Sentinel node biopsy for cT1 and cT2a gastric cancer

EJSO 32 (2006) 48–54

www.ejso.com

Sentinel node biopsy for cT1 and cT2a gastric cancer D.J. Parka,c, H.-J. Leea,c, H.S. Leeb, W.H. Kimb,c, H.-H. Kima, K.U. Leea, K.J. Choea, H.-K. Yanga,c,* a

Department of Surgery, Seoul National University College of Medicine, 28 Yongon-dong, Chongno-gu, Seoul 110-744, South Korea b Department of Pathology, Seoul National University College of Medicine, Seoul, South Korea c Cancer Research Institute, Seoul National University College of Medicine, Seoul, South Korea Accepted 5 September 2005 Available online 2 November 2005

Abstract Aims: To evaluate the feasibility and accuracy of sentinel node (SN) biopsy for gastric cancer. Patients and methods: One hundred patients with gastric cancer diagnosed as cT1 (nZ80) or cT2a (nZ20) were enrolled. Indocyanine greenstained SNs were analysed by hematoxylin and eosin staining (nZ100) and by cytokeratin immunohistochemistry (nZ50). Results: SNs were identified in 94 of the 100 patients and the mean number of SNs was 4.4 (range, 1—12). Of these 94 patients, 14 patients had lymph node metastases. Two patients with T1 and one patient with T2 had metastases in non-SNs alone by hematoxylin and eosin staining (diagnostic accuracy Z97.3% in T1 and 95.0% in T2). All three patients with a false negative result had a tumour, which was more than 4 cm in size and signet ring cell histology. In two of them, the tumour was located at lesser curvature. By immunohistochemical staining, three patients with T1 and one patient with T2 were found to have lymph node micrometastases in non-SNs alone among 45 patients (diagnostic accuracy Z92.1% in T1, 85.7% in T2). Conclusion: SN biopsy using indocyanine green can be performed rapidly and easily with a high detection rate and accuracy in patients with T1 gastric cancer. However, it should be performed with caution for large tumours with a signet ring cell histology located at lesser curvature due to the possibility of a false negative result. q 2005 Elsevier Ltd. All rights reserved. Keywords: Sentinel node biopsy; Gastric cancer; Accuracy; Indocyanine green; Micrometastasis

Introduction A sentinel node (SN) is defined as the first draining node from a primary lesion, and it represents the first site of metastasis or micrometastasis.1 SN biopsy was introduced to determine the extent of lymph node dissection in early gastric cancer. If SN biopsies have high accuracy, limited lymph node dissection can be performed for early gastric cancer without SN metastasis. Clinical researches on SN biopsy have been conducted for various solid tumours,2–9 and the diagnostic validity of SN biopsy has been confirmed for breast cancer and melanoma.1,2,10 Although gastric lymphatic flows are multidirectional and complex, several

* Corresponding author. Department of Surgery, Seoul National University College of Medicine, 28 Yongon-dong, Chongno-gu, Seoul 110-744, South Korea. Tel.: C82 2 2072 3797; fax: C82 2 3672 0047. E-mail address: [email protected] (H.-K. Yang).

0748-7983/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejso.2005.09.006

studies reported that SN biopsy is feasible in gastric cancer.11–13 However, studies about the feasibility of SN biopsy in gastric cancers involving micrometastasis analysis are rare. This study aimed to evaluate the feasibility and accuracy of SN biopsy for cT1 and cT2a gastric cancer by conventional hematoxylin and eosin staining, and by immunohistochemical staining. Patients and methods Patients One hundred patients who underwent resection with curative intent for cT1 or cT2a gastric cancer without distant metastasis from March 2002 to June 2004 were enrolled. Age, sex, tumour location, gross tumour appearance, histologic type, Lauren classification,14 tumour size, depth of invasion, and lymph node metastasis were investigated.

