Immunohistochemically detected ... - Wiley Online Library

753

Immunohistochemically Detected Micrometastases of the Lymph Nodes in Patients with Gastric Carcinoma Takeo Fukagawa, M.D.1 Mitsuru Sasako, M.D.,2 Gregory Bruce Mann, M.D.3 Takeshi Sano, M.D.2 Hitoshi Katai, M.D.2 Keiichi Maruyama, M.D.2 Yukihiro Nakanishi, M.D.4 Tadakazu Shimoda, M.D.5 1

Department of Surgical Oncology, Faculty of Medicine, University of Tokyo, Tokyo, Japan.

2

Department of Surgical Oncology, National Cancer Center Hospital, Tokyo, Japan.

3

Department of Surgery, The Royal Melbourne Hospital, University of Melbourne, Melbourne, Australia.

4

Pathology Division, National Cancer Center Research Institute, Tokyo, Japan.

5

Clinical Laboratory Division, National Cancer Center Hospital, Tokyo, Japan.

BACKGROUND. Lymph node status is a major determinant of disease recurrence after patients undergo curative resection for gastric carcinoma. A proportion of patients without lymph node metastasis develop systemic recurrences. Recent studies in a range of solid tumors have found a high incidence of micrometastases in the regional lymph nodes of patients with apparently negative lymph nodes. In patients with breast and colorectal carcinoma, the presence of micrometastases has been associated with a poorer prognosis. In patients with gastric carcinoma, the significance of micrometastases in lymph nodes remains controversial. Most published reports on this subject suffer from the problems of small sample size and selection bias. METHODS. One hundred seven patients with pathologic T2N0M0 (tumor invades muscularis propria or subserosa [T2], no regional lymph node metastasis [N0], and no distant metastasis [M0]; pT2N0M0) gastric carcinoma who underwent gastric resection between 1984 and 1990 at the National Cancer Center Hospital were studied. Two consecutive sections were newly prepared from each lymph node for hematoxylin and eosin staining and immunohistochemical staining (IHC) with antibody against cytokeratin. Associations between clinicopathologic factors and the presence of micrometastases as well as micrometastases and survival were sought. RESULTS. Micrometastases were identified in 38 of 107 patients (35.5%) and in 87 of 4484 lymph nodes (1.94%) by IHC. The incidence of micrometastases was significantly higher in patients with infiltrative tumors than in patients with expansive, growing tumors (P ⫽ 0.02). Other clinicopathologic findings had no statistically significant correlation with the incidence of micrometastases. The 5-year survival rates of patients with and without micrometastases were 94% and 89%, respectively. Similarly, the 10-year survival rates were 79% and 74%, respectively. The survival curves of patients with or without micrometastasis were nearly superimposed (P ⫽ 0.86). CONCLUSIONS. The presence of immunohistochemically detected micrometastases in the regional lymph nodes did not affect the survival of Japanese patients with pT2N0M0 gastric carcinoma who had undergone gastrectomy with D2 lymph node dissection. Cancer 2001;92:753– 60. © 2001 American Cancer Society. KEYWORDS: micrometastases, gastric carcinoma, immunohistochemistry, surgical treatment.

Address for reprints: Mitsuru Sasako, M.D., Ph.D., Department of Surgical Oncology, National Cancer Center Hospital, 5-1-1 Tsukiji, Chuo-ku, Tokyo, Japan; Fax; 81-3-3542-3815; E-mail; msasako@ gan2.ncc.go.jp. Received December 4, 2000; revision received April 16, 2001; accepted May 2, 2001. © 2001 American Cancer Society

T

he major prognostic factors for patients with gastric carcinoma are the depth of the primary tumor and the presence or absence of lymph node, peritoneal, and distant metastases.1,2 The complete removal of tumors is the only potentially curative treatment for patients with gastric carcinoma, although the value of extended lymph node dissection remains controversial, and some recent reports may revive interest in adjuvant therapy.3 Locally advanced gastric carcinoma frequently recurs after patients undergo curative operation (Interna-

