Br. J. Surg. 1992, Vol. 79, October, 1091-1 094
M. Oka, S. Yoshino, S. Hazama, K. Shimoda, M. Suzuki and T. Suzuki Second Department of Surgery, Yamaguchi University School of Medicine, 1 744 Kogushi, Ube, Yarnaguchi 755, Japan Correspondence to: Dr M. Oka
Prognostic significance of regional lymph node reaction after curative resection of advanced gastric cancer The prognosis of patients with advanced gastric cancer who undergo curative resection is still unsatisfactory. The relationship between prognosis and various factors such as stage, lymph node metastasis, serosal invasion and regional lymph node reaction uollicular hyperplasia and sinus histiocytosis) was evaluated. Of the factors studied, the only one that correlated well with survival was sinus histiocytosis. Lymph node metastasis was related weakly to prognosis. No correlation between prognosis and stage, serosal invasion or follicular hyperplasia was observed. Sinus histiocytosis may represent the morphological tumour-host immune reaction. Lymph node metastases or histological types were not related to regional lymph node reaclion. These results suggest that sinus histiocytosis could be a useful prognostic factor for gastric cancer. Intensive postoperative follow-up for recurrence may be necessary even for patients with low-grade sinus histiocytosis in whom curative surgery is attempted.
The survival rate of patients with gastric cancer has improved owing to the early detection ofcancer and to progress in surgery, chemotherapy and immunotherapy . The 5-year survival rate for early gastric cancer'-3 is > 90 per cent. However, in patients with advanced gastric cancer who undergo curative resection the 5-year survival rateL6 varies from 28 to 61 per cent. There are a number of prognostic factors for gastric cancer. Maruyama et 01.' reported that factors such as peritoneal, liver and distant lymph node metastases, and Borrmann type IV cancer (diffusely infiltrating carcinoma ) could reduce the success of surgical treatment. Kodama et a/.' reported that lymph node metastases and serosal invasion significantly affect the 5-year survival rate. It is known that the tumour-bearing host mounts an immune response against its own neoplasm, chiefly by cell-mediated immunity'. Experimentally, regional lymph nodes are capable of interfering with the growth of tumour cells8. Regional lymph nodes may block the implantation and growth of distant metastases'. These results suggest that regional lymph nodes may play an important role in the development of cancer. Initially Black et a/.'' reported that sinus histiocytosis in regional lymph nodes in breast cancer correlated well with the survival rate. They also observed that follicular hyperplasia in the regional lymph nodes was associated with longer survival' I . In a previous study it has been shown that postoperative survival of patients with oesophageal cancer correlated with sinus histiocytosis responses in regional lymph nodes". The aim of this study was t o evaluate the relationship between prognosis and factors such as cancer stage, lymph node metastases, serosal invasion and regional lymph node reaction in patients with gastric cancer undergoing curative surgery.
Patients and methods From 1975 to 1986, 102 patients with stage I1 and 111 gastric cancer underwent curative gastrectomy in this department (Table I ). All had primary adenocarcinoma of the stomach and no synchronous or metachronous carcinoma. No patient received any anticancer therapy before surgery. After surgery 41 patients received chemotherapy with the 5-Ruorouracil analogue tegafur (ftorafur FT: Taiho, Tokyo, Japan), 46 received immunochemotherapy with OK-432 and tegafur, and 15 had no additional therapy. The patients' ages ranged from 32 to 81
0007-1323/92/10109144
0 1992 Butterworth-Heinemann Ltd
