- Email: [email protected]
Sentinel lymph node mapping for 385 gastric cancer patients
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Sentinel lymph node mapping for 385 gastric cancer patients Masahiro Niihara, MD,a Hiroya Takeuchi, MD, PhD,a,* Tadaki Nakahara, MD,b Yoshiro Saikawa, MD, PhD,a Tsunehiro Takahashi, MD, PhD,a Norihito Wada, MD, PhD,a Makio Mukai, MD, PhD,c and Yuko Kitagawa, MD, PhDa a
Department of Surgery, School of Medicine, Keio University, Tokyo, Japan Department of Radiology, School of Medicine, Keio University, Tokyo, Japan c Division of Diagnostic Pathology, School of Medicine, Keio University, Tokyo, Japan b
article info
abstract
Article history:
Background: The objectives were to investigate the accuracy of sentinel lymph node (SLN)
Received 19 February 2015
biopsy, detect the predictors for undetected or false-negative cases, evaluate the in-
Received in revised form
dications for SLN-navigated gastrectomy, and characterize the problems of SLN mapping in
25 June 2015
gastric cancer. The SLN concept may be applicable to early gastric cancer, particularly
Accepted 26 June 2015
clinical T1N0M0 or T2N0M0 with tumor diameter 4 cm.
Available online xxx
Methods: A total of 385 consecutive patients diagnosed with cT1N0M0 or cT2aN0M0 operable gastric cancer from April 1999 to December 2007 underwent radical gastrectomy with
Keywords:
SLN mapping. SLNs were identified using radio-guided and dye-guided methods. Predictors
Learning curve
for undetected or false-negative cases on SLN mapping were examined by multivariate
Sentinel basin
regression analysis.
Sentinel navigation surgery
Results: The detection rate of hot and/or blue nodes was 96.6% (372 of 385). The accuracy of
Long-term survival
metastatic status based on SLNs was 98.9% (368 of 372) for all cases in whom SLNs could be
Stabilization
detected. Furthermore, the accuracy of metastatic status based on SLNs was 99.1% (344 of 347)
Predictor
in cT1 gastric cancer and 96.0% (24 of 25) in cT2 gastric cancer. Pathologically, the tumors invaded to the muscularis propria or deeper in three of four false-negative cases. All but one case had metastatic lymph nodes within the sentinel basins. In terms of 5-y recurrence free survival, positive SLN cases (SLN(þ)) had a worse prognosis than negative SLN cases (SLN(); P ¼ 0.008). Moreover, SLN(þ) and non-SLN() cases (SLN(þ)/non-SLN()) had a similar prognosis as SLN(þ) and non-SLN(þ) cases (SLN(þ)/non-SLN(þ)) (P ¼ 0.511). On multivariate regression analysis, undetected or false-negative cases were significantly associated with the time period. Conclusions: The present results appeared to validate the SLN concept for untreated cT1 gastric cancer with tumor diameter 4 cm. SLN mapping may provide an effective method of staging the lymph node status of patients undergoing minimized gastrectomy. Sentinel basin dissection guards against the possibility of leaving positive lymph nodes. Stabilization of the procedure and experience with SLN mapping in gastric cancer might decrease undetected or false-negative cases. ª 2015 Elsevier Inc. All rights reserved.
* Corresponding author. Department of Surgery, School of Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. Tel.: þ81 3 3353 1211x62334; fax: þ81 3 3355 4707. E-mail address: [email protected] (H. Takeuchi). 0022-4804/$ e see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2015.06.064
2
1.