D.J. Park et al. / EJSO 32 (2006) 48–54

SN biopsy In early gastric cancer, preoperative endoscopic clipping was done to palpate the lesion intraoperatively. One metallic clip is usually placed at the mucosal side about 1 cm proximally from the lesion before the operation. After the lesion had been identified in the operative field, indocyanine green (ICG, 25 mg/5 ml, Diagnogreenw, Tokyo, Japan) was injected into four different sites of the subserosal layer around the main lesion with a 26 gauge needle. Five minutes after injection, green-stained nodes were considered SNs and tagged with black silk (Fig. 1). After gastrectomy with D2 lymph node dissection had been performed, SNs and non-SNs were collected according to the Japanese classification of lymph node stations.15 The collected nodes were examined by hematoxylin and eosin (H&E) staining in 100 patients and by immunohistochemical (IHC) staining (standard technique) in the first 50 patients. Definitions Sensitivity (number of patients with a positive SN biopsy/number of patients with lymph-node metastases), specificity (number of patients with a negative SN biopsy/number of patients with no lymph-node metastases), positive predictive value (number of patients with lymphnode metastases/number of patients with a positive SN biopsy), negative predictive value (number of patients

49

without lymph node metastases/number of patients with a negative SN biopsy), overall accuracy (number of patients with true-positive and true-negative SN biopsies/number of patients in whom SNs were identified), and false-negative rate (number of patients with a negative SN biopsy/number of patients with lymph-node metastases) were calculated. Results One hundred patients were enrolled and their characteristics are summarized in Table 1. Subtotal gastrectomy was performed for 98 patients and total gastrectomy for two. The mean (GSD) number of resected lymph nodes was 37.3G12.4. ICG dye was injected around a tumour and a green line was observed moving toward the greater curvature, lesser curvature, or both. A green node was observed within 5 min of dye injection. Of 100 patients, SNs were identified in 94 patients. SN was unidentifiable grossly in one patient with fatty omentum and microscopically in other five. The mean number of SNs was 4.4. When the 100 patients were divided into two groups based on the chronologic order of operation (1—50 vs 51—100), detection rate was higher in the second group than in the first group. Although it was not statistically significant, the mean number of SNs was greater in the second group (nZ4.6) than in the first (nZ4.3). Table 1 Patient characteristics

Figure 1. Sentinel nodes (arrows) were stained green within 5 min after injecting indocyanine green into the subserosal layer around the lesion.

Factors

Number

Sex ratio (M:F) Age (years) Tumour location Upper Middle Lower Tumour size (cm) ! 2 cm 2—4 cm R4 cm Depth of invasion Mucosa Submucosa Proper muscle Subserosa Lauren classification Intestinal Diffuse Mixed Histologic type Well differentiated Moderately differentiated Poorly differentiated Signet ring cell Mucinous

2.2:1 (69:31) 55.6 (range, 27—78) 2 26 72 23 54 23 49 31 12 8a 60 34 6 19 41 17 20 3

a These patients were diagnosed as cT2a but were later found to be T2b by pathologic examination.

50

D.J. Park et al. / EJSO 32 (2006) 48–54

Table 2 Location of sentinel nodes Tumour(SN)

Anterior wall

Posterior wall

Lesser curvature

Greater curvature

Total

Lesser curvature Greater curvature Both Total

6 3 9 18

10 5 5 20

26 1 7 34

0 19 3 22

42 28 24 94

SN, sentinel node.

Table 3 Results of sentinel node biopsies Depth of invasion

No.

Hematoxylin and eosin staining Mucosa 44 Submucosa 30 T2 20 Total 94a Immunohistochemical staining Mucosa 24 Submucosa 14 T2 7 Total 45b

Lymph node metastasis (%)

Metastasis in SN/non-SN K/K

C/K

C/C

K/C

2 (4.5) 4 (13.3) 8 (40.0) 14(14.9)

42 26 12 80

0 3 2 5

1 0 5 6

1 1 1 3

3 (12.5) 5 (35.7) 6 (85.7) 14 (31.1)

21 9 1 31

0 2 0 2

1 2 5 8

2 1 1 4

SN, sentinel node. a Out of a total of 100 patients, six patients with detection failure were excluded. b Out of the first 50 patients, four patients with detection failure and one patient with an unavailable paraffin block were excluded.

Locations of SNs according to primary lesions are described in Table 2. SNs were distributed evenly at the lesser or greater curvature when the primary tumour was located at the anterior or posterior wall. When the primary

tumour was located at lesser or greater curvature, the SNs were mainly distributed on the same side as the tumour. Most SNs were located in the perigastric (N1) area, except three cases where SNs were located in the station eight area.

Figure 2. Two examples of SNs with metastasis on hematoxylin and eosin staining. Left: A type IIbCIIc, 7.8 cm sized, submucosal cancer with lymph node metastasis in SN alone. Right: A type IIc, 2.5 cm sized, mucosal cancer with lymph node metastasis in both SN and non-SN. LN, lymph node; Non-SN, nonsentinel node; SN, sentinel node; T1, a tumour whose T stage is T1; open circle, SN without metastasis; closed circle, SN with metastasis; closed triangle, nonSN with metastasis. Fractions designate the number of metastatic LNs/total number of examined LNs.