years. Operative findings, microscopic findings and curability were assessed and described according to the criteria of the Japanese Research Society for Gastric CancerI3. Briefly, stage I1 includes microscopic nodal metastasis surrounding the stomach ( n l ( + ) ) and/or microscopic wall invasion in the subserosal layer with no macroscopic liver metastasis and no peritoneal dissemination: stage 111 includes microscopic nodal metastasis along thz main arteries including the coeliac axis (n2( + ))and/ormicroscopic wall invasion in the serosal layer with no macroscopic liver metastases or peritoneal dissemination. Tumours with invasion of the gastric wall no deeper than the subserosal layer were included in the category of cancers with negative serosal invasion. Any cancer with invasion of a serosal surface exposed to the peritoneal cavity was classified as a cancer with positive serosal invasion. The criteria for curative resection included: ( 1 ) no macroscopic liver metastases or peritoneal dissemination and no other distant metastases: ( 2 )no residual cancer cells at the resection margins; and ( 3 ) no lymph node metastasis or resectable positive nodes. Histological type was taken as that with the most extensive histological pattern in the specimen. Fifty-one patients are still alive and 51 have died from tumour recurrence. The evaluation of microscopic lymph node reaction was performed in all 102 patients. A total of 3267 lymph nodes were examined. Sinus
Table 1 Characteristics of 102 patients wiih udvanced gastric cancer No. of patients Sex ratio ( M : F ) Postoperative therapy None Chemotherapy Immunochemotherapy Stage
I1 111
Serosal invasion Negative Positive Lymph node metastasis Negative Positive nl n2
63 :39 15 41 46 48 54 47 55 17 52 33
1091
Prognosis following resection of advanced gastric cancer: M. Oka et al. Table 2 Grade ?/regional lvnrplr node reuctiori Grade
Description
Follicular hyperplasia 0 Germinal centres I Germinal centres Germinal centres 2 Germinal centres 3 Sinus histiocytosis Sinus not 0 Sinus 1 Sinus 4-4 2 Sinus > 4 3
rare recognizable in cortex recognizable in cortex and medulla recognizable in all lymph node areas
seen in lymph node area
46 per cent, while that of 25 with follicular hyperplasia and There was lymph node metastases was 51 per cent (Figure 41). no significant difference between the two groups. The cumulative survival rate at 5 years in 33 patients with lymph node metastases and no sinus histiocytosis was 28 per cent, while that of 52 with lymph node metastases and sinus histiocytosis was 62 per cent (Figure 4b) ( P < 0.05). Multivariate analysis was performed to determine the most useful prognostic factor. Only sinus histiocytosis was significantly related to survival (Table 3).
5
0
10
I
5
0
a
Time a f t e r resection ( y e a r s )
100
10
Time a f t e r r e s e c t i o n ( y e a r s ) 100
-s W
Y
m
50
-
50
.-?
> J
m
I 5
I
0
'11 . 1
I
10
0
b
Time a f t e r r e s e c t i o n ( y e a r s )
Figure 3 Rrlurionsliip her wren s u r r i i d rute and regional lymph node rcwtion. a Folliculur liypc~rplusiu: e, negutioe cu.sr.s ( n = 76 ); 0 , positiw c'o.ses ( n = 2 6 ) . b Sinus lii.stioc~rtosi.s:a, nqutii:e cuses (n = 3 7 ) ; 0.
I
p o s i t i i ~cuses ~ (n = 65)
Variable*
z2
P
Stage ( I 1 rersus I11 ) Serosal invasion Lymph node metastasis Follicular hyperplasia Sinus histiocytosis
0.0 I3 1.551 0.185 0.105 7.0 19
0.836 0. I24 0.538 0.147 0.009
*The proportional hazards model was used to test the significance of the presence or absence of each variable (except stage). I d.f.
Lymph node reaction Follicular hyperplasia
Sinus histiocytosis
Lymph node metastases
Negative
Positive
Negative
Positive
Negative ( 1 1 = 17) Positive ( n = 85)
16 60
I 25
4 33
13 52
Lymph node reaction was compared between 17 patients with negative nodes and 85 with positive nodes. There were no correlations between lymph node reaction and lymph node metastases (follicular hyperplasia: xz = 2.98; sinus histiocytosis: x2 = 0.85; 4 d.f.) (Table 4 ) . Histologically, the tumours showed tubular adenocarcinoma in 41 patients, poorly differentiated adenocarcinoma in 36,
Br. J. Surg., Vol. 79. No. 10. October 1992
'1 J
I
5
I
10
Time a f t e r r e s e c t i o n (Years)
Figure 4
Relutionship hrtwrrn surriral rutc and rrgional lyniph node rruction iri patients with nodual n~rtusta.sis.a Patirnts without (e,n = 61)) = .?5),folliculur Iijprrplusiu. b PLitirnts without (a, n = 33) w i ~ h( 0 , n = 51) s i n i i . ~his~iocyrosis
ond ~ v i t h( 0 ,n tintl
Table 3 Multirarirrtr unulysis of' curious prognostic jirctors reluted to lrngth uf'sirrrircrl in 102 pcttieiits ivith yctstric cuiiccr
I
signet-ring cell carcinoma in 17, mucinous adenocarcinoma in five and papillary adenocarcinoma in three. There was no relationship between lymph node reaction and histological type (follicular hyperplasia: x2 = 6.14; sinus histiocytosis: xz = 050; 4 d.f. ) ( Table 5 ).