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Introduction
In the past decades, radical resection with lymphadenectomy for gastric cancer improved locoregional disease control and also improved survival in Far East [1]. The necessary extent of lymph node dissection has been heavily debated and controversial. Recently in Western countries also, D2 lymphadenectomy is recommended as the appropriate surgical approach for patients with resectable gastric cancer because of its long-term survival benefit. However, the routine D2 procedure has also been associated with significantly higher postoperative mortality, morbidity, and reoperation rates [2]. Meanwhile, the popularization of medical examinations and improvements in diagnostic technology make it possible to detect gastric cancer at earlier stages. Many of these patients have been cured by surgical treatment. Recently, laparoscopic-assisted gastrectomy has been improved and standardized. Therefore, the short-term quality of life of these patients is now considered to be better than before. However, such patients have suffered from the sequelae of gastrectomy, such as dumping syndrome, postvagotomy symptoms, and so on, over extended periods, even though some cases had no pathological lymph node metastases. The rate of lymph node metastasis reaches 10% in pathological mucosal gastric carcinoma, 25% in submucosal tumor, 45% in muscularis propria tumor, and 70% in subserosa tumor [3,4]. In fact, many resectable early-stage gastric cancers have no lymph node metastases. Accurate assessment of node status is mandatory to avoid unnecessary extended lymphadenectomy in nodenegative patients who may not get the full benefit of the standard D2 dissection. Given these problems, an individualized surgical procedure is needed, taking into account both curability and after effects, instead of a uniform method. The sentinel lymph node (SLN) concept may make it possible to select the appropriate extent of lymph node dissection to avoid unnecessary lymphadenectomy in node-negative patients. The SLN first receives lymphatic drainage from the tumor site [5]. The SLN concept that the SLNs are the possible site of first metastasis via the lymphatic route and that the pathological status of the SLN can predict metastases to all other downstream lymph nodes, has been evaluated in various solid tumors. We previously developed SLN mapping in gastric cancer and reported the procedure for the first time [6e9]. Some investigators demonstrated that the SLN concept can be applied in patients with early gastric cancer, in particular clinical T1N0M0 or T2N0M0 with a tumor diameter of 4 cm [10e12]. The present study is the largest single institutional report of SLN mapping in gastric cancer, to the best of our knowledge. The primary aim of the present work was to investigate the accuracy of the SLN of gastric cancer in our institution and detect the predictors for undetected or false-negative cases, focusing on the indication for SLN-navigated gastrectomy. These results supported the fundamental direction of sentinel node navigation surgery in gastric cancer, which involves functional preservation and is a minimally invasive surgical technique.
2.
Materials and methods
2.1.
Patients
All patients diagnosed with cT1N0M0 or cT2N0M0 operable primary gastric cancer at Keio University Hospital (Tokyo, Japan) during the period from April 1999 to December 2007 were enrolled in this study. Criteria for eligibility were, in principle, clinical T1N0M0 or T2aN0M0 gastric cancer according to the Union for International Cancer Control (sixth edition, 2002). The patients who obviously had surgical T3 tumor and those for whom endoscopic treatment was indicated were excluded from this analysis. The part of this cohort has been previously reported [13]. All 385 patients underwent a radical gastrectomy and lymphadenectomy with SLN mapping with their informed consent. (Table 1) The average age of the patients was 60.0 y (range, 26e86 y). Pathologic tumor size was <4.0 cm in 286 (74.3%) patients and 4.0 cm in 99 (25.7%) patients. Sixty-one (15.8%) tumors were in the upper third of the stomach, 224 (58.2%) in the middle third, and 100 (26.0%) in the lower third. Thirty-four (8.8%) patients underwent total gastrectomy, 256 (66.5%) distal gastrectomy, 44 (11.4%) proximal gastrectomy, and 51 (13.2%) underwent transected gastrectomy or pyloruspreserving gastrectomy.
2.2.
