumour stages (UICC 1992), number of prepared lymph nodes and lymph node metastases) are shown in Tables 1 and 2. All data were analysed using SPSS, a standard statistical software package. Survival curves, reflecting all observed deaths from all causes, were calculated according to Kaplan–Meier and differences in long-term survival were assessed as statistically significant when P-values were <0.05 (log rank). Significances were calculated by the Mann–Whitney U-test for quantitative data and by chi-squared test for qualitative data. Significance was assumed for P<0.05.
pTx pN0/1 <65 years (n=60)
0.001
0.001
0.222
patients in the department of general surgery. Half of these patients were referred to our clinic in an advanced tumour stage (48% in the stage IIIb and IV). Therefore in 124 patients (40%) only palliative (R1/2-) resections were possible. Resections with curative intention could be performed in 185 patients (60%) with a mean age of 64 years (18–90 years) and a ratio of males to females of 2:1. In accordance to tumour localization and tumour size, total gastrectomies (77%) were more frequent than subtotal gastrectomies (23%). A systematic lymphadenectomy (SLA) was performed in 44% of patients (81/185) with an average number of 36 resected lymph nodes per patient. Fifty-six per cent of patients (104/185) were operated on without SLA, with an average number of 18 lymph nodes per specimen (Tables 1 and 2). The overall complication rate of all R0-resected gastric cancer patients amounted to 38%. In 14% of all cases relaparotomies were indicated due to post-operative complications. The 30 days mortality was 6% and in-hospital mortality was 7%. Patients who underwent SLA did not demonstrate a higher overall complication rate (38% vs 39%) or a higher frequency of surgical (22% vs 30%), pulmonary (12% vs 13%) or cardiovascular complications (3% vs 6%) compared to the group operated on without SLA. No differences in peri-operative mortality between both groups were seen (Table 3). However, multivisceral surgery (mvs) with simultaneous resections of adjacent structures or organs, such as distal oesophagus, spleen or part of pancreas, caused a significant increase of perioperative complications (51% vs 27%), especially surgical (38% vs 16%) and pulmonary complications (20% vs 5%). Subsequently the need for relaparotomies was higher in the mvs-group than in the non-mvs-group. The higher rate of peri-operative complications is also expressed in a
500
N. M. Bo¨sing et al. Table 2. Tumour related data of curatively (R0-) resected gastric cancer patients All patients (n=185)
LN-dissection Number of patients
n
Localization Cardia Fundus, corpus Antrum >2/3 stomach
D2
D1
81
104
32 27 26 15
24 28 39 10
% % % % % % % % % % % %
24 46 28 3 31 24 45 90 28 26 36 11
37 42 16 5 56 20 24 95 54 16 24 7
% % %
37 40 25
49 37 13
LNs prepared
n
36
18
LN metastases
n
pN pM Stage
P
D2
D1
39
73
0.492 % % % %
TNM/UICC pT
All pN0/1-patients (n=112)
pT1 pT2 pT3 pT4 pN0 pN1 pN2/3 pM0 I II III IV
Lauren Intestinal Diffuse/mixed Not classified
pN0/1-patients <65 years (n=60) P
D2
D1
27
33
33 26 33 7
30 33 31 6
37 44 15 4 37 63 — 96 55 30 11 4
45 36 12 6 30 70 — 100 69 15 15 0
48 37 15
45 42 12
36
18
0.314
0.094
<0.001 0.172 0.002
33 28 31 8
23 29 41 7
33 51 13 3 56 44 — 97 48 41 8 3
51 34 12 3 77 23 — 100 74 18 9 0
49 29 23
49 38 12
35
15
0.534
0.260
0.027 0.171 0.043
P
0.913
0.019
0.773
0.585 0.269 0.469
0.332
LNs per patient 5·9
<0.001
3·3
0.003
1·8
1·3
<0.001 0.118
1·5
1·5
<0.001 0.920
Table 3. Comparison of perioperative morbidity and mortality, relaparotomies and hospital stay in 185 curatively resected gastric cancer patients in relation to SLA and extent of surgery SLA
Extent of surgery
All
With
Without
185
81
104
Complications (%) Overall Surgical Cardiac Pulmonary
38 27 4 12
38 22 3 12
39 30 6 13
Relaparotomy (%)
13
11
7 6
16 20
Number of patients (n)
Mortality (%) In hospital In 30 days
Extended
Non-extended
89
96
0.979 0.247 0.275 0.975
51 38 6 20
27 16 3 5
<0.001 <0.001 0.406 0.002
15
0.400
20
7
0.010
6 6
7 6
0.879 0.909
8 9
4 4
0.185 0.289
16 20
16 21
18 24
15 17
P
P
Hospital stay (days) 0.804 Median Mean
significantly increased length of hospital stay. Peri-operative morbidity was twice as high (8% vs 4%) in the mvs-group but without reaching statistical significance (Table 3). Within a long observation period (37–146 months) 44% of all 185 R0-resected gastric cancer patients were still alive at time of the last follow-up (31.01.1999). All following calculations of 5-year survival rates according to the Kaplan–Meier test
<0.001
have been done under exclusion of peri-operative morbidity, which did not differ between D1- and D2-dissected patients. Calculated 5-year survival of all 173 R0-resected gastric cancer patients was rated to be approximately 49%. Long-term prognosis of all patients (n=173) did not show any significant difference between gastrectomies done with or without SLA. Five-year survival rates rose to 53% for
501
Gastric cancer
80
80
60
60 %
100
%
100
40
40
20
20
24
48 72 96 Survival time (months)
120
144
Fig. 1. Prognostic relevance of SLA in 112 curatively resected pTxpN0/1-gastric cancer patients. Patients resected with SLA (Μ: n=39) demonstrated a significantly better (P=0.001) long-term survival than patients resected without SLA (Κ: n=73).
D2- (n=76) and to 47% for D1- (n=97) dissected patients respectively (P=0.344). Particularly in patients with pN2/3lymph node metastases, SLA did not improve long-term prognosis demonstrating 5-year survival rates less than 30% in each group (P=0.941). However, patients without LN metastases (pN0) and patients with restricted LN-involvement (pN1) showed a significantly better long-term prognosis (P=0.001) when SLA was done achieving 5-year survival rates of 84% in patients with SLA and 51% in patients without SLA respectively (Fig. 1). Patients with concomitant diseases were more frequently encountered and the mean age was found to be higher in the D1-group than in the D2-group (Table 1). On the other hand, regarding tumour specific data (Table 2), D1- and D2-groups showed significant differences with respect to pN-category and distribution of tumour stages demonstrating more pT1- and stage-Icarcinomas and less LN metastases in the D1-group. Differences in long-term survival between patients operated on with and without SLA are not only caused by a higher mean age and more frequent occurrence of concomitant diseases in D1-group. When analysing patients younger than 65 years we did not find any significant differences between the D1- and D2-groups (Table 1). In these patients median age amounted to 57 years and the share of patients suffering from concomitant diseases was almost identical in both groups. Tumour specific data were also very similar in both groups (Table 2). Patients operated on with SLA again demonstrated a significantly better long-term prognosis (P= 0.001) in comparison to patients operated on without SLA, showing 5-year survival rates of 92% for D2- and 54% for D1-dissected patients respectively (Fig. 2). The extent of LN metastases is of crucial prognostic importance. Forty-five per cent of our patients did not have LN metastases (pN0) and 55% had histologically verified
0
24
48 72 96 Survival time (months)
120
144
Fig. 2. Prognostic relevance of SLA in gastric cancer patients. In 60 curatively resected pTxpN0/1-gastric cancer patients younger than 65 years of age patients resected with SLA (Μ: n=27) demonstrated a significantly better (P=0.001) long-term survival than patients resected without SLA (Κ: n=33).