Discussion The analysis of postoperative prognostic factors for gastric cancer may be useful for the selection of postoperative therapy and follow-up study. If a patient is at high risk for tumour recurrence, aggressive chemotherapy and intensive follow-up may be required after operation. Lymph node metastases and depth of invasion can be considered the major prognostic factors after attempted curative resection of advanced gastric cancer. Baba et d.'' carried out univariate and multivariate analyses in patients who underwent curative resection for advanced carcinoma of the stomach and identified tumour stage, lymph node metastases and depth of invasion as important prognostic factors. Boku rt a/.' also reported that the higher the aggregate total of serosal invasion and lymph node metastasis factors, the lower the 5-year survival rate. To analyse the prognostic factors for gastric cancer in this study, patients with stage I1 and I11 cancer were selected because half of these patients die from tumour recurrence. There was no significant difference in the survival rate between the two groups. Thus subsequent analyses were carried out on all patients combined. The prognosis of patients without nodal metastases may be slightly better than those with such metastases; however, a significant difference was not found. Moreover, the prognosis of patients with serosal invasion was similar to that in those without. Therefore neither lymph node metastases nor serosal invasion could be identified
1093
Prognosis following resection of advanced gastric cancer: M. Oka et al. Table 5 Relationship betueen lymph node reaction and hisrological tvpe in 102 patients with gastric cancer
Lymph node reaction Follicular hyperplasia Histological type of adenocarcinoma Papillary Tubular Poorly differentiated Mucinous Signet-ring cell Total
No. of patients
Negative
Positive
3 41 36 5 17
3 27 26 5 15
0 14
102
76
as significant prognostic factors in patients with stage I1 and I11 cancer. This failure to confirm well known prognostic factors may be due to small sample size. Black et al." presented an initial report that high-grade sinus histiocytosis in the axillary nodes of patients with breast cancer is almost invariably associated with a survival of 5 years or more regardless of the presence of axillary metastases. Many reports have suggested that sinus histiocytosis is correlated with a better prognosis in various cancers, such as those of the breast, . The evaluation of follicular colon and oesophagusi2~is-18 hyperplasia is different. Many authors have indicated that there is no correlation between follicular hyperplasia and prognosis12.i6-18, but some reports d o support such a relationship".". Black et a/.' I reported that in gastric cancer follicular hyperplasia was associated with longer survival and that sinus histiocytosis was seen infrequently. The present study indicates that the prognosis of patients with sinus histiocytosis is better than that of those without, while follicular hyperplasia has no effect on prognosis. Multivariate analysis demonstrated that sinus histiocytosis is the only independent variable associated with survival. In patients with nodal metastases, sinus histiocytosis was associated with longer survival. This lymph node reaction is considered to represent a tumour-host reactioni6. Experimental results have suggested that regional lymph nodes are capable of interfering with the growth of tumour cells8. Analysis of T cell subsets in the lymph node using monoclonal antibodies indicates that T cells predominate in the paracortical area, whereas B cells comprise the predominant cell type in follicular areas". Thus, sinus histiocytosis in the paracortical area may represent the activation of both T cells and assumed that sinus histiocytosis macrophages. Patt el represented morphological evidence of cell-mediated immunity, while follicular hyperplasia may be an aspect of the antibody-mediated immune response. Since T cells and macrophages (histiocytes) play a major role in cancer immunity, sinus histiocytosis in regional lymph nodes may be a local antitumour reaction. Lymph node metastases or histological type may influence lymph node reaction. The relationship between lymph node reaction and lymph node metastases or histological type was examined in the present study and no correlation was found. Thus, these factors do not appear to be related to regional lymph node reaction. Finally, in the present study sinus histiocytosis was found to be a more sensitive prognostic factor than either lymph node metastases or serosal invasion in patients with advanced gastric cancer. Patients with low-grade sinus histiocytosis may require more intensive follow-up for recurrence after surgery.