SLN mapping procedures
A double tracer method with radioactive tracer and blue dye (1% isosulfan blue (Lymphazurin; U.S. Surgical (a division of Tyco Healthcare Group LP), Norwalk, CT or indocyanine green; Daiichi Sankyo, Tokyo, Japan)) was used, as previously described [9e11]. The day before surgery, a 2.0-mL volume of technetium-99m tin colloid solution (150 Bq) was injected at four points around the primary lesion into the submucosal layer endoscopically. At the beginning of surgery, the blue dye was injected in a similar way. Intraoperative radio-labeled SLN sampling was performed using a handheld gamma probe (GPS Navigator; Sheen Man Co, Ltd., Osaka, Japan). Radio-labeled SLNs were defined as hot nodes containing ten times more radioactivity than surrounding tissue. Blue-dyed SLN (blue node) sampling was based on visual evaluation; blue nodes were identified through the afferent lymphatic flow. Finally, the hot and/or blue nodes were defined as the SLNs. Not all SLNs were picked up. SLN biopsy was followed by conventional D1 to D2 gastrectomy based on the Japanese guideline [14]. A skilful surgical team has sequentially performed all gastrectomy and SLN mapping.
2.3.
Lymph node specimens
SLNs and the other lymph nodes were examined separately. In virtually all cases, the SLN was immediately bisected during surgery, and frozen sections were examined histologically. The remaining SLN was then processed routinely for H&E staining. All lymph nodes including SLNs were examined in multiple permanent sections.
3
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 1 e Patients’ characteristics. Characteristics
SLN biopsy (n ¼ 385) No. of Patients
Mean, y Mean Two standard deviations Sex Male Female cT 1 (M or SM) 2a (MP) Location U M L Tumor size (mm; pathologic) 40> 40 pT 1 (M) 1 (SM) 2a (MP) 2b (SS) or 3 (SE) pN Negative Positive Lymphatic invasion þ Vascular invasion þ Infiltration A B C Operative methods Open surgery Laparoscopy assisted Procedure TG DG PG Transected Histologic type Differentiated Undifferentiated
%
Earlier period (n ¼ 159) No. of Patients
60.0 10.7
%
Later period (n ¼ 226) No. of Patients
58.8 10.9
P value
%
60.9 10.4
262 123
68.1 31.9
107 52
67.3 32.7
155 71
68.6 31.4
0.82
359 26
93.2 6.8
137 22
86.2 13.8
222 4
98.2 1.8
<0.05
61 224 100
15.8 58.2 26.0
29 93 37
18.2 58.5 23.3
32 131 63
14.2 58.0 27.9
0.43
286 99
74.3 25.7
119 40
74.8 25.2
167 59
73.9 26.1
0.91
209 139 22 13
54.3 36.1 5.7 3.4
83 55 13 8
52.2 34.6 8.2 5.0
126 84 9 5
55.8 37.1 4.0 2.2
0.14
349 38
90.6 9.9
136 23
85.5 14.5
211 15
93.4 6.6
<0.05
303 79
78.7 20.5
119 40
74.8 25.2
184 39
81.4 17.3
0.07
362 16
94.0 4.2
153 4
96.2 2.5
209 12
92.5 5.3
0.20
80 222 53
20.8 57.7 13.8
34 98 21
21.4 61.6 13.2
46 124 32
20.4 54.9 14.2
0.85
250 135
64.9 35.1
155 4
97.5 2.5
95 131
42.0 58.0
<0.05
34 256 44 51
8.8 66.5 11.4 13.2
23 117 16 3
14.5 73.6 10.1 1.9
11 139 28 48
4.9 61.5 12.4 21.2
<0.05
194 189
50.4 49.0
85 74
53.5 46.5
109 115
48.2 50.9
0.41
DG ¼ distal gastrectomy; Infiltration A ¼ The tumor shows expanding growth and a distinct border with the surrounding tissue; Infiltration B ¼ This category is between Infiltration Alpha and Infiltration C; Infiltration C ¼ The tumor shows infiltrating growth and an indistinct border with the surrounding tissue; L ¼ lower third of the stomach; M ¼ middle third of the stomach; M ¼ mucosa; MP ¼ muscularis propria; PG ¼ proximal gastrectomy; SE ¼ Tumor penetration of serosa; SM ¼ submucosa; SS ¼ subserosa; TG ¼ total gastrectomy; U ¼ upper third of the stomach.