100
80
60 %
0
40
20
0
24
48 72 96 Survival time (months)
120
144
Fig. 3. Prognostic relevance of the extent of lymph node metastases in 173 curatively resected gastric cancer patients. Gastric cancer patients suffering from pN0- and pN1-carcinomas demonstrated a significantly better prognosis (P<0.001) than patients suffering from pN2- and pN3-carcinomas (pN3=pM1(Lym)). Φ: pN3 (n=11); Ν: pN2 (n=50); Χ: pN1 (n=34); Α: pN0 (n=78).
LN-metastases; in 22% as pN1- and in 33% as pN2/3involvement (Table 2). The pN0/1-patients showed significantly higher 5-year survival rates than the pN2/3patients (65%, 52%, 26% and 18%, respectively), but 5-year survival rates between pN0- versus pN1-patients (P=0.369) on the one hand and pN2- versus pN3-patients on the other hand (P=0.368) did not differ significantly (Fig. 3). In particular, prognosis of pN1-patients was very similar to prognosis of pN0-patients despite less favourable tumour data in the pN1-group. In the pN0-group 60% of early
502
N. M. Bo¨sing et al. Table 4. Patient and tumour related data in pTxpN0- and pTxpN1-carcinomas Patient related data
Tumour related data pN1
Number of patients (n)
78
34
Sex (%) Male Female
63 37
65 35
Diseases (%) Present Absent
59 41
56 44
0.805
0.762
Nutrition (%) Malnutrition Normal ± Obesity
12 76 12
3 84 13
Karnofsky (%) <80 [80
12 88
0 100
Status (%) Living Deceased
60 40
53 47
Age (years) Median Mean
64 65
63 62
pN0
pN1
78
34
Localization (%) Cardia Fundus, corpus Antrum
21 30 44
41 27 24
>2/3 stomach
6
9
P
0.533
0.735
0.535
0.359
carcinomas (pT1) and 40% of advanced tumours (pT2–4) were found, whereas in the pN1-group early carcinomas (pT1) were only found in 9% and advanced carcinomas (pT2–4) in 91%. Moreover, the frequency of stage-Icarcinoma was 90% in the pN0-group and 9% in the pN1group. With respect to all other parameters (mean age, concomitant diseases, Karnofsky-Index) no differences were found between both groups (Table 4). However, in pN1patients SLA was significantly more often (P<0.05) performed than in pN0-patients (50% vs 28%). Both pN0patients and pN1-patients resected with SLA showed a significantly better long-term prognosis than patients operated on without SLA, demonstrating 5-year survival rates of 91% versus 56% (P<0.001) in pN0-patients (Fig. 4) and 74% versus 34% (P<0.05) in pN1-patients respectively (Fig. 5).
Discussion Up to now surgery represents the only therapy with a potential to cure gastric cancer. However, possibilities of surgery are limited and surgical radicality may be unfeasible due to tumour stage.13 The R0-resection, the complete removal of all tumour tissue, is essential for longterm prognosis.2,4,7,14–16 However, whether systematic lymphadenectomy (SLA) is necessary and to what extent, remains controversial. In the past Japanese authors have reported good experiences with SLA in gastric cancer patients.2,3,6,15,17–20 So far, in the USA and Europe these good results have not been confirmed. In contrast to Japan, where
P
0.062
pT (%) pT1 pT2 pT3 pT4 Tumour-stage (%) I II III IV
60 30 10 0
9 65 18 9
<0.001
90 9 0 1
9 65 27 0
SLA (%) D2-dissection D1-dissection
28 72
50 50
Lauren (%) Intestinal Diffuse/mixed Not class.
53 32 15
41 42 18
<0.001
0.027
0.313
100
80
60 %
pN0
40
20
0
24
48 72 96 Survival time (months)
120
144
Fig. 4. Prognostic relevance of SLA in curatively resected pTxpN0gastric cancer patients (n=78). Patients resected with SLA (Μ: n=22) demonstrated a significantly better (P=0.004) long-term survival than patients resected without SLA (Κ: n=56).
SLA is to a higher extent a routine procedure in resections for gastric cancer operated on with curative intent, in Europe and the USA controversy remains to the benefit and value of SLA. Some authors do not believe that patients may benefit
Gastric cancer 100
80
%
60
40
20
0
24
48 72 96 Survival time (months)
120
144
Fig. 5. Prognosis of 34 curatively resected pTxpN1-gastric cancer patients. Patients resected with SLA (Μ: n=17) showed a significantly better (P=0.038) long-term survival than patients resected without SLA (Κ: n=17).
from SLA, because in their studies SLA did not improve, or minimally improved, post-operative survival time. Moreover, in some studies SLA seemed to be associated with an elevated peri-operative morbidity and mortality.8,21 Bonenkamp et al.9 and Cushieri et al.10 recently published the results of their prospective randomized multicentre trials regarding SLA. A significant increase in peri-operative morbidity and mortality in patients with SLA was one essential observation in both studies. In contrast to these results other western authors in prior studies did not find, or found minimal, differences in peri-operative morbidity and mortality between patients with and without SLA.7,22,23 We alo found in our analysis of 185 curatively resected gastric cancer patients no differences regarding peri-operative morbidity or mortality between patients resected with or without SLA. The increased peri-operative morbidity and mortality, shown by Cushieri et al.10 as well as by Bonenkamp et al.,9 should not be attributed to SLA, but is due to the extent of resection required for sufficient local tumour control. We can confirm this observation with this analysis. If multivisceral surgery was necessary, the overall complication rate was doubled in comparison to resections without extended surgery with a particular increase in surgical and pulmonary complications. In addition 30 days mortality as well as inpatient mortality were also doubled in the multivisceral surgery group. The extent of surgery and not the SLA seems to be linked to elevated peri-operative morbidity and mortality. In the studies of Bonenkamp et al.25 as well as Cushieri et al.10 significant differences in perioperative morbidity and mortality between resections with and without SLA also disappeared when patients who underwent extended surgery, especially resections of the spleen and pancreas, were excluded. Viste et al.24 also showed, that in patients undergoing splenectomy, perioperative morbidity and mortality were doubled.