D.J. Park et al. / EJSO 32 (2006) 48–54

51

Figure 3. A false negative case. The gross tumour size was measured to be 3 cm, but microscopic mapping showed that the tumour size was 6.8 cm and the tumour had focal extension to the proper muscle layer. LN, lymph node; non-SN, non-sentinel node; SN, sentinel node; T2, a tumour whose T stage is T2; open circle, SN without metastasis; closed triangle, non-SN with metastasis. Fractions designate the number of metastatic LNs/total number of examined LNs.

The results of H&E staining and IHC staining are summarized in Table 3. Two examples of positive SNs were shown schematically in Fig. 2. Three patients had metastases in non-SNs but were without metastasis in SNs

by H&E staining. These false negative cases were as follows: a type IIc, 4.5 cm sized, mucosal cancer with a signet ring cell histology at the lower body lesser curvature, another type IIc, 4 cm sized, submucosal cancer with a

Table 4 Clinico-pathological features of the patients with lymph node metastases in non-sentinel nodes alone No.

Age/sex

Depth

Gross size

Pathologic Size

Type

Histology

Location

SN

Non-SN

Stain

1 2 3 4 5 6

49/M 37/F 38/M 54/F 62/M 42/M

M SM2 PM M SM2 SS

3 4.5 3 1.5 3 3

4.5 4 6.8 2 2 3.5

IIc IIc IIcC IIac IIc IIc B III

SRC SRC SRC SRC MD PD

L/LC L/PW Mi/LC L/GC Mi/LC L/PW

0a/5b 0/3 0/2 0/3 0/5 0/11

4/40 2/36 2/16 1/33 1/41 2/41

H, I H H I I I

SN, sentinel node; M, mucosa; SRC, signet ring cell; L, lower; LC, lesser curvature; H, hematoxylin and eosin; I, immunohistochemistry; SM, submucosa; PW, posterior wall; PM, proper muscle; Mi, middle; GC, greater curvature; SS, subserosa; BIII, Borrmann type III; PD, poorly differentiated. a Number of metastatic lymph nodes. b Number of total lymph nodes. c Gross tumour appearance mimicked type IIcCIIa early gastric cancer. Table 5 The accuracy, positive predictive value, negative predictive value, sensitivity, specificity, and false negative rate of sentinel node biopsy No. Hematoxylin and eosin staining T1 74 T2 20 Total 94 Immunohistochemical staining T1 38 T2 7 Total 45

Accuracy (%)

Positive predictive value (%)

Negative predictive value (%)

Sensitivity (%)

Specificity (%)

False negative rate (%)

72/74 (97.3) 19/20 (95.0) 91/94 (96.8)

4/4 (100) 7/7 (100) 11/11 (100)

68/70 (97.1) 12/13 (92.3) 80/83 (96.4)

4/6 (66.7) 7/8 (87.5) 11/14 (78.6)

68/68 (100) 12/12 (100) 80/80 (100)

2/6 (33.3) 1/8 (12.5) 3/14 (21.4)

35/38 (92.1) 6/7 (85.7) 41/45 (91.1)

5/5 (100) 5/5 (100) 10/10 (100)

30/33 (90.9) 1/2 (50.0) 31/35 (88.6)

5/8 (62.5) 5/6 (83.3) 10/14 (71.4)

30/30 (100) 1/1 (100) 31/31 (100)

3/8 (37.5) 1/6 (16.7) 4/14 (28.6)

52

D.J. Park et al. / EJSO 32 (2006) 48–54

Table 6 Accuracy according to tumour size and differentiation Tumour size Differentiation Differentiateda Undifferentiatedb Total

! 2 cm

2—4 cm

R4 cm

Total

14/14 (100) 6/6 (100) 20

32/32 (100) 19/19 (100) 51

9/9 (100) 11/14 (78.8) 23

55 39 94

Values in parentheses are percentages. a Includes well differentiated and moderately differentiated tumours. b Includes poorly differentiated, signet ring cell, and mucinous tumours.