In this study, micro-metastasis was defined as a metastatic focus ranging from 0.2 to 2 mm in diameter, and ITC was defined as a single tumor cell or a cluster of tumor cells <0.2 mm in diameter.
2.4.
Follow-up
All patients were surveyed after gastrectomy by performing computed tomography (CT) every 6 mo and upper
gastrointestinal endoscopy annually. The median follow-up period was 58.9 mo.
2.5.
Statistical analysis
The results of SLN biopsy were quantified using the following definitions: SLN detection rate ¼ (number of patients with successful identification of one or more SLNs)/(total number of patients with SLN biopsy procedure), false-negative
4
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
tumors (6.4%) invaded into the muscularis propria or deeper in clinical T1 cases. In 5 (19.2%) of 26 cT2 cases, the tumor invaded into the subserosa or deeper.
Table 2 e Pathologic status of SLNs (n [ 372). SLN
Nodal status from all dissected LN
Positive Negative Total
Positive
Negative
Total
34 (9.1%) 4 (1.1%) 38
0 (0%) 334 (89.8%) 334
34 338 372
The gastric cancer patients with cT1 or cT2N0M0, single lesion, and no previous treatment. The false-negative cases: 4 (1.1%). The accuracy rate based on the SLN biopsy: 98.9%.
rate ¼ false-negative/total metastatic cases, accuracy rate ¼ (number of patients with true positive and true negative patients)/(number of patients with successful identification of one or more SLNs). The success of SLN identification, accuracy, sensitivity, and specificity was compared between the earlier period (1999e2002) and the later period (2003e2007). The data are reported with 95% binomial confidence intervals (CIs); the chisquared test and the Fisher exact test were performed for comparisons of qualitative variables, whereas the Manne Whitney U test was used for comparing quantitative variables. Multivariate logistic regression analyses were used to calculate odds ratios (ORs) and their 95% CIs for undetected or falsenegative cases in SLN mapping. In survival analysis, recurrence free survival was measured from the date of surgery to the date of the first evidence of recurrence or death. Survival probabilities were estimated by KaplaneMeier method, and group differences were assessed by log-rank test. All statistical analyses were carried out using SPSS 17.0 software (SPSS Inc, Chicago, IL).
3.
Results
3.1.
Clinicopathologic background factors
3.2.
The SLN was successfully identified in 372 (96.6%, CI: 94.8%e98.4%) of the 385 patients using the dye-guided and radio-guided methods (Table 2). The mean number of identified SLNs per patient was 4.7 (range 1e24). Only one SLN was found in 38 (10.2%) patients, 3 SLNs were found in 149 (40.1%), and 4 SLNs were found in 221 (59.4%) patients. Meanwhile, in patients who underwent laparoscopic-assisted gastrectomy, SLNs were identified in 131 (97.0%) of 135 patients. The mean number of SLNs was 5.0 in patients undergoing open surgery and 4.2 in patients undergoing laparoscopic-assisted surgery. The mean number of detected SLNs was not significantly different between open surgery and laparoscopy-assisted surgery. Thirty-four (89.5%) of 38 cases with lymph node metastases showed positive SLNs. Eleven patients had positive SLN and positive non-SLN (SLN(þ)/non-SLN(þ)), and 23 patients had positive SLN and negative non-SLN (SLN(þ)/non-SLN()). The accuracy of SLN biopsy for the detection of lymph node metastases was 98.9% (368 of 372).
3.3.