503
In the Dutch Stomach Carcinoma Study 3-year survival rates rose to 60% for patients with D1-dissections and to 55% for patients with D2-dissections,9,25 and in the British Stomach Carcinoma Study 3-year survival rates rose to 30% for patients with pancreatico–splenectomy and to 50% for patients without splenectomy and without resection of pancreas.10 Consequently both studies could not show a better long-term prognosis for patients operated on with SLA. However, both studies restricted themselves to analysis of only the total of all patients operated on with or without SLA and relinquished on stage related evaluations of tumour data. Regarding all R0-resected gastric cancer patients we also could not show a significant advantage for patients operated on with SLA in comparison to patients resected without SLA. However, in a more detailed analysis subgroups of patients taking benefit from SLA could be defined. In our study SLA was of significant prognostic relevance in pN0- and pN1-gastric cancer patients. In 1993 the results of the German Stomach Carcinoma Study were published. In this study Siewert et al.7 identified SLA as an essential and important prognostic factor in gastric cancer patients in tumour stages II and IIIa. Gall and Hermanek22 demonstrated a significantly better longterm prognosis for gastric cancer patients in stage II and positive trends for patients in tumour stages Ia, Ib and IIIa, when SLA was performed. In the study of Pacelli et al.23 an advantage in long-term survival was noticed in stage III gastric cancer patients resected with SLA. Similar results were reported by Volpe et al.26 In this study long-term prognosis of gastric cancer patients with pT2–4 pN0–1-tumours was improved by SLA. With respect to our study, also mainly patients with negative lymph nodes (pN0) and patients with pN1-LN involvement benefit from SLA. Long-term survival of pTx pN0/1 gastric cancer patients operated on with SLA was significantly better than long-term survival of those patients resected without SLA (P=0.001). In our study the mean age of patients operated on with SLA was lower and the share of patients suffering from concomitant diseases was smaller than in patients operated on without SLA, but better prognosis of D2-dissected patients in comparison to D1-dissected patients should not be misunderstood as only related to biased selection, because on the other hand tumour specific data were significantly less favourable in D2dissected patients than in D1-dissected patients. However, analysis of patients younger than 65 years of age again showed a superior 5-year survival rate when SLA was performed (92% vs 54%; P=0.001) without any differences in concomitant diseases and with an identical median age in both groups. In 1995 Baba et al.27 demonstrated 373 histologically proven LN-negative early gastric cancer patients in whom SLA seemed to improve long-term prognosis, with 5- and 10-year survival rates of 90% and 81% in D1- and 97% and 95% in D2/3-dissected patients. Also in our study pN0-patients operated on with SLA showed a significantly better long-term prognosis than patients operated on without SLA (P=0.004) as well. There is no doubt that an extended lymphadenectomy contributes to the discovery of more LN metastases and hereby SLA causes an ‘up-staging’ of tumours. However, to assume the better prognosis of D2-dissected patients in comparison to D1-dissected patients only results from ‘stage
504
N. M. Bo¨sing et al.
migration phenomenon’ through SLA28–30 appears to be too restricted. The ‘stage migration phenomenon’ should not be considered as the sole cause of these results and in our opinion it is not a sufficient explanation, since it is regarded as unlikely that lymph node metastases jump a lymph node compartment. In less than 2% of all cases LN metastases may be expected in the second LN-compartment, if lymph nodes of the first LN-compartment are tumour free.22 Micrometastases escaping pathologists routine sections may be a possible explanation for the presented results. In a retrospective analysis of 100 gastric cancer patients (pT1–3 pN0–1) who had undergone surgery Siewert et al.12 could show that micro-involvement of the regional lymph nodes occurs very frequently. Both, micro-involvement and micrometastases respectively seem to be of prognostic relevance for long-term survival. The proof of three or more micrometastases in more than 10% of lymph nodes removed was connected with a significantly worse prognosis in pN0patients. In another study Isozaki et al.11 showed by serial sectioning of lymph nodes, which were assessed as tumour free in pathologists’ routine sections, that micrometastases, and especially micrometastases in more distant lymph nodes, occur often. In 1994 Natsugoe et al.31 published the results of their examinations regarding occult LN metastases. Using three additional sections of lymph nodes the rate of LN metastases in early gastric carcinomas involving the submucosal layer (pT1b-gastric cancer) rose from 19% to 30% on re-examination. All three studies may underline the necessity of SLA in gastric cancer patients resected with curative intention and may help to explain why even patients suffering from pTx pN0-gastric carcinomas may benefit from SLA.
Conclusion We conclude from our results that SLA is able to improve long-term prognosis in gastric cancer patients in some tumour stages without any increase of peri-operative morbidity or mortality. In our opinion, the removal of the first and second LN compartment is an essential part of surgical therapy in gastric cancer patients in order to avoid a false R0-resection. Therefore D2-dissection should be the standard of radicality in each gastric cancer patient operated on with curative intention, although it is likely that only some patients will profit from this procedure. Following our results particularly patients without (pN0) and with only perigastric (pN1) lymph node metastases (mainly patients in UICC stages Ia, Ib and II and possibly IIIa) will benefit from SLA. In each case SLA guarantees a more realistic N-classification, a more exact assessment of tumour stages and consequently enables a better judgement of long-term prognosis.
References 1. Hermanek P, Sobin LH. International Union Against Cancer (UICC), TNM classification of Malignant Tumours. 4th edn. 2nd revision. Berlin: Springer 1992.
2. Maruyama 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. Noguchi Y, Imada T, Matsumoto A, et al. Radical surgery for gastric cancer. A review of the Japanese experience. Cancer 1989; 64: 2053–62. 4. Meyers WC, Damiano RJ, Postlethwait RW, Rotolo FS. Adenocarcinoma of the stomach: changing patterns over the last 4 decades. Ann Surg 1987; 205: 1–8. 5. Wanebo HJ, Kennedy BJ, Chmiel J. Cancer of the stomach: a patient care study by the American College of Surgeons. Ann Surg 1993; 218: 583–92. 6. Maruyama K, Sasako M, Kinoshita T. Role of systematic extended lymph node dissection: Japanese experience. Langenbecks Arch Surg 1992; Suppl. II: 130–5. 7. Siewert JR, Bo¨ttcher K, Roder JD, et al. Prognostic relevance of systematic lymph node dissection in gastric carcinoma. Br J Surg 1993; 80: 1015–8. 8. Wanebo HJ, Kennedy BJ, Winchester DP, et al. Gastric carcinoma: does lymph node dissection alter survival? J Am Coll Surg 1996; 183: 616–24. 9. Bonenkamp JJ, Van De Velde CJH, Songun I, et al. Randomized comparison of morbidity after D1- and D2-dissection for gastric cancer in 996 Dutch patients. Lancet 1995; 345: 745–8. 10. Cushieri A, Fayers P, Fielding J, et al. Postoperative morbidity and mortality after D1 and D2 resections for gastric cancer: preliminary results of the MRC randomised controlled surgical trial. Lancet 1996; 347: 995–9. 11. Isozaki H, Okajima K, Fujii K. Histological evaluation of lymph node metastasis on serial sectioning in gastric cancer with radical lymphadenectomy. Hepatogastroenterology 1997; 44: 1133–6. 12. Siewert JR, Kestlmeier R, Busch R, et al. Benefits of D2 lymph node dissection for patients with gastric cancer and pN0 and pN1 lymph node metastases. Br J Surg 1996; 83: 1144–7. 13. Meyer HJ, Ja¨hne J, Pichlmayr R. Strategies in the surgical treatment of gastric carcinoma. Ann Oncol 1994; Suppl. 3: 33–6. 14. Jatzko GR, Lisborg PH, Denk H, et al. A 10-year experience with Japanese-type radical lymph node dissection for gastric cancer outside of Japan. Cancer 1995; 76: 1302–12. 15. Lee WJ, Lee WC, Houng SJ, et al. Survival after resection of gastric cancer and prognostic relevance of systematic lymph node dissection: twenty years experience in Taiwan. World J Surg 1995; 19: 707–13. 16. Meyer HJ, Ja¨hne J, Wilke H, Pichlmayr R. Operative Behandlung des Magenkarzinoms. Langenbecks Arch Surg 1990; Suppl. II: 117–23. 17. Inada T, Ogata Y, Andoh J, et al. Significance of paraaortic lymph node dissection in patients with advanced and recurrent gastric cancer. Anticancer Res 1994; 14: 677–82. 18. Kodama Y, Sugimachi K, Soejima K, et al. Evaluation of extensive lymph node dissection for carcinoma of the stomach. World J Surg 1981; 5: 241–8. 19. Soga J, Kobayashi K, Saito J, et al. The role of lymphadenectomy in curative surgery for gastric cancer. World J Surg 1979; 3: 701–8. 20. Soga J, Ohyama S, Miyashita K, et al. Statistical evaluation of advancement in gastric cancer surgery with special reference to the significance of lymphadenectomy for cure. World J Surg 1988; 12: 398–405. 21. Robertson CS, Chung SC, Woods SD, et al. A prospective randomized trial comparing R1 subtotal gastrectomy with R3 total gastrectomy for antral cancer. Ann Surg 1994; 220: 176–82. 22. Gall FP, Hermanek P Die systematische erweiterte Lymphknoten-Dissektion in der kurativen Therapie des Magenkarzinoms. Chirurg 1993; 64: 1024–31. 23. Pacelli F, Doglietto GB, Bellantone R, et al. Extensive versus limited lymph node dissection for gastric cancer: a comparative study of 320 patients. Br J Surg 1993; 80: 1153–6. 24. Viste A, Svanes K, Janssen CW, et al. Prognostic importance of radical lymphadenectomy in curative resections for gastric cancer. Eur J Surg 1994; 160: 497–502.