754

CANCER August 15, 2001 / Volume 92 / Number 4

tional Union Against Cancer R0 resection), and even early gastric tumors occasionally recur.4 In such patients with recurrent disease, there must have been residual tumor cells in the form of occult micrometastases that were left behind at the time of apparently curative surgery. Recent advances in immunohistochemistry (IHC) and molecular biology allow the identification of discrete and occult tumor cells, which remain undetected by standard hematoxylin and eosin (H&E) staining, in the lymph nodes of patients with breast,5,6 esophageal,7,8 and colorectal9 –15 carcinoma. Various methods have been developed and are being assessed for the detection of such tumor cells in bone marrow,16 –19 lymph nodes,20 –23 or peritoneal lavage fluid.24 Although the natural history of micrometastases is not well known,25 the existence of tumor cells in such sites is correlated with patient prognosis in various malignancies. Cytokeratin is a component of the cytoskeleton of epithelial cells26 that is not present in normal lymph nodes; therefore, IHC using an antibody that recognizes cytokeratin27 enables the detection of minute deposits of tumor cells in lymph nodes. Numerous reports have assessed the correlation between the presence of such micrometastases and patient prognosis in gastric carcinoma.28 –33 Conclusive results are not available. The prognosis of patients with gastric carcinoma is affected strongly by the T-stage of the tumor. Once the primary tumor invades through the serosa (T3), peritoneal dissemination and subsequent peritoneal recurrence is common. Patients with pathologic T2N0M0 (tumor invades muscularis propria or subserosa [T2], no regional lymph node metastasis [N0], no distant metastasis [M0]; pT2N0M0) gastric carcinoma were selected for this study to determine the clinical significance of regional lymph nodal micrometastases and to avoid the confounding effect of peritoneal recurrence.

MATERIALS AND METHODS One thousand seven hundred fifty-one patients underwent gastric resection for gastric carcinoma from January 1984 through December 1990 at the National Cancer Center Hospital, Tokyo. During this period, 118 patients were classified with pT2N0M0 disease on the prospective data base maintained within the Gastric Surgery Department. Five patients whose paraffin blocks were not available, two patients whose T stage was revised at this review (one to T3 and one to T1), two patients who were found to have lymph node involvement by H&E staining at this review, and another two patients who were completely lost to follow-

up were excluded from the analysis. The remaining 107 consecutive patients were studied. When followup details were not recorded in the data base, the data were obtained from the patient’s local medical officer. All patients underwent partial or total gastrectomy with systematic lymphadenectomy involving the complete dissection of perigastric lymph nodes with complete or partial removal of the second-tier lymph nodes. Some patients underwent para-aortic lymph node dissection. All specimens were formalin fixed and paraffin embedded. Lymph nodes were examined by one cross section through the center of each lymph node. Tumors were classified histologically into differentiated and undifferentiated types according to the World Health Organization tumor classification system.34 The differentiated type includes well or moderately differentiated tubular adenocarcinoma and papillary adenocarcinoma of Japanese classification, whereas the undifferentiated type includes poorly differentiated adenocarcinoma, signet ring cell carcinoma, and mucinous adenocarcinoma. Patients with tumors that were classified as T2N0M0 were selected for this study. All available paraffin blocks were obtained, and two consecutive sections measuring 4 ␮m in thickness were newly cut for H&E staining and immunostaining. Immunohistochemistry was performed using AE1/AE3 (Boehringer Mannheim, Indianapolis, IN), a monoclonal antibody that is reactive with a broad spectrum of human cytokeratins. The tissue sections were deparaffinized and dehydrated, then microwaved for 10 minutes for antigen retrieval, and incubated with AE1/AE3 at a 1:50 dilution, followed by second antibodies against mouse immunoglobulin (Vector Laboratories, Burlingame, CA). Immunostaining was performed using the avidin-biotin-peroxidase complex (ABC) method.35 In our preliminary study, the sensitivities and specificities of three different antibodies—AE1/AE3, KL-1 (Immunotech, Marseilles, France), and CAM5.2 (Becton Dickinson, San Jose, CA)—were evaluated using primary tumor tissues. After comparison of immunostaining and H&E staining, AE1/AE3 was selected as the most sensitive for the detection of micrometastases. The immunostained lymph node sections were evaluated by three observers (T.F., T.S., and G.B.M.) without knowledge of clinicopathologic information. Micrometastasis was defined as the presence of tumor cells detected only by cytokeratin specific immunostaining that could not be detected by ordinary H&E staining. AE1/AE3 positive tumor cells had diffuse cytoplasmic staining with a more intensely stained peripheral band. There were two patterns of micrometastases: Micrometastases consisting of a sin-

Micrometastases in Gastric Carcinoma/Fukagawa et al.