signet ring cell histology at the lower body posterior wall, and the third 6.8 cm sized, proper muscle tumour mimicking a type IIaCIIc early gastric cancer, with a signet ring cell histology (Fig. 3). Four cases were false negative by IHC staining but they had no specific common characteristics (Table 4). The overall accuracy, positive predictive value, negative predictive value, sensitivity, specificity and false negative rate of sentinel node biopsy by H&E staining and IHC staining are summarized in Table 5. The overall accuracies according to tumour size and differentiation are shown in Table 6. Discussion SN biopsy in gastric cancer was considered difficult for clinical application because of the complex lymphatic pathways of the stomach and frequently skipped metastases, that is, the fact that cancer cells do not stop in every node in their lymphatic stream.16 However, the clinical application of SN biopsy in gastric cancer was carefully attempted in Japan after 2000. Hiratsuka et al.11 performed SN biopsy for patients with T1 and T2 gastric cancer using ICG, and reported that the sensitivity of SN biopsy was 100% in T1 and 88% in T2. Kitagawa et al.12 also performed sentinel node biopsy by using a radio-guided method and reported a diagnostic accuracy for SNs of 98.6%. Successive studies from many institutes have reported high detection rates or accuracies by SN biopsy in gastric cancer.13,17,18 Dye-guided SN mapping Methods of identifying SNs are categorized as dye method, radio-guided method, or both. Carbon particle and many vital dyes such as isosulfan blue, patent blue, methylene blue, indigo carmine, Evans blue, and ICG are available for SN mapping. We chose ICG for the present study because ICG was shown as one of good dyes for SN mapping of gastric cancer on previous studies11,19,23 and it is easily available at our institute. The dye can be injected into the submucosal layer or the subserosal layer. To inject dye into the submucosal layer, intraoperative gastroscopy or gastrotomy should be

done, but gastrotomy may block lymphatic pathway, which passes through the gastrotomy site and change the original lymphatic pathway. However, dye can be injected into the submucosal layer well by intraoperative gastroscopy and Miwa et al.17 reported a 96.2% detection rate and 98.0% accuracy by this method. Moreover, the method of injecting into the subserosal layer can be performed easily and Hiratsuka et al.11 reported a 99% detection rate and a 99% accuracy by this method. Although no randomized study has been undertaken to compare submucosal and subserosal injection, it is generally accepted that the two methods are equivalent. The amount of injected dye is an important factor that affects both the detection rate and sensitivity. Ichikura et al.19 reported that a larger number of SNs were observed in patients injected with 8 ml of ICG than in those injected with 4 ml. The dose range of injected dye in several studies was 0.8—8 ml.11,19,20 If the dose is too small, SN detection is difficult, whereas if the dose is too large, SN is stained too fast. This study was conducted by injecting 5 ml of ICG. The time of inspection after injection is another important factor in SN biopsy. There is possibility of missing original SNs as time elapses after injection because the dye passes through but does not stay in the node. Most studies using this dye method detected SNs within 5 min of injection.11,19,20 The dye method has a shortcoming that the identification of SN is difficult at deep sites and in fatty omentum, and becomes more difficult with time. Radio-guided SN mapping The radio-guided method involves injecting a radioisotope into the submucosal layer by preoperative gastroscopy. There is no unified method of using radio-isotope, and a variety of radio-isotope types and quantities, colloid types, and intervals from injection to surgical operation are used. When SNs are identified using an isotope, it is important to collimate the gamma probe properly to avoid interference due to strong background radioactivity emanating from the isotope injection site.8

D.J. Park et al. / EJSO 32 (2006) 48–54

Combined dye- and radio-guided SN mapping Some investigators have tried to increase sensitivity by using a combination of the dye and radio-guided methods.13, 21 Kitagawa et al.22 found that the SN detection rate could be increased if these two methods were used together, because the radio-guided method excelled at finding SNs in an unexpected site with a wide coverage, whereas the dye method allowed lymphatic flow to be observed in real time. In addition, Nimura et al.23 reported that infrared ray electronic endoscopy combined with ICG efficiently detects SNs. Detection failure and false negativity of SN biopsy Of six patients with detection failure in the present study, one patient had fatty omentum. Therefore, it can be postulated that the SN biopsy may be difficult in obese patients with fatty omentum. The other five patients had green nodes, but no lymph node was observed by microscopic examination. The detection failures may have resulted from technical problems while SNs were collected or preserved, or due to an error, e.g. one of the green-stained lymphatic lines where ICG passed through may have been regarded as an SN in a patient with poorly developed lymph nodes. The rate of detection failure in the first group (1—50 patients) was higher than that in the second (51—100 patients). Therefore, there exists a learning curve in SN biopsy for gastric cancer as has been noted in melanoma.1 Collecting SNs from the lymphatic basin rather than picking up one node may be effective at preventing any SN being overlooked and thus reducing false negative rates.17 On analysing false negative cases in this study, the tumour sizes on pathology were 4, 4.5, and 6.8 cm and the histologic types were all of the signet ring cell type. The tumours were mainly located at lesser curvature of stomach. In addition, the intraoperative gross sizes of these tumours with pathology determined sizes of 4.5 and 6.8 cm, were 3 and 3 cm, respectively, and, therefore, there was possibility of injecting into the wrong site depending on palpation of a clip; moreover, because the tumour size was large, the previous dose of dye would not be enough to cover the primary tumour.