Gastric cancer 25. Bonenkamp JJ, Van De Velde CJH, Hermans J. Randomised trial of extended lymph node dissection for gastric cancer. In: JR Siewert, JD Roder (eds). Progress in Gastric Cancer Research—Proceedings of the 2nd IGCC. Bologna: Monduzzi Editore S.p.A., 1997: 1111–1116. 26. Volpe CM, Koo J, Miloro SM, et al. The effect of extended lymphadenectomy on survival in patients with gastric adenocarcinoma. J Am Col Surg 1995; 181: 56–64. 27. Baba H, Maehara Y, Takauchi H, et al. Effect of lymph node dissection on the prognosis in patients with node negative early gastric cancer. Surgery 1995; 117: 165–9. 28. Ja¨hne J, Tusch G, Breitmeier D, et al. Will Rogers—Not just a phenomenon: Stage migration as an explanation for the improved prognosis after D2-lymphadenectomy in
505
gastric carcinoma. Langenbecks Arch Surg 1996; Suppl. I: 383–7. 29. Bunt AMG, Hermans J, Smit VTH, et al. Surgical/pathologicalstage migration confounds comparisons of gastric cancer survival between Japan and Western countries. J Clin Oncol 1995; 13: 19–25. 30. Feinstein AR, Sosin DM, Wells CK. The Will Rogers phenomenon—Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med 1985; 312: 1604–8. 31. Natsugoe S, Aiko T, Shimada M, et al. Occult lymph node metastases in gastric cancer with submucosal invasion. Surg Today 1994; 24: 870–5. Accepted for publication 12 November 1999
European Society of Surgical Oncology FELLOWSHIPS The Education and Training Committee of ESSO is pleased to offer a number of Training Fellowships (Minor) to support trainees in surgical oncology. The aim of these followships is to allow young surgeons to spend a time in another specialist centre to either expand their experience or learn new techniques. Applicants require support from their own Head of Department and also from the Head of Department of the Centre which they wish to visit. Fellowships are offered for short periods of training, i.e. 2–3 weeks’ duration and will be supported at the level of 2500 EUR.
In addition, one Major Fellowship to the value of 6.250 EUR will be offered to support an individual who wishes to spend 3 months in a specialist centre to expand his or her clinical or research experience in surgical oncology. The successful applicant will be expected to write a short paper describing their training experiences and this may be published in the European Journal of Surgical Oncology. Deadline for all applications: 22nd September 2000 Applications, including a full CV, letters of support and an outline of what the applicant hopes to gain from the Training Fellowships should be sent to: Mr. James Geraghty Professorial Unit of Surgery Nottingham City Hospital Nottingham, NG5 1PB, UK Fax: +44 115 8402673 Email: [email protected] Further information on our website: www.esso-surgeonline.be
Gastric cancer: which patients benefit from systematic lymphadenectomy? N. M. Bo¨sing, P. E. Goretzki and H. D. Ro¨her Department of General and Trauma Surgery, Heinrich Heine University, Moorenstraße 5, 40225 Du¨sseldorf, Federal Republic Germany
Aims: The purpose of this study was to evaluate the value of systematic lymphadenectomy (SLA) in curative resected gastric cancer patients with respect to long-term survival, peri-operative morbidity and mortality. Methods: We reviewed our prospectively gathered database of 309 resected gastric cancer patients and analysed the outcome of 185 R0-resected patients (60%) with respect to peri-operative morbidity, mortality and long-term survival by comparing 81 patients resected with SLA (D2-group) versus 104 patients resected without SLA (D1-group). Results: Overall 5-year survival rates of R0-resected patients (n=173; exclusion of peri-operative mortality) amounted to 49% and did not differ significantly between D2- and D1-groups (53% vs 47%; P=0.344). Nevertheless, subgroups of patients taking a benefit from SLA could be defined. Gastric cancer patients without LN metastases (pTx pN0; n=78) and patients with LN metastases only in perigastric lymph nodes (pTx pN1; n=34) showed a significantly better long-term prognosis when SLA was performed (84% vs 51%; P=0.001). Regarding peri-operative morbidity (38% vs 39%) and mortality (6% in each case) we could not find any differences between the D2- and D1-groups. Conclusions: We conclude that SLA is able to improve long-term survival for some tumour stages. Therefore SLA should be recommended as a standard procedure in all gastric cancer patients resected with curative intention. 2000 Harcourt Publishers Ltd Key words: gastric cancer; curative resection; systematic lymphadenectomy (sla); long-term survival.
Introduction There is no doubt that complete tumour resection (according to R0-resection of UICC)1 is the strongest prognostic parameter influencing long-term prognosis. In Japan prognosis of gastric cancer has clearly improved in the last 30 years and overall 5-year survival rates of more than 50% are reported today.2,3 However, in western industrial countries long-term prognosis of gastric cancer has barely changed over the same time period. In Europe and the USA fewer than 30% of all gastric cancer patients survive 5 years or more after gastrectomy with curative intention.4,5 The superior results in Japan may partially be explained by a higher proportion of early tumour stages, but also separate analyses for different tumour stages in eastern countries demonstrated a significantly improved long-term prognosis of gastric cancer patients during the last three decades.6 In the past Japanese authors have reported good experiences with SLA, whereas communications from authors in Europe and the USA regarding SLA are still inconsistent, with respect to long-term survival, as well as peri-operative morbidity and mortality.7,8 Moreover, recently reported Correspondence to: Prof. Dr. H.-D. Ro¨her, Department of General and Trauma Surgery, Heinrich Heine University, Moorenstraße 5, 40225 Du¨sseldorf, Federal Republic Germany. Tel: 0211-81-17351; Fax: 0211-81-17359. 0748–7983/00/050498+08 $35.00/0
results from two randomized prospective European multicentre trials demonstrated no survival benefit for patients resected with SLA but an increased peri-operative morbidity and mortality and consequently the value of SLA is questioned again.9,10 However, on the background of frequently existing micrometastases in clinically inconspicuous lymph nodes, which are classified in pathologists’ routine sections as pN0,11,12 discussion about the impact of SLA appears in a quite new light. Therefore we reviewed our gastric cancer patients after curative resection in order to evaluate whether there are patient subgroups that benefit from SLA.