755

TABLE 1 Frequency of Micrometastases Measure

Patients

Lymph nodes

Micrometastases Total no. Percentage of total

38 107 35.5

87 4484 1.96

Statistical analyses were performed using the chisquare test to examine the correlation between micrometastases and clinicopathologic factors. The Kaplan– Meier method was used for calculating the survival curves, and the log-rank test was used to evaluate the statistical difference between survival curves.

RESULTS Micrometastases

FIGURE 1. Micrometastases of the single-cell type. Cytokeratin positive cells are involved in the lymph nodes (original magnification, ⫻200).

Micrometastases were identified in 38 of 107 patients (35.5%) and in 87 of 4484 lymph nodes (1.94%) (Table 1). With respect to the 38 patients with micrometastases, the median number of involved lymph nodes was 1 (range, 1–10 lymph nodes), and the mean was 2.3 lymph nodes per patient. Seventeen patients had micrometastases of the single-cell type, and 21 had micrometastases of the cluster type.

Correlation between Micrometastases and Clinical Factors

FIGURE 2. Micrometastases of the cluster type. Small clusters consisted of cytokeratin positive cells (original magnification, ⫻200).

gle cell were classified as the single-cell type (Fig. 1). Micrometastases consisting of clusters of two or more tumor cells were classified as the cluster type (Fig. 2). When both single tumor cells and clusters were observed in a lymph node, they were classified as the cluster type.

The demographic characteristics of patients and the pathologic features of tumors with and without micrometastases are shown in Table 2. Females were more likely to have micrometastases, but this was not statistically significant (16 of 34 females vs. 22of 73 males; P ⫽ 0.13). The frequency of micrometastases was related to the depth of the primary tumor. Micrometastases were more frequent in tumors invading the subserosal layer (12 of 28 tumors; 43%) compared with tumors that were limited to the muscularis propria (26 of 79 tumors; 33%), but this difference was not statistically significant (P ⫽ 0.37). Undifferentiated tumors were less likely to have micrometastases compared with differentiated tumors (16 of 55 undifferentiated tumors vs. 22 of 52 differentiated tumors; P ⫽ 0.15). Both the mean size and the proportion of lesions measuring ⬎ 30 mm in greatest dimension were larger in micrometastasis positive patients compared with negative patients, but the P value was 0.07 for both groups. The presence of lymphatic invasion (P ⫽ 0.26) and venous invasion (P ⫽ 0.69) was not associated with micrometastases, although it should be noted that only seven tumors had vascular invasion. Lesions with infiltrative growth characteristics were significantly more likely to have micrometastases compared

756


TABLE 2 Correlation between Micrometastases and Clinicopathologic Findings

TABLE 3 Correlation between Type of Micrometastases and Clinicopathologic Findings

Micrometastases Micrometastases Variable Age in yrs (mean ⫾ SD) Gender Male Female Depth of tumor MP SS Type of tumor 1 2 3 4 5 Pathologic type Differentiated Undifferentiated Greatest tumor dimension (in mm) (mean ⫾ SD) ⬉ 30 ⬎ 30 Lymphatic invasion Negative Positive Vascular invasion Negative Positive Growth characteristics Expansive Infiltrative

Positive (n ⴝ 38)

Negative (n ⴝ 69)

P value

57.8 ⫾ 10.8

59.8 ⫾ 12.1

—

Variable

22 16

51 18

0.13 —

26 12

53 16

0.37 —

3 9 24 1 1

7 19 40 0 3

0.27 — — — —

16 22

39 30

0.15 —

46.6 ⫾ 26.3 10 28

38.2 ⫾ 20.1 31 38

— 0.07 —

20 18

44 25

0.26 —

36 2

64 5

0.69 —

23 15

56 13

0.02 —

Age in yrs (mean ⫾ SD) Gender Male Female Depth of tumor MP SS Type of tumor 1 2 3 4 5 Pathologic type Differentiated Undifferentiated Greatest tumor dimension (in mm) (mean ⫾ SD) ⬉ 30 ⬎ 30 Lymphatic invasion Negative Positive Vascular invasion Negative Positive Growth characteristics Expansive Infiltrative

SD: standard deviation; MP: muscularis propria; SS: subserosa.