1.
Kodama Y, Sugimachi K, Soejima K e f nl. Evaluation of extensive lymph node dissection for carcinoma of the stomach. World J Surg 1981; 5 : 241-8.
1094
Negative 1
2. 3.
Positive
0 2
15 13 2 6
2 26 23 3 11
26
37
65
10
Mishima Y, Hirayama R. The role of lymph node surgery in gastric cancer. World J Surg 1987; 11: 406-11. Soga J, Ohyama S, Miyashita K e f a!. A statistical evaluation of advancement in gastric cancer surgcry with special reference to the significance of lymphadenectomy for cure. World J Surg 1988; 12: 398-405.
4.
Breaux JR, Bringaze W, Chappuis C, Cohn I. Adenocarcinoma of the stomach: a review of 35 years and 1710 cases. World J Surg 1990; 14: 580-6.
5. 6. 7.
Maruyama K, Okabayashi K, Kinoshita T. Progress in gastric surgery in Japan and its limits of radicality. World J Surg 1987; 11: 418-25. Boku T, Nakane Y, Minoura T et al. Prognostic significance of
serosal invasion and free intraperitoneal cancer cells in gastric cancer. Br J Surg 1990; 11:436-9. Friedman H, Southam CM. Area of relationship between immunology and clinical oncology. Am J C h Pathol 1976; 62: 224-42.
8.
Fisher B, Saffer E, Fisher ER. Studies concerning the regional lymph in cancer IV. Tumor inhibition by regional lymph node cells. Cancer 1974; 33: 631-6.
9.
Crile G. The smaller the cancer, the bigger the operation? JAMA
10.
Black MM, Kerpe S, Speer FD. Lymph node structure in patients with cancer of the breast. Am J Pathoi 1953; 29: 505-21. Black MM, Freeman C, Mork T, Harvei S, Cutler SJ. Prognostic significanceof microscopic structure of gastric carcinomas and their regional lymph nodes. Cancer 1971; 27: 703-11. Oka M. Immunological studies on esophageal cancer - cellular immunocompetence and histological responses in main tumor and regional lymph nodes in esophageal cancer patients. Arch
1967; 199: 736-8. 11.
12.
Jpn Chir 1981; 50: 29-44.
13.
Japanese Research Society for Gastric Cancer. The general rules for gastric cancer study in surgery and pathology.-Jpn J Surg
14.
Baba H. Korenaga D, Okamura T, Saito A, Sugimachi K. Prognostic factors in gastric cancer with serosal invasion. Arch
1981; 11: 127-45. Sury 1989; 124: 1061-4.
15. 16.
17.
Black MM, Barclay THC, Hankey BF. Prognosis in breast cancer utilizing histologic characteristics of the primary tumor. Cuncer 1975; 36: 2048-55. Patt DJ, Brynes RK, Vardiman JW, Coppelson LW. Mesocolic
lymph node histology is an important prognostic indicator for patients with carcinoma of the sigmoid colon: an immunomorphologic study. Cuncer 1975; 35: 1388-97. Syranen KJ, Hjelt LH. Tumor-host interrelationship in carcinoma of the female breast. Sury G ~ ~ n e cOhsrer ~ o l 1978; 147: 43-8.
18.
19.
20.
References
Sinus histiocytosis
Urdiales-Viedma M, Nogales-Fernandez F, Martos-Padilla S, Sanchez-Cantalejo E. Correlation of histologic grade and lymph node status with some histopathologic discriminants in breast cancer. Tuntori 1986; 72: 43-51. Nacopoulou L, Azaris P, Napacharalampous N, Davaris P. Prognostic significanceof histologic host response in cancer of the large bowel. Cuncer 1981; 47: 930-6. Poppema S, Bhan AK, Reinherz EL, McCluskey RT, Schlossman SF. Distribution of cell subsets in human lymph nodes. J ESP Med 1981; 153: 30-41.