Patients and methods In this analysis of a time period of 10 years (4/86–12/95) the significance of systematic lymphadenectomy (SLA) in 185 curatively resected gastric cancer patients was evaluated. Long-term survival, frequency of peri-operative complications and peri-operative mortality were the main criteria for outcome evaluation. We defined curative resections (R0-resections) as total or subtotal gastrectomies including the resection of the omentum majus and omentum minus with histological tumour free margins on the one hand and with no intraoperative evidence for macroscopic tumour residue on the other hand. Patient and tumour 2000 Harcourt Publishers Ltd
499
Gastric cancer Table 1. Patient related data of all curatively (R0-) resected gastric cancer patients All patients (n=185) LN-dissection
D2
D1
Number of patients (n)
81
104
Sex (%) Male Female
70 30
64 36
Diseases (%) Present Absent
44 56
61 39
Nutrition (%) Malnutrition Normal (±) Obesity
8 82 10
8 77 15
Karnofsky (%) <80% [80%
5 95
12 88
Status (%) Living Deceased
49 51
39 61
Age (years) Median Mean
62 60
66 66
pTx pN0/1 (n=112) P
D2
D1
39
73
0.325 72 28
59 41
44 56
66 34
11 84 5
9 76 16
D1
27
33
70 30
73 27
44 56
51 49
8 85 8
6 75 19
5 95
10 90
4 96
6 94
80 20
47 53
93 7
52 48
60 58
67 67
57 52
57 56
0.589
0.325
0.058
0.183
0.097
0.001
0.008
P
0.842
0.393
0.042
Between April 1986 and December 1995, 309 total or subtotal gastrectomies were performed in gastric cancer
D2
0.024
0.358
Results
P
0.179
0.045
related data were collected prospectively. During the entire time period total or subtotal gastrectomy including SLA of the first and second LN-compartment (D2-dissection) was the basic principle of surgical radicality in our clinic. In consideration of a supposed elevated peri-operative morbidity and maybe even mortality after SLA deviation from this principle was allowed in patients with actual or at least alleged clearly elevated peri-operative risks. Two comparable large groups of patients—either resected without SLA (D1-group: dissection of at least the perigastric lymph nodes (LN 1–6); n=104) or resected with SLA (D2-group: dissection of at least the first and second LN compartment (LN 1–11); n=81)—resulted from this strategy and were analysed. Patient related data (age, sex, frequency of concomitant diseases, Karnofsky index), as well as tumour specific data (tumour localization, Lauren- and TNM-classification, distribution of tumour stages (UICC 1992), number of prepared lymph nodes and lymph node metastases) are shown in Tables 1 and 2. All data were analysed using SPSS, a standard statistical software package. Survival curves, reflecting all observed deaths from all causes, were calculated according to Kaplan–Meier and differences in long-term survival were assessed as statistically significant when P-values were <0.05 (log rank). Significances were calculated by the Mann–Whitney U-test for quantitative data and by chi-squared test for qualitative data. Significance was assumed for P<0.05.
pTx pN0/1 <65 years (n=60)
0.001
0.001
0.222
patients in the department of general surgery. Half of these patients were referred to our clinic in an advanced tumour stage (48% in the stage IIIb and IV). Therefore in 124 patients (40%) only palliative (R1/2-) resections were possible. Resections with curative intention could be performed in 185 patients (60%) with a mean age of 64 years (18–90 years) and a ratio of males to females of 2:1. In accordance to tumour localization and tumour size, total gastrectomies (77%) were more frequent than subtotal gastrectomies (23%). A systematic lymphadenectomy (SLA) was performed in 44% of patients (81/185) with an average number of 36 resected lymph nodes per patient. Fifty-six per cent of patients (104/185) were operated on without SLA, with an average number of 18 lymph nodes per specimen (Tables 1 and 2). The overall complication rate of all R0-resected gastric cancer patients amounted to 38%. In 14% of all cases relaparotomies were indicated due to post-operative complications. The 30 days mortality was 6% and in-hospital mortality was 7%. Patients who underwent SLA did not demonstrate a higher overall complication rate (38% vs 39%) or a higher frequency of surgical (22% vs 30%), pulmonary (12% vs 13%) or cardiovascular complications (3% vs 6%) compared to the group operated on without SLA. No differences in peri-operative mortality between both groups were seen (Table 3). However, multivisceral surgery (mvs) with simultaneous resections of adjacent structures or organs, such as distal oesophagus, spleen or part of pancreas, caused a significant increase of perioperative complications (51% vs 27%), especially surgical (38% vs 16%) and pulmonary complications (20% vs 5%). Subsequently the need for relaparotomies was higher in the mvs-group than in the non-mvs-group. The higher rate of peri-operative complications is also expressed in a
500
N. M. Bo¨sing et al. Table 2. Tumour related data of curatively (R0-) resected gastric cancer patients All patients (n=185)
LN-dissection Number of patients
n
Localization Cardia Fundus, corpus Antrum >2/3 stomach
D2
D1
81
104
32 27 26 15
24 28 39 10
% % % % % % % % % % % %
24 46 28 3 31 24 45 90 28 26 36 11
37 42 16 5 56 20 24 95 54 16 24 7
% % %
37 40 25
49 37 13
LNs prepared
n
36
18
LN metastases
n
pN pM Stage
P
D2
D1
39
73
0.492 % % % %
TNM/UICC pT
All pN0/1-patients (n=112)
pT1 pT2 pT3 pT4 pN0 pN1 pN2/3 pM0 I II III IV
Lauren Intestinal Diffuse/mixed Not classified
pN0/1-patients <65 years (n=60) P
D2
D1
27
33
33 26 33 7
30 33 31 6
37 44 15 4 37 63 — 96 55 30 11 4
45 36 12 6 30 70 — 100 69 15 15 0
48 37 15
45 42 12
36
18
0.314
0.094
<0.001 0.172 0.002
33 28 31 8
23 29 41 7
33 51 13 3 56 44 — 97 48 41 8 3
51 34 12 3 77 23 — 100 74 18 9 0
49 29 23
49 38 12
35
15
0.534
0.260
0.027 0.171 0.043
P
0.913
0.019
0.773
0.585 0.269 0.469
0.332
LNs per patient 5·9
<0.001
3·3
0.003
1·8
1·3
<0.001 0.118
1·5
1·5
<0.001 0.920
Table 3. Comparison of perioperative morbidity and mortality, relaparotomies and hospital stay in 185 curatively resected gastric cancer patients in relation to SLA and extent of surgery SLA
Extent of surgery
All
With
Without
185
81
104
Complications (%) Overall Surgical Cardiac Pulmonary
38 27 4 12
38 22 3 12
39 30 6 13
Relaparotomy (%)
13
11
7 6
16 20
Number of patients (n)
Mortality (%) In hospital In 30 days
Extended
Non-extended
89
96
0.979 0.247 0.275 0.975
51 38 6 20
27 16 3 5
<0.001 <0.001 0.406 0.002
15
0.400
20
7
0.010
6 6
7 6
0.879 0.909
8 9
4 4
0.185 0.289
16 20
16 21
18 24
15 17
P
P
Hospital stay (days) 0.804 Median Mean
significantly increased length of hospital stay. Peri-operative morbidity was twice as high (8% vs 4%) in the mvs-group but without reaching statistical significance (Table 3). Within a long observation period (37–146 months) 44% of all 185 R0-resected gastric cancer patients were still alive at time of the last follow-up (31.01.1999). All following calculations of 5-year survival rates according to the Kaplan–Meier test
<0.001
have been done under exclusion of peri-operative morbidity, which did not differ between D1- and D2-dissected patients. Calculated 5-year survival of all 173 R0-resected gastric cancer patients was rated to be approximately 49%. Long-term prognosis of all patients (n=173) did not show any significant difference between gastrectomies done with or without SLA. Five-year survival rates rose to 53% for
501
Gastric cancer
80
80
60
60 %
100
%
100
40
40
20
20
24
48 72 96 Survival time (months)
120
144
Fig. 1. Prognostic relevance of SLA in 112 curatively resected pTxpN0/1-gastric cancer patients. Patients resected with SLA (Μ: n=39) demonstrated a significantly better (P=0.001) long-term survival than patients resected without SLA (Κ: n=73).