Single-cell type (n ⴝ 17)

Cluster type (n ⴝ 21)

P value

57.7 ⫾ 9.81

57.8 ⫾ 11.7

—

11 6

11 10

0.44 —

11 6

15 6

0.66 —

1 4 12 0 0

2 5 12 1 1

0.73 — — — —

7 10

9 12

0.92 —

40.0 ⫾ 20.3 7 10

52.0 ⫾ 29.8 3 18

— 0.06 —

11 6

9 12

0.18 —

15 2

21 0

0.11 —

10 7

13 8

0.85 —

SD: standard deviation; MP: muscularis propria; SS: subserosa.

with lesions that had expansive characteristics. (15 of 28 lesions vs. 23 of 79 lesions, respectively; P ⫽ 0.02). Similarly, the correlation between the type of micrometastases and clinicopathologic factors is shown in Table 3. The size of the primary tumor was greater in patients who had cluster type micrometastases compared with patients who had single-cell type micrometastases. Lymphatic invasion was seen more frequently in patients with cluster type micrometastases. The location of micrometastases (N1, N2, or N3 lymph nodes) was not associated significantly with the depth, pathologic type, or size of the primary tumor. Micrometastases were observed beyond first-tier lymph nodes (N2 and N3) in 6 of 28 patients in whom the tumor extended into the subserosa compared with 9 of 79 patients in whom the tumor was limited to the muscularis propria (P ⫽ 0.37). Micrometastases were found more frequently beyond N1 lymph nodes when the histologic type was undifferentiated (10 of 52 tu-

mors) compared with those of the differentiated histologic type (5 of 55 tumors), but the difference was not significant (P ⫽ 0.38). The incidence of micrometastases beyond N1 lymph nodes was not related to tumor size: Four of 41 patients had tumors measuring up to 30 mm in greatest dimension, whereas 11 of 66 patients had tumors measuring ⬎ 30 mm in greatest dimension (P ⫽ 0.32). The type of micrometastases was associated with the location of the micrometastases: Nine of 17 patients with single-cell micrometastases had metastases beyond N1 lymph nodes compared with only 6 of 21 patients with cluster micrometastases (P ⫽ 0.12) (Table 4). Skip micrometastases were found in four patients. In three patients, there were micrometastases in second-tier lymph nodes without perigastric lymph node involvement, and one patient had isolated para-aortic lymph node micrometastases.


757

TABLE 4 Stage of Micrometastases Variable Depth MP SS Pathologic type Differentiated Undifferentiated Greatest tumor dimension (mm) ⬉ 30 31–60 ⭌ 61 Micrometastases type Single cell Cluster Totals (NS)

Group 1 (perigastric)

Group 2 (regional)

Group 3 (distant)

17 6

7 5

2 1

11 12

3 9

2 1

6 12 5

4 6 2

0 3 0

8 15 23

6 6 12

3 0 3

MP: muscularis propria; SS: subserosa; NS: not significant.

FIGURE 3. TABLE 5 Correlation between Micrometastases and Recurrence of Tumor Micrometastases Recurrence

Positive

Negative

Positive Negative

2a 36

4 65

a

One patient with each type of micrometastasis developed recurrence. The overall results were not significant.

Kaplan–Meier survival curves for 107 patients with pT2N0M0 gastric carcinoma. (a) Thirty-eight patients with micrometastases to the lymph nodes. (b) Sixty-nine patients without micrometastases to the lymph nodes. There was no significant difference between the two survival curves (log-rank test; P ⫽ 0.86).

89%, and 74%, respectively. The survival curves of patients with or without micrometastases were nearly superimposed (P ⫽ 0.87) (Fig. 3). If the patients who died of other diseases are excluded, then there is still no significant difference. The type of micrometastases did not affect the survival curve (Fig. 4).