Paper accepted 14 March 1992
Br. J. Surg., Vol. 79, No. 10, October 1992
M. Oka, S. Yoshino, S. Hazama, K. Shimoda, M. Suzuki and T. Suzuki Second Department of Surgery, Yamaguchi University School of Medicine, 1 744 Kogushi, Ube, Yarnaguchi 755, Japan Correspondence to: Dr M. Oka
Prognostic significance of regional lymph node reaction after curative resection of advanced gastric cancer The prognosis of patients with advanced gastric cancer who undergo curative resection is still unsatisfactory. The relationship between prognosis and various factors such as stage, lymph node metastasis, serosal invasion and regional lymph node reaction uollicular hyperplasia and sinus histiocytosis) was evaluated. Of the factors studied, the only one that correlated well with survival was sinus histiocytosis. Lymph node metastasis was related weakly to prognosis. No correlation between prognosis and stage, serosal invasion or follicular hyperplasia was observed. Sinus histiocytosis may represent the morphological tumour-host immune reaction. Lymph node metastases or histological types were not related to regional lymph node reaclion. These results suggest that sinus histiocytosis could be a useful prognostic factor for gastric cancer. Intensive postoperative follow-up for recurrence may be necessary even for patients with low-grade sinus histiocytosis in whom curative surgery is attempted.
The survival rate of patients with gastric cancer has improved owing to the early detection ofcancer and to progress in surgery, chemotherapy and immunotherapy . The 5-year survival rate for early gastric cancer'-3 is > 90 per cent. However, in patients with advanced gastric cancer who undergo curative resection the 5-year survival rateL6 varies from 28 to 61 per cent. There are a number of prognostic factors for gastric cancer. Maruyama et 01.' reported that factors such as peritoneal, liver and distant lymph node metastases, and Borrmann type IV cancer (diffusely infiltrating carcinoma ) could reduce the success of surgical treatment. Kodama et a/.' reported that lymph node metastases and serosal invasion significantly affect the 5-year survival rate. It is known that the tumour-bearing host mounts an immune response against its own neoplasm, chiefly by cell-mediated immunity'. Experimentally, regional lymph nodes are capable of interfering with the growth of tumour cells8. Regional lymph nodes may block the implantation and growth of distant metastases'. These results suggest that regional lymph nodes may play an important role in the development of cancer. Initially Black et a/.'' reported that sinus histiocytosis in regional lymph nodes in breast cancer correlated well with the survival rate. They also observed that follicular hyperplasia in the regional lymph nodes was associated with longer survival' I . In a previous study it has been shown that postoperative survival of patients with oesophageal cancer correlated with sinus histiocytosis responses in regional lymph nodes". The aim of this study was t o evaluate the relationship between prognosis and factors such as cancer stage, lymph node metastases, serosal invasion and regional lymph node reaction in patients with gastric cancer undergoing curative surgery.
Patients and methods From 1975 to 1986, 102 patients with stage I1 and 111 gastric cancer underwent curative gastrectomy in this department (Table I ). All had primary adenocarcinoma of the stomach and no synchronous or metachronous carcinoma. No patient received any anticancer therapy before surgery. After surgery 41 patients received chemotherapy with the 5-Ruorouracil analogue tegafur (ftorafur FT: Taiho, Tokyo, Japan), 46 received immunochemotherapy with OK-432 and tegafur, and 15 had no additional therapy. The patients' ages ranged from 32 to 81
0007-1323/92/10109144
0 1992 Butterworth-Heinemann Ltd