Clinicopathologic features of false-negative cases
In four of the 38 patients, SLN examination revealed a falsenegative diagnosis. The primary tumor was <4 cm in diameter in all four false-negative cases (Table 3). Three of these tumors invaded into the muscularis propria or deeper on pathologic examination. Three of these cases had metastases in non-SLN within the lymphatic area of the SLN (the sentinel basin). Only one case (case 3) had a metastatic lymph node in the node along the lesser curvature (#3), although the SLN was in the nodes along the common hepatic artery (#8a). This tumor invaded to the subserosa on pathologic examination.
3.4. A total of 385 consecutive patients were enrolled in this study. The clinicopathologic characteristics of the patients enrolled in this study are summarized in Table 1. On pathology, 23
SLN identification
Results in the earlier and later periods
The SLN detection rate was 93.7% in the earlier period (1999e2002) and 98.8% in the later period (2003e2007; Table 4).
Table 3 e False-negative cases. Case
Age
Sex
SLN#*
1
57
F
2 3
55 52
F M
3, 4d, 6, 7, 11p 1, 3 8a
4
63
M
3, 5, 7
Non-SLN met#y
cT
4d, 11p
1
3 3 3
Location
Location
Type
Size (mm)
Histology
Depth
Recurrence
M
Greater
0-IIc
30
Undiff.
MP
d
1 2
U M
Less Less
0-IIc 3
30 25
Undiff. Diff.
SM SS
1
M
Less
0-IIc
29
Diff.
MP
d Liver, po 1Y3M d
(U/M/L)
diff. ¼ differentiated type; L ¼ lower third of the stomach; M ¼ middle third of the stomach; MP ¼ muscularis propria; SM ¼ submucosa; SS ¼ subserosa; U ¼ upper third of the stomach; undiff. ¼ undifferentiated type. * SLN# ¼ station number of the SLN. y Non-SLN met# ¼ station number of the metastatic non-SLN (the Japanese Gastric Cancer Association, Japanese Classification of Gastric Carcinoma, second English Edition [14]).
5
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 4 e Results in the earlier period (1999e2002) and the later period (2003e2007).
SLN
Earlier period (159 patients)
Later period (226 patients)
The detection rate: 149 of159 (93.7%)
The detection rate: 223 of 226 (98.8%)
Nodal status from all dissected LN Positive
Negative
Positive 20 (13.4%) 0 (0%) Negative 3 (2.0%) 126 (84.6%) Total 23 126 False-negative cases: 3 (2.0%) The accuracy rate based on the SLN biopsy: 98.0%
20 129 149
Tumor depth
The accuracy rates based on the SLN biopsy by cT1 and cT2 are listed in Table 5. The false-negative rate was 10.7% and 10.0% in cT1 and cT2 cases, respectively. The accuracy rate based on the SLN biopsy was 99.2% in cases with cT1 and 96.0% in cases with cT2, which was not significantly different between the two groups (P ¼ 0.24). Meanwhile, the detection rate was 96.6% in 348 cases with pT1 and 97.1% in 35 cases with pT2 or deeper. The accuracy rate based on the SLN biopsy was 99.7% in cases with pT1 and 91.2% in cases with pT2 or deeper, and the accuracy rate was significantly better in pT1 cases than pT2 or deeper cases (P < 0.01).
3.6.
Nodal status from all dissected LN
Total
The SLN detection rates were significantly better in the later period than in the earlier period (P < 0.01; Table 4). Clinical tumor invasion (cT), pathological lymph node metastasis (pN), and operative method were significantly different between the two periods (Table 1). The histological type and tumor size were not significantly different. There were three falsenegative cases (2.0%) in the earlier period and only one case (0.4%) in the later period. The percentage of the undetected and false-negative cases each year was shown in Figure 1. In particular, most of the falsenegative cases were concentrated in the initial 2-y period.
3.5.