D2- (n=76) and to 47% for D1- (n=97) dissected patients respectively (P=0.344). Particularly in patients with pN2/3lymph node metastases, SLA did not improve long-term prognosis demonstrating 5-year survival rates less than 30% in each group (P=0.941). However, patients without LN metastases (pN0) and patients with restricted LN-involvement (pN1) showed a significantly better long-term prognosis (P=0.001) when SLA was done achieving 5-year survival rates of 84% in patients with SLA and 51% in patients without SLA respectively (Fig. 1). Patients with concomitant diseases were more frequently encountered and the mean age was found to be higher in the D1-group than in the D2-group (Table 1). On the other hand, regarding tumour specific data (Table 2), D1- and D2-groups showed significant differences with respect to pN-category and distribution of tumour stages demonstrating more pT1- and stage-Icarcinomas and less LN metastases in the D1-group. Differences in long-term survival between patients operated on with and without SLA are not only caused by a higher mean age and more frequent occurrence of concomitant diseases in D1-group. When analysing patients younger than 65 years we did not find any significant differences between the D1- and D2-groups (Table 1). In these patients median age amounted to 57 years and the share of patients suffering from concomitant diseases was almost identical in both groups. Tumour specific data were also very similar in both groups (Table 2). Patients operated on with SLA again demonstrated a significantly better long-term prognosis (P= 0.001) in comparison to patients operated on without SLA, showing 5-year survival rates of 92% for D2- and 54% for D1-dissected patients respectively (Fig. 2). The extent of LN metastases is of crucial prognostic importance. Forty-five per cent of our patients did not have LN metastases (pN0) and 55% had histologically verified
0
24
48 72 96 Survival time (months)
120
144
Fig. 2. Prognostic relevance of SLA in gastric cancer patients. In 60 curatively resected pTxpN0/1-gastric cancer patients younger than 65 years of age patients resected with SLA (Μ: n=27) demonstrated a significantly better (P=0.001) long-term survival than patients resected without SLA (Κ: n=33).
100
80
60 %
0
40
20
0
24
48 72 96 Survival time (months)
120
144
Fig. 3. Prognostic relevance of the extent of lymph node metastases in 173 curatively resected gastric cancer patients. Gastric cancer patients suffering from pN0- and pN1-carcinomas demonstrated a significantly better prognosis (P<0.001) than patients suffering from pN2- and pN3-carcinomas (pN3=pM1(Lym)). Φ: pN3 (n=11); Ν: pN2 (n=50); Χ: pN1 (n=34); Α: pN0 (n=78).
LN-metastases; in 22% as pN1- and in 33% as pN2/3involvement (Table 2). The pN0/1-patients showed significantly higher 5-year survival rates than the pN2/3patients (65%, 52%, 26% and 18%, respectively), but 5-year survival rates between pN0- versus pN1-patients (P=0.369) on the one hand and pN2- versus pN3-patients on the other hand (P=0.368) did not differ significantly (Fig. 3). In particular, prognosis of pN1-patients was very similar to prognosis of pN0-patients despite less favourable tumour data in the pN1-group. In the pN0-group 60% of early
502
N. M. Bo¨sing et al. Table 4. Patient and tumour related data in pTxpN0- and pTxpN1-carcinomas Patient related data
Tumour related data pN1
Number of patients (n)
78
34
Sex (%) Male Female
63 37
65 35
Diseases (%) Present Absent
59 41
56 44
0.805
0.762
Nutrition (%) Malnutrition Normal ± Obesity
12 76 12
3 84 13
Karnofsky (%) <80 [80
12 88
0 100
Status (%) Living Deceased
60 40
53 47
Age (years) Median Mean
64 65
63 62
pN0
pN1
78
34
Localization (%) Cardia Fundus, corpus Antrum
21 30 44
41 27 24
>2/3 stomach
6
9
P
0.533
0.735
0.535
0.359
carcinomas (pT1) and 40% of advanced tumours (pT2–4) were found, whereas in the pN1-group early carcinomas (pT1) were only found in 9% and advanced carcinomas (pT2–4) in 91%. Moreover, the frequency of stage-Icarcinoma was 90% in the pN0-group and 9% in the pN1group. With respect to all other parameters (mean age, concomitant diseases, Karnofsky-Index) no differences were found between both groups (Table 4). However, in pN1patients SLA was significantly more often (P<0.05) performed than in pN0-patients (50% vs 28%). Both pN0patients and pN1-patients resected with SLA showed a significantly better long-term prognosis than patients operated on without SLA, demonstrating 5-year survival rates of 91% versus 56% (P<0.001) in pN0-patients (Fig. 4) and 74% versus 34% (P<0.05) in pN1-patients respectively (Fig. 5).
Discussion Up to now surgery represents the only therapy with a potential to cure gastric cancer. However, possibilities of surgery are limited and surgical radicality may be unfeasible due to tumour stage.13 The R0-resection, the complete removal of all tumour tissue, is essential for longterm prognosis.2,4,7,14–16 However, whether systematic lymphadenectomy (SLA) is necessary and to what extent, remains controversial. In the past Japanese authors have reported good experiences with SLA in gastric cancer patients.2,3,6,15,17–20 So far, in the USA and Europe these good results have not been confirmed. In contrast to Japan, where
P
0.062
pT (%) pT1 pT2 pT3 pT4 Tumour-stage (%) I II III IV
60 30 10 0
9 65 18 9
<0.001
90 9 0 1
9 65 27 0
SLA (%) D2-dissection D1-dissection
28 72
50 50
Lauren (%) Intestinal Diffuse/mixed Not class.
53 32 15
41 42 18
<0.001
0.027
0.313
100
80
60 %
pN0
40
20
0
24
48 72 96 Survival time (months)
120
144
Fig. 4. Prognostic relevance of SLA in curatively resected pTxpN0gastric cancer patients (n=78). Patients resected with SLA (Μ: n=22) demonstrated a significantly better (P=0.004) long-term survival than patients resected without SLA (Κ: n=56).