years. Operative findings, microscopic findings and curability were assessed and described according to the criteria of the Japanese Research Society for Gastric CancerI3. Briefly, stage I1 includes microscopic nodal metastasis surrounding the stomach ( n l ( + ) ) and/or microscopic wall invasion in the subserosal layer with no macroscopic liver metastasis and no peritoneal dissemination: stage 111 includes microscopic nodal metastasis along thz main arteries including the coeliac axis (n2( + ))and/ormicroscopic wall invasion in the serosal layer with no macroscopic liver metastases or peritoneal dissemination. Tumours with invasion of the gastric wall no deeper than the subserosal layer were included in the category of cancers with negative serosal invasion. Any cancer with invasion of a serosal surface exposed to the peritoneal cavity was classified as a cancer with positive serosal invasion. The criteria for curative resection included: ( 1 ) no macroscopic liver metastases or peritoneal dissemination and no other distant metastases: ( 2 )no residual cancer cells at the resection margins; and ( 3 ) no lymph node metastasis or resectable positive nodes. Histological type was taken as that with the most extensive histological pattern in the specimen. Fifty-one patients are still alive and 51 have died from tumour recurrence. The evaluation of microscopic lymph node reaction was performed in all 102 patients. A total of 3267 lymph nodes were examined. Sinus
Table 1 Characteristics of 102 patients wiih udvanced gastric cancer No. of patients Sex ratio ( M : F ) Postoperative therapy None Chemotherapy Immunochemotherapy Stage
I1 111
Serosal invasion Negative Positive Lymph node metastasis Negative Positive nl n2
63 :39 15 41 46 48 54 47 55 17 52 33
1091
Prognosis following resection of advanced gastric cancer: M. Oka et al. Table 2 Grade ?/regional lvnrplr node reuctiori Grade
Description
Follicular hyperplasia 0 Germinal centres I Germinal centres Germinal centres 2 Germinal centres 3 Sinus histiocytosis Sinus not 0 Sinus 1 Sinus 4-4 2 Sinus > 4 3
rare recognizable in cortex recognizable in cortex and medulla recognizable in all lymph node areas
seen in lymph node area
46 per cent, while that of 25 with follicular hyperplasia and There was lymph node metastases was 51 per cent (Figure 41). no significant difference between the two groups. The cumulative survival rate at 5 years in 33 patients with lymph node metastases and no sinus histiocytosis was 28 per cent, while that of 52 with lymph node metastases and sinus histiocytosis was 62 per cent (Figure 4b) ( P < 0.05). Multivariate analysis was performed to determine the most useful prognostic factor. Only sinus histiocytosis was significantly related to survival (Table 3).
5
0
10
I
5
0
a
Time a f t e r resection ( y e a r s )
100
10
Time a f t e r r e s e c t i o n ( y e a r s ) 100
-s W
Y
m
50
-
50
.-?
> J
m
I 5
I
0
'11 . 1
I
10
0
b
Time a f t e r r e s e c t i o n ( y e a r s )
Figure 3 Rrlurionsliip her wren s u r r i i d rute and regional lymph node rcwtion. a Folliculur liypc~rplusiu: e, negutioe cu.sr.s ( n = 76 ); 0 , positiw c'o.ses ( n = 2 6 ) . b Sinus lii.stioc~rtosi.s:a, nqutii:e cuses (n = 3 7 ) ; 0.
I
p o s i t i i ~cuses ~ (n = 65)
Variable*
z2
P
Stage ( I 1 rersus I11 ) Serosal invasion Lymph node metastasis Follicular hyperplasia Sinus histiocytosis
0.0 I3 1.551 0.185 0.105 7.0 19
0.836 0. I24 0.538 0.147 0.009
*The proportional hazards model was used to test the significance of the presence or absence of each variable (except stage). I d.f.
Lymph node reaction Follicular hyperplasia
Sinus histiocytosis
Lymph node metastases
Negative
Positive
Negative
Positive
Negative ( 1 1 = 17) Positive ( n = 85)
16 60
I 25
4 33
13 52
Lymph node reaction was compared between 17 patients with negative nodes and 85 with positive nodes. There were no correlations between lymph node reaction and lymph node metastases (follicular hyperplasia: xz = 2.98; sinus histiocytosis: x2 = 0.85; 4 d.f.) (Table 4 ) . Histologically, the tumours showed tubular adenocarcinoma in 41 patients, poorly differentiated adenocarcinoma in 36,
Br. J. Surg., Vol. 79. No. 10. October 1992
'1 J
I
5
I
10
Time a f t e r r e s e c t i o n (Years)
Figure 4
Relutionship hrtwrrn surriral rutc and rrgional lyniph node rruction iri patients with nodual n~rtusta.sis.a Patirnts without (e,n = 61)) = .?5),folliculur Iijprrplusiu. b PLitirnts without (a, n = 33) w i ~ h( 0 , n = 51) s i n i i . ~his~iocyrosis
ond ~ v i t h( 0 ,n tintl
Table 3 Multirarirrtr unulysis of' curious prognostic jirctors reluted to lrngth uf'sirrrircrl in 102 pcttieiits ivith yctstric cuiiccr
I
signet-ring cell carcinoma in 17, mucinous adenocarcinoma in five and papillary adenocarcinoma in three. There was no relationship between lymph node reaction and histological type (follicular hyperplasia: x2 = 6.14; sinus histiocytosis: xz = 050; 4 d.f. ) ( Table 5 ).