Prognosis The median follow-up was 120 months (range, 71–185 months). Disease recurrence was observed in only six patients. Two patients had micrometastases, and four patients did not. Recurrence was not related to the presence of micrometastases (P ⫽ 0.91) (Table 5). Among the patients without micrometastases, four died of disease recurrence. The mode of recurrence in these patients was unclear. Five died of unrelated malignancies, three died of benign disease, three died of unknown reasons, and one died from a postoperative complication. Only two of the patients with micrometastases died of disease recurrence. Both of these patients had peritoneal recurrence. Two patients died of unrelated malignancies. One patient died of benign disease. Four patients died of unknown causes but had no evidence of recurrent gastric carcinoma on clinical examination and liver ultrasound shortly before their death. The 5-year and 10-year survival rates of patients with and without micrometastases were 94%, 79%,

DISCUSSION There is increasing interest in the presence and prognostic relevance of micrometastases in lymph nodes in various malignant diseases. There have been reports on patients with breast and colorectal carcinoma and patients with melanoma that micrometastases detected using IHC and/or reverse transcriptase-polymerase chain reaction analysis can identify a subset of patients with negative lymph node who are at greater risk of disease recurrence.5,14,20 There is debate regarding whether the identification of isolated tumor cells is of clinical significance and whether it should be incorporated into staging systems. Some point to a large body of evidence showing that tumor cells can be found in a variety of sites with no clinical impact on patient prognosis,36 whereas others suggest that the finding of viable tumor cells should be included in the staging system.37 In this setting, it is important to search for associations between micrometastases and clinical outcomes. There also is debate regarding the terminology of

758


FIGURE 4.

Kaplan–Meier survival curves for 107 patients with pT2N0M0 gastric carcinoma. (a) Twenty-one patients with cluster type micrometastases to the lymph nodes. (b) Seventeen patients with single cell type micrometastases to the lymph nodes. (c) Sixty-nine patients without micrometastases to the lymph nodes.

micrometastases. Some regard occult metastases as those not observed on routine histologic examination,38 whereas others consider that all metastases measuring ⬍ 2 mm in greatest dimension are micrometastases.39 Several authors use the term micrometastasis to denote established metastasis accompanying surrounding tissue reaction, distinguishing them from single cells in lymph nodes without reaction, which are called microinvolvement.30 –32 Our definition of micrometastases was the presence of tumor cells in regional lymph nodes not detected on routine H&E staining. Many lymph node metastases detected at our institute measure ⬍ 2 mm in greatest dimension.40 We arbitrarily divided the metastases into single-cell and cluster types in light of previous suggestions that this differentiation may be important. However, we found no differences in outcome between the two types. We found that large numbers of our patients (35.5%) had micrometastases, with approximately equal numbers of single-cell type and cluster type metastases. Only 2 of 118 patients had lymph node metastases on this review, attesting to the accuracy of the original pathologic reporting. The antibody we used (AE1/AE3) is well known to have high specificity, unlike other antibodies.41 We excluded some cells that were stained weakly with AE1/AE3, which were identifiable as dendritic reticulum cells in H&E sections.42 Most of the micrometastases were found in N1 lymph

nodes, supporting the concept that lymph nodal spread is orderly and predictable. In 4 of 38 patients, the micrometastases were found beyond the first tier of lymph nodes, with no disease found in the first-tier lymph nodes. It is interesting to note that more of the single-cell type metastases were found beyond the first lymph node tier. The current study did not address the viability of tumor cells, so we cannot draw any conclusions regarding the clinical significance of this finding. In patients with gastric carcinoma, the percentage of lymph nodes with micrometastases among all lymph nodes examined has been reported as 4.8%29 to 13%30 for pN0 patients. In our study, 1.96% of the examined lymph nodes were involved by micrometastases. The relatively low overall incidence of micrometastasis is mainly because this study was restricted to pT2, pN0 patients. The incidence of micrometastases depends on the depth of the primary tumor, as does the incidence of lymph node metastasis.28,29,43 In the current study, micrometastases were found in 26 of 79 patients (32.9%) with invasion limited to the muscularis propria and in 12 of 28 patients (42.9%) with invasion to the subserosal layer. These results correspond well to the earlier reports: 65% for patients with pT3pN0 disease,30 33% for patients with pT2pN0 disease,29 and 23.5% for patients with pT1 disease.33 The only clinicopathologic characteristic that had a significant association with micrometastases was infiltrative growth. Of other features that intuitively would be expected to be associated with an increased propensity to lymph node spread, the statistically insignificant trend of the results was in the predicted direction. Thus, micrometastases were more likely to be found in subserosal tumors, larger tumors, and tumors with lymphatic or vascular invasion. It would be interesting to analyze results from a larger cohort of patients. Many authors have reported the existence of lymph node micrometastases in patients with gastric carcinoma, but their prognostic significance remains controversial. The majority of papers are retrospective analyses of small numbers of selected patients. The most frequent pattern of recurrence of T3 and T4 tumors is peritoneal dissemination.2 If locally advanced tumors are included, then the prognostic significance of micrometastasis in lymph nodes is likely to be confounded by frequent peritoneal recurrences. Multivariate analysis of large numbers of patients may address the impact of micrometastatic disease. However, a strong correlation between lymph node micrometastases and peritoneal dissemination would confound multivariate analysis as well. For this reason, we used exclusively patients with disease that was staged