SLA is to a higher extent a routine procedure in resections for gastric cancer operated on with curative intent, in Europe and the USA controversy remains to the benefit and value of SLA. Some authors do not believe that patients may benefit
Gastric cancer 100
80
%
60
40
20
0
24
48 72 96 Survival time (months)
120
144
Fig. 5. Prognosis of 34 curatively resected pTxpN1-gastric cancer patients. Patients resected with SLA (Μ: n=17) showed a significantly better (P=0.038) long-term survival than patients resected without SLA (Κ: n=17).
from SLA, because in their studies SLA did not improve, or minimally improved, post-operative survival time. Moreover, in some studies SLA seemed to be associated with an elevated peri-operative morbidity and mortality.8,21 Bonenkamp et al.9 and Cushieri et al.10 recently published the results of their prospective randomized multicentre trials regarding SLA. A significant increase in peri-operative morbidity and mortality in patients with SLA was one essential observation in both studies. In contrast to these results other western authors in prior studies did not find, or found minimal, differences in peri-operative morbidity and mortality between patients with and without SLA.7,22,23 We alo found in our analysis of 185 curatively resected gastric cancer patients no differences regarding peri-operative morbidity or mortality between patients resected with or without SLA. The increased peri-operative morbidity and mortality, shown by Cushieri et al.10 as well as by Bonenkamp et al.,9 should not be attributed to SLA, but is due to the extent of resection required for sufficient local tumour control. We can confirm this observation with this analysis. If multivisceral surgery was necessary, the overall complication rate was doubled in comparison to resections without extended surgery with a particular increase in surgical and pulmonary complications. In addition 30 days mortality as well as inpatient mortality were also doubled in the multivisceral surgery group. The extent of surgery and not the SLA seems to be linked to elevated peri-operative morbidity and mortality. In the studies of Bonenkamp et al.25 as well as Cushieri et al.10 significant differences in perioperative morbidity and mortality between resections with and without SLA also disappeared when patients who underwent extended surgery, especially resections of the spleen and pancreas, were excluded. Viste et al.24 also showed, that in patients undergoing splenectomy, perioperative morbidity and mortality were doubled.
503
In the Dutch Stomach Carcinoma Study 3-year survival rates rose to 60% for patients with D1-dissections and to 55% for patients with D2-dissections,9,25 and in the British Stomach Carcinoma Study 3-year survival rates rose to 30% for patients with pancreatico–splenectomy and to 50% for patients without splenectomy and without resection of pancreas.10 Consequently both studies could not show a better long-term prognosis for patients operated on with SLA. However, both studies restricted themselves to analysis of only the total of all patients operated on with or without SLA and relinquished on stage related evaluations of tumour data. Regarding all R0-resected gastric cancer patients we also could not show a significant advantage for patients operated on with SLA in comparison to patients resected without SLA. However, in a more detailed analysis subgroups of patients taking benefit from SLA could be defined. In our study SLA was of significant prognostic relevance in pN0- and pN1-gastric cancer patients. In 1993 the results of the German Stomach Carcinoma Study were published. In this study Siewert et al.7 identified SLA as an essential and important prognostic factor in gastric cancer patients in tumour stages II and IIIa. Gall and Hermanek22 demonstrated a significantly better longterm prognosis for gastric cancer patients in stage II and positive trends for patients in tumour stages Ia, Ib and IIIa, when SLA was performed. In the study of Pacelli et al.23 an advantage in long-term survival was noticed in stage III gastric cancer patients resected with SLA. Similar results were reported by Volpe et al.26 In this study long-term prognosis of gastric cancer patients with pT2–4 pN0–1-tumours was improved by SLA. With respect to our study, also mainly patients with negative lymph nodes (pN0) and patients with pN1-LN involvement benefit from SLA. Long-term survival of pTx pN0/1 gastric cancer patients operated on with SLA was significantly better than long-term survival of those patients resected without SLA (P=0.001). In our study the mean age of patients operated on with SLA was lower and the share of patients suffering from concomitant diseases was smaller than in patients operated on without SLA, but better prognosis of D2-dissected patients in comparison to D1-dissected patients should not be misunderstood as only related to biased selection, because on the other hand tumour specific data were significantly less favourable in D2dissected patients than in D1-dissected patients. However, analysis of patients younger than 65 years of age again showed a superior 5-year survival rate when SLA was performed (92% vs 54%; P=0.001) without any differences in concomitant diseases and with an identical median age in both groups. In 1995 Baba et al.27 demonstrated 373 histologically proven LN-negative early gastric cancer patients in whom SLA seemed to improve long-term prognosis, with 5- and 10-year survival rates of 90% and 81% in D1- and 97% and 95% in D2/3-dissected patients. Also in our study pN0-patients operated on with SLA showed a significantly better long-term prognosis than patients operated on without SLA (P=0.004) as well. There is no doubt that an extended lymphadenectomy contributes to the discovery of more LN metastases and hereby SLA causes an ‘up-staging’ of tumours. However, to assume the better prognosis of D2-dissected patients in comparison to D1-dissected patients only results from ‘stage
504
N. M. Bo¨sing et al.
migration phenomenon’ through SLA28–30 appears to be too restricted. The ‘stage migration phenomenon’ should not be considered as the sole cause of these results and in our opinion it is not a sufficient explanation, since it is regarded as unlikely that lymph node metastases jump a lymph node compartment. In less than 2% of all cases LN metastases may be expected in the second LN-compartment, if lymph nodes of the first LN-compartment are tumour free.22 Micrometastases escaping pathologists routine sections may be a possible explanation for the presented results. In a retrospective analysis of 100 gastric cancer patients (pT1–3 pN0–1) who had undergone surgery Siewert et al.12 could show that micro-involvement of the regional lymph nodes occurs very frequently. Both, micro-involvement and micrometastases respectively seem to be of prognostic relevance for long-term survival. The proof of three or more micrometastases in more than 10% of lymph nodes removed was connected with a significantly worse prognosis in pN0patients. In another study Isozaki et al.11 showed by serial sectioning of lymph nodes, which were assessed as tumour free in pathologists’ routine sections, that micrometastases, and especially micrometastases in more distant lymph nodes, occur often. In 1994 Natsugoe et al.31 published the results of their examinations regarding occult LN metastases. Using three additional sections of lymph nodes the rate of LN metastases in early gastric carcinomas involving the submucosal layer (pT1b-gastric cancer) rose from 19% to 30% on re-examination. All three studies may underline the necessity of SLA in gastric cancer patients resected with curative intention and may help to explain why even patients suffering from pTx pN0-gastric carcinomas may benefit from SLA.
Conclusion We conclude from our results that SLA is able to improve long-term prognosis in gastric cancer patients in some tumour stages without any increase of peri-operative morbidity or mortality. In our opinion, the removal of the first and second LN compartment is an essential part of surgical therapy in gastric cancer patients in order to avoid a false R0-resection. Therefore D2-dissection should be the standard of radicality in each gastric cancer patient operated on with curative intention, although it is likely that only some patients will profit from this procedure. Following our results particularly patients without (pN0) and with only perigastric (pN1) lymph node metastases (mainly patients in UICC stages Ia, Ib and II and possibly IIIa) will benefit from SLA. In each case SLA guarantees a more realistic N-classification, a more exact assessment of tumour stages and consequently enables a better judgement of long-term prognosis.