Discussion The analysis of postoperative prognostic factors for gastric cancer may be useful for the selection of postoperative therapy and follow-up study. If a patient is at high risk for tumour recurrence, aggressive chemotherapy and intensive follow-up may be required after operation. Lymph node metastases and depth of invasion can be considered the major prognostic factors after attempted curative resection of advanced gastric cancer. Baba et d.'' carried out univariate and multivariate analyses in patients who underwent curative resection for advanced carcinoma of the stomach and identified tumour stage, lymph node metastases and depth of invasion as important prognostic factors. Boku rt a/.' also reported that the higher the aggregate total of serosal invasion and lymph node metastasis factors, the lower the 5-year survival rate. To analyse the prognostic factors for gastric cancer in this study, patients with stage I1 and I11 cancer were selected because half of these patients die from tumour recurrence. There was no significant difference in the survival rate between the two groups. Thus subsequent analyses were carried out on all patients combined. The prognosis of patients without nodal metastases may be slightly better than those with such metastases; however, a significant difference was not found. Moreover, the prognosis of patients with serosal invasion was similar to that in those without. Therefore neither lymph node metastases nor serosal invasion could be identified
1093
Prognosis following resection of advanced gastric cancer: M. Oka et al. Table 5 Relationship betueen lymph node reaction and hisrological tvpe in 102 patients with gastric cancer
Lymph node reaction Follicular hyperplasia Histological type of adenocarcinoma Papillary Tubular Poorly differentiated Mucinous Signet-ring cell Total
No. of patients
Negative
Positive
3 41 36 5 17
3 27 26 5 15
0 14
102
76
as significant prognostic factors in patients with stage I1 and I11 cancer. This failure to confirm well known prognostic factors may be due to small sample size. Black et al." presented an initial report that high-grade sinus histiocytosis in the axillary nodes of patients with breast cancer is almost invariably associated with a survival of 5 years or more regardless of the presence of axillary metastases. Many reports have suggested that sinus histiocytosis is correlated with a better prognosis in various cancers, such as those of the breast, . The evaluation of follicular colon and oesophagusi2~is-18 hyperplasia is different. Many authors have indicated that there is no correlation between follicular hyperplasia and prognosis12.i6-18, but some reports d o support such a relationship".". Black et a/.' I reported that in gastric cancer follicular hyperplasia was associated with longer survival and that sinus histiocytosis was seen infrequently. The present study indicates that the prognosis of patients with sinus histiocytosis is better than that of those without, while follicular hyperplasia has no effect on prognosis. Multivariate analysis demonstrated that sinus histiocytosis is the only independent variable associated with survival. In patients with nodal metastases, sinus histiocytosis was associated with longer survival. This lymph node reaction is considered to represent a tumour-host reactioni6. Experimental results have suggested that regional lymph nodes are capable of interfering with the growth of tumour cells8. Analysis of T cell subsets in the lymph node using monoclonal antibodies indicates that T cells predominate in the paracortical area, whereas B cells comprise the predominant cell type in follicular areas". Thus, sinus histiocytosis in the paracortical area may represent the activation of both T cells and assumed that sinus histiocytosis macrophages. Patt el represented morphological evidence of cell-mediated immunity, while follicular hyperplasia may be an aspect of the antibody-mediated immune response. Since T cells and macrophages (histiocytes) play a major role in cancer immunity, sinus histiocytosis in regional lymph nodes may be a local antitumour reaction. Lymph node metastases or histological type may influence lymph node reaction. The relationship between lymph node reaction and lymph node metastases or histological type was examined in the present study and no correlation was found. Thus, these factors do not appear to be related to regional lymph node reaction. Finally, in the present study sinus histiocytosis was found to be a more sensitive prognostic factor than either lymph node metastases or serosal invasion in patients with advanced gastric cancer. Patients with low-grade sinus histiocytosis may require more intensive follow-up for recurrence after surgery.