as pT2, pN0 by conventional H&E. staining. Analysis of a large group of uniform tumors allows more definitive conclusions to be drawn regarding the impact of micrometastases. Cai et al.30 reported that lymph node micrometastasis led to a significant worse prognosis in patients with pT3, pN0 gastric carcinoma. The larger and more poorly differentiated tumors had significantly more micrometastases. The pattern of recurrence was not reported, but it is possible that lymph nodal micrometastases in these patients identifies a biologically distinct group of T3 tumors that are at high risk of peritoneal recurrence. Our study provides support for this view, because our infiltrative tumors had more micrometastases than expansive tumors (P ⫽ 0.02). Infiltrative tumors have a strong tendency to develop peritoneal seeding when they evolve into T3–T4 tumors.44 One of the problems with assessing the outcome of patients who undergo potentially curative treatment for gastric carcinoma is that standardization of surgical treatment is difficult. In the recently reported Dutch trial comparing the effectiveness of D1 and D2 lymph node dissection, there was a significant incidence of contamination or noncompliance. The advantage of this single institutional study is that all of the operations were done by or under the supervision of a specialist gastric surgeon who performed standardized surgery. Occult micrometastases to regional lymph nodes were not associated with poorer survival in patients with pT2N0 gastric carcinoma who underwent extended lymph node dissection. Although the very good prognosis in our patients meant that survival differences were unlikely to be found, we had expected that there would be some association between the presence of micrometastases and recurrence of gastric carcinoma. This was not the case. Our significant incidence of micrometastases means that this remains local disease, and the micrometastases are not a marker of systemic disease. This supports our policy of performing comprehensive lymph node dissection. This study was designed to address the questions of whether the presence of micrometastases in lymph nodes is an indicator of a worse prognosis. Further studies will involve comparing our data with data from similarly selected and analyzed patients from a Western institution. Furthermore, we are considering analyzing the N2 lymph nodes in patients who have been found to have macrometastases in N1 lymph nodes. This study may shed more light on the influence of such minimal disease. There has been a recent suggestion that analysis of the selected gastric lymph

759

nodes may be able to provide a means of rationally selecting those patients who may benefit from D2 lymph node dissection.45

CONCLUSIONS The current results show that the presence of immunohistochemically detected micrometastases in the lymph nodes of patients with pT2N0 gastric carcinoma did not affect their prognosis, and that such micrometastases do not imply that the disease is beyond cure with local treatment. Because all of our patients underwent extended lymph node dissection, our data could not justify the practice. The relatively high incidence of micrometastases in lymph nodes of patients with pT2N0 disease and the excellent prognosis of patients after undergoing extended dissection confirm our view that surgery is quite appropriate if it is done with a very low operative mortality.

REFERENCES 1.