References 1. Hermanek P, Sobin LH. International Union Against Cancer (UICC), TNM classification of Malignant Tumours. 4th edn. 2nd revision. Berlin: Springer 1992.
2. Maruyama 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. Noguchi Y, Imada T, Matsumoto A, et al. Radical surgery for gastric cancer. A review of the Japanese experience. Cancer 1989; 64: 2053–62. 4. Meyers WC, Damiano RJ, Postlethwait RW, Rotolo FS. Adenocarcinoma of the stomach: changing patterns over the last 4 decades. Ann Surg 1987; 205: 1–8. 5. Wanebo HJ, Kennedy BJ, Chmiel J. Cancer of the stomach: a patient care study by the American College of Surgeons. Ann Surg 1993; 218: 583–92. 6. Maruyama K, Sasako M, Kinoshita T. Role of systematic extended lymph node dissection: Japanese experience. Langenbecks Arch Surg 1992; Suppl. II: 130–5. 7. Siewert JR, Bo¨ttcher K, Roder JD, et al. Prognostic relevance of systematic lymph node dissection in gastric carcinoma. Br J Surg 1993; 80: 1015–8. 8. Wanebo HJ, Kennedy BJ, Winchester DP, et al. Gastric carcinoma: does lymph node dissection alter survival? J Am Coll Surg 1996; 183: 616–24. 9. Bonenkamp JJ, Van De Velde CJH, Songun I, et al. Randomized comparison of morbidity after D1- and D2-dissection for gastric cancer in 996 Dutch patients. Lancet 1995; 345: 745–8. 10. Cushieri A, Fayers P, Fielding J, et al. Postoperative morbidity and mortality after D1 and D2 resections for gastric cancer: preliminary results of the MRC randomised controlled surgical trial. Lancet 1996; 347: 995–9. 11. Isozaki H, Okajima K, Fujii K. Histological evaluation of lymph node metastasis on serial sectioning in gastric cancer with radical lymphadenectomy. Hepatogastroenterology 1997; 44: 1133–6. 12. Siewert JR, Kestlmeier R, Busch R, et al. Benefits of D2 lymph node dissection for patients with gastric cancer and pN0 and pN1 lymph node metastases. Br J Surg 1996; 83: 1144–7. 13. Meyer HJ, Ja¨hne J, Pichlmayr R. Strategies in the surgical treatment of gastric carcinoma. Ann Oncol 1994; Suppl. 3: 33–6. 14. Jatzko GR, Lisborg PH, Denk H, et al. A 10-year experience with Japanese-type radical lymph node dissection for gastric cancer outside of Japan. Cancer 1995; 76: 1302–12. 15. Lee WJ, Lee WC, Houng SJ, et al. Survival after resection of gastric cancer and prognostic relevance of systematic lymph node dissection: twenty years experience in Taiwan. World J Surg 1995; 19: 707–13. 16. Meyer HJ, Ja¨hne J, Wilke H, Pichlmayr R. Operative Behandlung des Magenkarzinoms. Langenbecks Arch Surg 1990; Suppl. II: 117–23. 17. Inada T, Ogata Y, Andoh J, et al. Significance of paraaortic lymph node dissection in patients with advanced and recurrent gastric cancer. Anticancer Res 1994; 14: 677–82. 18. Kodama Y, Sugimachi K, Soejima K, et al. Evaluation of extensive lymph node dissection for carcinoma of the stomach. World J Surg 1981; 5: 241–8. 19. Soga J, Kobayashi K, Saito J, et al. The role of lymphadenectomy in curative surgery for gastric cancer. World J Surg 1979; 3: 701–8. 20. Soga J, Ohyama S, Miyashita K, et al. Statistical evaluation of advancement in gastric cancer surgery with special reference to the significance of lymphadenectomy for cure. World J Surg 1988; 12: 398–405. 21. Robertson CS, Chung SC, Woods SD, et al. A prospective randomized trial comparing R1 subtotal gastrectomy with R3 total gastrectomy for antral cancer. Ann Surg 1994; 220: 176–82. 22. Gall FP, Hermanek P Die systematische erweiterte Lymphknoten-Dissektion in der kurativen Therapie des Magenkarzinoms. Chirurg 1993; 64: 1024–31. 23. Pacelli F, Doglietto GB, Bellantone R, et al. Extensive versus limited lymph node dissection for gastric cancer: a comparative study of 320 patients. Br J Surg 1993; 80: 1153–6. 24. Viste A, Svanes K, Janssen CW, et al. Prognostic importance of radical lymphadenectomy in curative resections for gastric cancer. Eur J Surg 1994; 160: 497–502.
Gastric cancer 25. Bonenkamp JJ, Van De Velde CJH, Hermans J. Randomised trial of extended lymph node dissection for gastric cancer. In: JR Siewert, JD Roder (eds). Progress in Gastric Cancer Research—Proceedings of the 2nd IGCC. Bologna: Monduzzi Editore S.p.A., 1997: 1111–1116. 26. Volpe CM, Koo J, Miloro SM, et al. The effect of extended lymphadenectomy on survival in patients with gastric adenocarcinoma. J Am Col Surg 1995; 181: 56–64. 27. Baba H, Maehara Y, Takauchi H, et al. Effect of lymph node dissection on the prognosis in patients with node negative early gastric cancer. Surgery 1995; 117: 165–9. 28. Ja¨hne J, Tusch G, Breitmeier D, et al. Will Rogers—Not just a phenomenon: Stage migration as an explanation for the improved prognosis after D2-lymphadenectomy in
505
gastric carcinoma. Langenbecks Arch Surg 1996; Suppl. I: 383–7. 29. Bunt AMG, Hermans J, Smit VTH, et al. Surgical/pathologicalstage migration confounds comparisons of gastric cancer survival between Japan and Western countries. J Clin Oncol 1995; 13: 19–25. 30. Feinstein AR, Sosin DM, Wells CK. The Will Rogers phenomenon—Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med 1985; 312: 1604–8. 31. Natsugoe S, Aiko T, Shimada M, et al. Occult lymph node metastases in gastric cancer with submucosal invasion. Surg Today 1994; 24: 870–5. Accepted for publication 12 November 1999
European Society of Surgical Oncology FELLOWSHIPS The Education and Training Committee of ESSO is pleased to offer a number of Training Fellowships (Minor) to support trainees in surgical oncology. The aim of these followships is to allow young surgeons to spend a time in another specialist centre to either expand their experience or learn new techniques. Applicants require support from their own Head of Department and also from the Head of Department of the Centre which they wish to visit. Fellowships are offered for short periods of training, i.e. 2–3 weeks’ duration and will be supported at the level of 2500 EUR.
In addition, one Major Fellowship to the value of 6.250 EUR will be offered to support an individual who wishes to spend 3 months in a specialist centre to expand his or her clinical or research experience in surgical oncology. The successful applicant will be expected to write a short paper describing their training experiences and this may be published in the European Journal of Surgical Oncology. Deadline for all applications: 22nd September 2000 Applications, including a full CV, letters of support and an outline of what the applicant hopes to gain from the Training Fellowships should be sent to: Mr. James Geraghty Professorial Unit of Surgery Nottingham City Hospital Nottingham, NG5 1PB, UK Fax: +44 115 8402673 Email: [email protected] Further information on our website: www.esso-surgeonline.be