53

was believed to be that although micrometastases were newly found in SNs of patients with no metastasis in SNs by H&E staining, micrometastases were also found in nonSNs. Sensitivity of SN biopsy The predictive value of SN biopsy has been reported to be high enough to accurately predict regional lymph node metastasis. However, in terms of sensitivity, the SN biopsy is found wanting. Morton25 emphasized the significance of sensitivity and suggested that the reason why regional and distant recurrence occurred when regional SNs are apparently tumour free lies in the lack of sensitivity of the technique used to sample and analyse the SNs. The sensitivity of SN biopsy in gastric cancer has been reported to be the range of 61.1—100%.11–13,17–20 Hiratsuka et al.11 reported a sensitivity of SN biopsy in T1 gastric cancer of 100% (2/2), but the denominator of the formula was just two. The largest denominator of the sensitivity formula reported to date was 35, and even this included T2 and T3 gastric cancer.17 In breast cancer, the denominator of sensitivity in a randomized controlled study or a large trial was nearly close to or substantially over 100.26,27 Therefore, we suggest that a multicenter large-scale trial should be conducted to properly evaluate the sensitivity of SN biopsy in T1 gastric cancer. In conclusion, SN biopsies using ICG can be performed rapidly and easily with a high detection rate and accuracy in patients with T1 gastric cancer. However, it should be performed with caution for large tumours with signet ring cell histology located at lesser curvature of stomach due to the possibility of a false negative result. The sensitivity of SN biopsy in this study was somewhat low, thus we suggest that the method should be subjected to a large-scale multicenter prospective validation study. Although the overall diagnostic accuracy of SN biopsy by IHC staining was not higher than that of H&E staining, because micrometastases were also found in non-SNs, IHC staining usefully detected additional micrometastases in SNs.

Acknowledgements SN micrometastasis Attempts have been made to increase the accuracy of SN biopsy by IHC staining with anti-cytokeratin antibody to detect micrometastasis in addition to conventional H&E staining.18,21 In addition, reverse transcription-polymerase chain reaction using tumour markers such as CEA, cytokeratin, or CD44 has been used to detect SNs more accurately.24 The present study also showed four additional SN micrometastases by IHC staining, but the overall diagnostic accuracy of SN biopsy by IHC staining was lower than that of SN biopsy by H&E staining. The reason

This study was supported by a grant no.04-2003-022-0 from the Seoul National University Hospital Research Fund.

References 1. Morton DL, Wen DR, Wong JH, Economou JS, Cagle LA, Storm FK, et al. Technical details of intraoperative lymphatic mapping for early stage melanoma. Arch Surg 1992;127:392–9. 2. Giuliano AE, Kirgan DM, Guenther JM, Morton DL. Lymphatic mapping and sentinel lymphadenectomy for breast cancer. Ann Surg 1994;220:391–401.