1.
Kodama Y, Sugimachi K, Soejima K e f nl. Evaluation of extensive lymph node dissection for carcinoma of the stomach. World J Surg 1981; 5 : 241-8.
1094
Negative 1
2. 3.
Positive
0 2
15 13 2 6
2 26 23 3 11
26
37
65
10
Mishima Y, Hirayama R. The role of lymph node surgery in gastric cancer. World J Surg 1987; 11: 406-11. Soga J, Ohyama S, Miyashita K e f a!. A statistical evaluation of advancement in gastric cancer surgcry with special reference to the significance of lymphadenectomy for cure. World J Surg 1988; 12: 398-405.
4.
Breaux JR, Bringaze W, Chappuis C, Cohn I. Adenocarcinoma of the stomach: a review of 35 years and 1710 cases. World J Surg 1990; 14: 580-6.
5. 6. 7.
Maruyama K, Okabayashi K, Kinoshita T. Progress in gastric surgery in Japan and its limits of radicality. World J Surg 1987; 11: 418-25. Boku T, Nakane Y, Minoura T et al. Prognostic significance of
serosal invasion and free intraperitoneal cancer cells in gastric cancer. Br J Surg 1990; 11:436-9. Friedman H, Southam CM. Area of relationship between immunology and clinical oncology. Am J C h Pathol 1976; 62: 224-42.
8.
Fisher B, Saffer E, Fisher ER. Studies concerning the regional lymph in cancer IV. Tumor inhibition by regional lymph node cells. Cancer 1974; 33: 631-6.
9.
Crile G. The smaller the cancer, the bigger the operation? JAMA
10.
Black MM, Kerpe S, Speer FD. Lymph node structure in patients with cancer of the breast. Am J Pathoi 1953; 29: 505-21. Black MM, Freeman C, Mork T, Harvei S, Cutler SJ. Prognostic significanceof microscopic structure of gastric carcinomas and their regional lymph nodes. Cancer 1971; 27: 703-11. Oka M. Immunological studies on esophageal cancer - cellular immunocompetence and histological responses in main tumor and regional lymph nodes in esophageal cancer patients. Arch
1967; 199: 736-8. 11.
12.
Jpn Chir 1981; 50: 29-44.
13.
Japanese Research Society for Gastric Cancer. The general rules for gastric cancer study in surgery and pathology.-Jpn J Surg
14.
Baba H. Korenaga D, Okamura T, Saito A, Sugimachi K. Prognostic factors in gastric cancer with serosal invasion. Arch
1981; 11: 127-45. Sury 1989; 124: 1061-4.
15. 16.
17.
Black MM, Barclay THC, Hankey BF. Prognosis in breast cancer utilizing histologic characteristics of the primary tumor. Cuncer 1975; 36: 2048-55. Patt DJ, Brynes RK, Vardiman JW, Coppelson LW. Mesocolic
lymph node histology is an important prognostic indicator for patients with carcinoma of the sigmoid colon: an immunomorphologic study. Cuncer 1975; 35: 1388-97. Syranen KJ, Hjelt LH. Tumor-host interrelationship in carcinoma of the female breast. Sury G ~ ~ n e cOhsrer ~ o l 1978; 147: 43-8.
18.
19.
20.
References
Sinus histiocytosis
Urdiales-Viedma M, Nogales-Fernandez F, Martos-Padilla S, Sanchez-Cantalejo E. Correlation of histologic grade and lymph node status with some histopathologic discriminants in breast cancer. Tuntori 1986; 72: 43-51. Nacopoulou L, Azaris P, Napacharalampous N, Davaris P. Prognostic significanceof histologic host response in cancer of the large bowel. Cuncer 1981; 47: 930-6. Poppema S, Bhan AK, Reinherz EL, McCluskey RT, Schlossman SF. Distribution of cell subsets in human lymph nodes. J ESP Med 1981; 153: 30-41.
Paper accepted 14 March 1992
Br. J. Surg., Vol. 79, No. 10, October 1992