SLN
Intraoperative diagnosis of SLN metastases
In 5 (13.2%) of 38 node-positive cases, all SLNs had no metastases on intraoperative rapid diagnosis. However, seven
Positive
Positive 14 (6.3%) 0 (0%) Negative 1 (0.4%) 208 (93.3%) Total 15 208 False-negative cases: 1 (0.4%) The accuracy rate based on the SLN biopsy: 99.6%
Total 14 214 223
SLNs in the five cases had micrometastases, not macrometastases, on histopathologic diagnosis of the permanent sections. Three cases were ultimately evaluated for the presence of isolated tumor cells (ITCs). Remarkably, non-SLNs beyond the sentinel basin had no metastases in these cases with micrometastasis and/or ITCs.
3.7. The relationship between the location of the primary lesion and the SLNs The station numbers of the SLNs and the frequencies by tumor site are listed in Tables 6 and 7. The tumors in the upper third (U area) had SLNs at LNs along the left gastric artery (#3, #1, and #7) in many cases and few or no SLNs at the suprapyloric LN (#5), infrapyloric LN (#6), and the LN along the common hepatic artery (#8a). The tumors in the middle third (M area) had SLNs at the right paracardial LN and along the lesser curvature and left gastric and gastroepiploic arteries. The tumors in the lower third (L area) had SLNs at #6 and LNs along the right gastroepiploic artery (#4d) with high frequency, and SLNs along the superior mesenteric vein (#14v) with low frequency, but one had a positive SLN at #14v. The tumors in L area had no SLNs at the left paracardial LN (#2) and the LN along the short gastric artery (#4sa).
3.8. Predictors for undetected or false-negative cases on SLN mapping Multivariate logistic regression analyses identified the factors associated with undetected or false-negative cases. Candidate variables included location, size, and depth of the tumor, tumor differentiation, lymphatic invasion, venous invasion, operative method (open surgery versus laparoscopic surgery), sex, and period (earlier period versus later period). The best predictor for undetected or false-negative cases on SLN mapping was the period (odds ratio ¼ 0.056; 95% CI ¼ 0.01e0.42), as listed in Table 8.
3.9.
Fig. 1 e Changes in percentage of the undetected cases and false-negative cases.
Negative
Survival
The 5-y recurrence free survival (RFS) rates were 96.2% in all patients. Moreover, 5y-RFS were 99.0% in negative SLN (SLN()) cases compared with 86.3% in positive SLN (SLN(þ)) cases. SLN(þ) cases had a worse prognosis than SLN() cases (P ¼ 0.008; Fig. 2). In SLN(þ) cases, RFS was not significantly different between the negative non-SLN cases
6
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 5 e Results as per tumor depth and size. cT1 (365 patients)
cT2 (26 patients)
The detection rate: 353/365 (96.7%)
The detection rate: 25/26 (96.1%)
SLN
Non-SLN Positive
Negative
Positive 8 (2.3%) 17 (4.8%) Negative 3 (0.8%) 325 (92.1%) Total 11 342 False-negative cases: 3 (0.8%) The accuracy rate based on SLN biopsy: 99.2% pT1 (354 patients)
SLN Total
Positive
25 328 353
Positive 3 (12.0%) 6 (24.0%) Negative 1 (4.0%) 15 (60.0%) Total 4 21 False-negative cases: 1 (4.0%) The accuracy rate based on SLN biopsy: 96.0% pT2 and pT3 (35 patients)
The detection rate: 342/354 (96.6%) SLN
Negative
Positive 6 (1.8%) 14 (4.1%) Negative 1 (0.3%) 321 (93.9%) Total 7 335 False-negative cases: 1 (0.3%) The accuracy rate based on SLN biopsy: 99.7%
Negative
Total 9 16 25
The detection rate: 34/35 (97.1%)
Non-SLN Positive
Non-SLN
SLN
Non-SLN
Total
Positive
Negative
20 322 342
Positive 5 (14.3%) 9 (26.5%) Negative 3 (8.8%) 17 (50.0%) Total 8 26 False-negative cases: 3 (8.8%) The accuracy rate based on SLN biopsy: 91.2%
Total 14 20 34
cT1 ¼ clinically tumor invaded lamina propria and submucosa; cT2 ¼ clinically invaded muscularis propria; pT1 ¼ pathologically tumor invaded lamina propria and submucosa; pT2 ¼ pathologically invaded muscularis propria; pT3 ¼ pathologically tumor penetrated serosa without invasion of adjacent structures.