54

D.J. Park et al. / EJSO 32 (2006) 48–54

3. Saha S, Ganatra BK, Gauthier J, Beutler T, Nora D, Ganatra BK, et al. Localization of sentinel lymph node in colon cancer. A feasibility study. SSO 50th Ann Cancer Symp 1997;80:54. 4. Keleman PR, van Herle AJ, Giuliano AE. Sentinel lymphadenectomy in thyroid malignant neoplasms. Arch Surg 1998;133:288–92. 5. de Hullu JA, Doting ME, Piers DA, Hollema H, Aalders JG, Koops HS, et al. Sentinel lymph node identification with technetium-99m-labeled nanocolloid in squamous cell cancer of the vulva. J Nucl Med 1998;39: 1381–5. 6. Wawroschek F, Vogt H, Weckermann D, Wagner T, Harzmann R. The sentinel lymph node concept in prostate cancer—first results of gamma probe-guided sentinel lymph node identification. Eur Urol 1999;36: 595–600. 7. Little AG, DeHoyos A, Kirgan DM, Arcomano TR, Murray KD. Intraoperative lymphatic mapping for non-small cell lung cancer: the sentinel node technique. J Thorac Cardiovasc Surg 1999;117:220–4. 8. Kitagawa Y, Fugii H, Mukai M, Kubota T, Ando N, Watanabe M, et al. The role of the sentinel lymph node in gastrointestinal cancer. Surg Clin North Am 2000;80:1799–809. 9. Keshtgar MR, Amin A, Taylor I, Ell PJ. The sentinel node in anal carcinoma. Eur J Surg Oncol 2001;27:113–4. 10. Morton DL, Thompson JF, Essner R, Elashoff R, Stern SL, Nieweg OE, et al. Validation of the accuracy of intraoperative lymphatic mapping and sentinel lymphadenectomy for early-stage melanoma: a multicenter trial. Multicenter Selective Lymphadenectomy Trial Group. Ann Surg 1999;230:453–63. 11. Hiratsuka M, Miyashiro I, Ishikawa O, Furukawa H, Motomura K, Ohigashi H, et al. Application of sentinel node biopsy to gastric cancer surgery. Surgery 2001;129:335–40. 12. Kitagawa Y, Fujii H, Mukai M, Kubota T, Otani Y, Kitajima M. Radioguided sentinel node detection for gastric cancer. Br J Surg 2002;89: 604–8. 13. Hayashi H, Ochiai T, Mori M, Karube T, Suzuki T, Gunji Y, et al. Sentinel lymph node mapping for gastric cancer using a dual procedure with dye- and gamma probe-guided techniques. J Am Coll Surg 2003; 196:68–74. 14. Lauren P. The two histological main types of gastric carcinoma: Diffuse and so-called intestinal-type carcinoma. An attempt at a histo-clinical classification. Acta Pathol Microbiol Scand 1965;64: 31–49.

15. Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma. Gastric Cancer 1998;1:10–24 [ 2nd English Edition]. 16. Maruyama K, Sasako M, Kinoshita T, Sano T, Katai H. Can sentinel node biopsy indicate rational extent of lymphadenectomy in gastric cancer surgery? Fundamental and new information on lymph-node dissection Langenbecks Arch Surg 1999;384:149–57. 17. Miwa K, Kinami S, Taniguchi K, Fushida S, Fujimura T, Nonomura A. Mapping sentinel nodes in patients with early-stage gastric carcinoma. Br J Surg 2003;90:178–82. 18. Kim MC, Kim HH, Jung GJ, Lee JH, Choi SR, Kang DY, et al. Lymphatic mapping and sentinel node biopsy using 99mTc tin colloid in gastric cancer. Ann Surg 2004;239:383–7. 19. Ichikura T, Morita D, Uchida T, Okura E, Majima T, Ogawa T, et al. Sentinel node concept in gastric carcinoma. World J Surg 2002;26: 318–22. 20. Ryu KW, Lee JH, Kim HS, Kim YW, Choi IJ, Bae JM. Prediction of lymph nodes metastasis by sentinel node biopsy in gastric cancer. Eur J Surg Oncol 2003;29:895–9. 21. Aikou T, Higashi H, Natsugoe S, Hokita S, Baba M, Tako S. Can sentinel node navigation surgery reduce the extent of lymph node dissection in gastric cancer? Ann Surg Oncol 2001;8:90S–93. 22. Kitagawa Y, Ohgami M, Fujii H. Laparoscopic detection of sentinel lymph nodes in gastrointestinal cancer: a novel and minimally invasive approach. Ann Surg Oncol 2001;8:86S–89. 23. Nimura H, Narimiya N, Mitsumori N, Yamazaki Y, Yanaga K, Urashima M. Infrared ray electronic endoscopy combined with indocyanine green injection for detection of sentinel nodes of patients with gastric cancer. Br J Surg 2004;91:575–9. 24. Ajisaka H, Miwa K. Micrometastases in sentinel nodes of gastric cancer. Br J Cancer 2003;89:676–80. 25. Morton DL. Sentinel node mapping and an International Sentinel Node Society: current issues and future directions. Ann Surg Oncol 2004;11: 137S–143. 26. Veronesi U, Paganelli G, Viale G, Galimberti V, Luini A, Zurrida S, et al. Sentinel lymph node biopsy and axillary dissection in breast cancer: results in a large series. J Natl Cancer Inst 1999;91:368–73. 27. Veronesi U, Paganelli G, Viale G, Luini A, Zurrida S, Galimberti V, et al. A randomized comparison of sentinel-node biopsy with routine axillary dissection in breast cancer. N Engl J Med 2003;349:546–53.