Moriguchi S, Maehara Y, Korenaga D, Sugimachi K, Nose Y. Risk factors which predict pattern of recurrence after curative surgery for patients with advanced gastric cancer. Surg Oncol 1992;1:341– 6. 2. Katai H, Maruyama K, Sasako M, Sano T, Okajima K, Kinoshita T, et al. Mode of recurrence after gastric cancer surgery. Digest Surg 1994;11:99 –103. 3. Shimada K, Ajani JA. Adjuvant therapy for gastric carcinoma patients in the past 15 years. Cancer 1999;86:1657– 68. 4. Sano T, Sasako M, Kinoshita T, Maruyama K. Recurrence of early gastric cancer. Cancer 1993;72:3174 – 8. 5. McGuckin MA, Cummings MC, Walsh MD, Hohn BG, Bennett IC, Wright RG. Occult axillary node metastases in breast cancer: their detection and prognostic significance. Br J Cancer 1996;73:88 –95. 6. Dowlatshahi K, Fan M, Snider HC, Habib FA. Lymph node micrometastases from breast carcinoma. Cancer 1997;80: 1188 –97. 7. Natsugoe S, Mueller J, Stein HJ, Feith M, Hofler H, Siewert JR. Micrometastasis and tumor cell microinvolvement of lymph nodes from esophageal squamous cell carcinoma. Cancer 1998;83:858 – 66. 8. Glickman JN, Torres C, Wang HH, Turner JR, Shahsafaei A, Richards WG, et al. The prognostic significance of lymph node micrometastasis in patients with esophageal carcinoma. Cancer 1999;85:769 –78. 9. Liefers GJ, Jansen AMC, van de Velde CJH, Hermans J, van Krieken JHJM, Cornelisse CJ, Tollenaar RAEM. Micrometastases and survival in Stage II colorectal cancer. N Engl J Med 1998;339:223– 8. 10. Oberg A, Stenling R, Tavelin B, Lindmark G. Are lymph node micrometastases of any clinical significance in Dukes Stages A and B colorectal cancer? Dis Colon Rectum 1998;41: 1244 –9. 11. Jeffers MD, O’Dowd GM, Mulcahy H, Stagg M, O’Donoghu DP, Toner M. The prognostic significance of immunohistochemically detected lymph node micrometastases in colorectal carcinoma. J Pathol 1994;172:183–7.

760


12. Nicholson AG, Marks CG, Cook MG. Effect on lymph node status of triple leveling and immunohistochemistry with CAM5.2 on node negative colorectal carcinomas. Gut 1994; 35:1447– 8. 13. Davidson BR, Sams VR, Styles J, Deane C, Boulos PB. Detection of occult nodal metastases in patients with colorectal carcinoma. Cancer 1990;65:967–70. 14. Sasaki M, Watanabe H, Jass JR, Ajioka Y, Kobayashi M, Matsuda K, et al. Occult lymph node metastases detected by cytokeratin immunohistochemistry predicted recurrence in “node-negative” colorectal cancer. J Gastroenterol 1997;32: 758 – 64. 15. Nakanishi Y, Ochiai A, Yamauchi Y, Moriya Y, Yoshimura K, Hirohashi S. Clinical implications of lymph node micrometastases in patients with colorectal cancers. Oncology 1999; 57:276 – 80. 16. Maehara Y, Yamamoto M, Oda S, Baba H, Kusumoto T, Ichiyoshi Y, et al. Cytokeratin-positive cells in bone marrow for identifying distant micrometastasis of gastric cancer. Br J Cancer 1996;73:83–7. 17. Lindeman F, Schlimok G, Dirschedl P, Witte J, Riethmuller G. Prognostic significance of micrometastatic tumour cells in bone marrow of colorectal cancer patients. Lancet 1992; 340:685–9. 18. Cote RJ, Rosen PP, Lesser ML, Old LJ, Osborne MP. Prediction of early relapse in patients with operable breast cancer by detection of occult bone marrow micrometastases. J Clin Oncol 1991;9:1749 –56. 19. Kakeji Y, Maehara Y, Shibahara K, Hasuda S, Tokunaga E, Oki E, Sugimachi K. Clinical significance of micrometastasis in bone marrow of patients with gastric cancer and its relation to angiogenesis. Gastric Cancer 1999;2:46 –51. 20. Hayashi N, Ito I, Yanagisawa A, Kato Y, Imaoka S, Watanabe H, et al. Genetic diagnosis of lymph-node metastasis in colorectal cancer. Lancet 1995;345:1257–9. 21. Noguchi S, Hiratsuka M, Furukawa H, Aihara T, Kasugai T, Tamura S, et al. Detection of gastric cancer micrometastases in lymph nodes by amplification of keratin 19 Mrna with reverse transcriptase-polymerase chain reaction. Jpn J Cancer Res 1996;87:650 – 4. 22. Hayashi N, Arakawa H, Nagase H, Yanagisawa A, Kato Y, Ohta H, et al. Genetic diagnosis identifies occult lymph node metastases undetectable by the histopathological method. Cancer Res 1994;54:3853– 6. 23. Yamamoto N, Kato Y, Yanagisawa A, Ohta H, Takanashi T, Kitagawa T. Predictive value of genetic diagnosis for cancer micrometastasis. Cancer 1997;80:1393– 8. 24. Ikeguchi M, Oka A, Tsujitani S, Maeta M, Kaibara N. Relationship between area of serosal invasion and intraperitoneal free cancer cells in patients with gastric cancer. Anticancer Res 1994;14:2131– 4. 25. O’Sullivan GC, Collins JK, Kelly J, Morgan J, Madden M, Shanahan F. Micrometastases: marker of metastatic potential or evidence of residual disease. Gut 1997;40:512–5. 26. Moll R, Franke WW, Schiller DL, Geiger B, Krepler R. The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells. Cell 1982;31:11–24. 27. Listrom MB, Dalton LW. Comparison of keratin monoclonal antibodies MAK-6, AE1:AE3, and CAM-5.2. Am J Clin Pathol 1987;88:297–301. 28. Ishida K, Katsuyama T, Sugiyama A, Kawasaki S. Immunohistochemical evaluation of lymph node micrometastases from gastric carcinomas. Cancer 1997;79:1069 –76.

29. Kikuchi Y, Tsuchiya A, Ando Y, Yoshida T, Takenoshita S. Immunohistochemical detection of lymph node microinvolvement in node-negative gastric cancer. Gastric Cancer 1999;2:173– 8. 30. Cai J, Ikeguchi M, Maeta M, Kaibara N, Sakatani T. Clinicopathological value of immunohistochemical detection of occult involvement in pT3N0 gastric cancer. Gastric Cancer 1999;2:95–100. 31. Natsugoe S, Nakashima S, Matsumoto M, Nakajo A, Miyazono F, Kijima F, et al. Paraaortic lymph node micrometastases and tumor cell microinvolvement in advanced gastric carcinoma. Gastric Cancer 1999;2:179 – 85. 32. Siewert JR, Kestlmeier R, Busch R, Bottcher K, Roder JD, Muller J, 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. 33. Maehara Y, Oshiro T, Endo K, Baba H, Oda S, Ichiyoshi Y, et al. Clinical significance of occult micrometastasis in lymph nodes from patients with early gastric cancer who died of recurrence. Surgery 1996;119:397– 402. 34. Japanese Gastric Cancer Association. Japanese Classification of Gastric Carcinoma, 13th ed. Tokyo: Japanese Gastric Cancer Association, 1999. 35. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 1981;29:577– 80. 36. Hermanek P, Hutter RV, Sobin LH, Wittekind C. Classification of isolated tumor cells and micrometastasis. Cancer 1999;86:2668 –73. 37. Scheunemann P, Izbicki JR, Pantel K. Tumorigenic potential of apparently tumor-free lymph nodes. N Engl J Med 1999; 340:1687. 38. International (Ludwig) Breast Cancer Study Group. Prognostic importance of occult axillary lymph node micrometastases from breast cancers. Lancet 1990;335:1565– 8. 39. Turner RR, Ollila DW, Stern S, Giuliano AE. Optimal histopathologic examination of the sentinel lymph node for breast carcinoma staging. Am J Surg Pathol 1999;23:263–7. 40. Noda N, Sasako M, Yamaguchi N, Nakanishi Y. Ignoring small lymph nodes can be a major cause of staging error in gastric cancer. Br J Surg 1998;85:831– 4. 41. Xu X, Roberts SA, Pasha TL, Zang PJ. Undesirable cytokeratin immunoreactivity of native nonepithelial cells in sentinel lymph nodes from patients with breast carcinoma. Arch Pathol Lab Med 2000;124:1310 –3. 42. Latza U, Niedobitek G, Schwarting R, Nekarda H, Stein H. Ber-EP4: new monoclonal antibody which distinguishes epithelia from mesothelia. J Clin Pathol 1990;43:213–9. 43. Maruyama K, Gunven P, Okabayashi K, Sasako M, Kinoshita T. Lymph node metastases of gastric cancer. General pattern in 1931 patients. Ann Surg 1989;210:596 – 602. 44. Fujino K, Ichikura T, Hase K, Tomimatsu S, Uefuji K, Tamakura S. Peritoneal lavage cytology in gastric cancer operation—a study on the prognostic value using multivariate analysis [in Japanese with English abstract]. Jpn J Gastroenterol Surg 1996;25:1741–5. 45. Whiting JL, Hallissey MT, Rowlands DC, Fielding JWL. Redefining surgery for gastric cancer. Gastric Cancer 1999;2: 226 –9.