(SLN(þ)/non-SLN()) and positive non-SLN cases (SLN(þ)/ non-SLN(þ)) (P ¼ 0.511; Fig. 3).
Table 6 e Results as per tumor site. Upper third area (61 patients) Detection rate: 58/61 (95.1%) SLN
4.
Discussion
Non-SLN Positive
Negative
Total
Positive 1 (1.7%) 5 (8.6%) 6 Negative 1 (1.7%) 51 (87.9%) 52 Total 2 56 58 False-negative cases: 1 (1.7%) The specificity true positive rate based on SLN biopsy: 98.3% Middle third area (230 patients) The detection rate: 227/230 (98.7%) SLN
Non-SLN Positive
Negative
Total
Positive 6 (2.6%) 14 (6.2%) 20 Negative 3 (1.3%) 204 (89.9%) 207 Total 9 218 227 False-negative cases: 3 (1.3%) The specificity true positive rate based on SLN biopsy: 98.7% Lower third area (100 patients) The detection rate: 93/100 (93.0%) SLN
Non-SLN Positive
Negative
Total
Positive 4 (4.3%) 4 (4.3%) 8 Negative 0 (0%) 85 (91.4%) 85 Total 4 89 93 False-negative cases: 0 (0%) The specificity true positive rate based on SLN biopsy: 100%
We previously developed SLN mapping using radio-guided and dye-guided methods in gastric cancer and reported the procedure for the first time [6e8]. Similar results concerning the feasibility of the SLN concept in gastric cancer have been reported in published studies [9e12,15]. The present report is the largest single institutional report of SLN in gastric cancer to the best of our knowledge. The SLN concept has been modified for clinical practice in several institutions [10e12,16]. The eligibility criteria for the SLN concept were cT1N0M0 or cT2N0M0, 4 cm tumor, and single tumor gastric cancer patients in most reports [9,12]. However, before introducing SLN biopsy in gastric cancer, we must be concerned about the learning curve of the SLN mapping technique and the indications with respect to tumor depth in particular. SLN biopsy was performed for 385 untreated cT1N0M0 or cT2N0M0 gastric cancer patients from 1999 to 2007. For all cases, the accuracy rate was 98.9%. The result was a match even for melanoma or breast cancer cases, in that the SLN concept has been shown to be generally applicable clinically. Furthermore, false-negative cases were seen in four patients in whom the SLNs were detected intraoperatively. From the perspective of tumor depth, the tumors penetrated the muscularis propria or deeper pathologically in three of the four false-negative cases. Given this, SLN mapping has to be adapted for gastric cancer patients whose tumor penetrates the mucosa or submucosa. The present study showed that the
7
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 7 e Location of the SLN by tumor site. U (n ¼ 58)
Station No.
1 2 3 4sa 4sb 4d 5 6 7 8a 9 11p 11d 12a 14v
M (n ¼ 221)
L (n ¼ 93)
SLN
Metastasis
SLN
Metastasis
SLN
Metastasis
19 (33) 5 (9) 44 (76) 3 (5) 3 (5) 2 (3) 2 (3) d 17 (29) d 1 (2) 2 (3) 2 (3) d d
1 (2) 1 (2) 1 (2) 0 0 0 0 d 1 (2) d 0 0 1 (2) d d
40 (18) 2 (1) 160 (72) 2 (1) 22 (10) 79 (36) 14 (6) 32 (14) 70 (32) 18 (8) 2 (1) 2 (1) 1 (0) d d
3 (1) 0 10 (5) 1 (0) 1 (0) 5 (2) 0 0 6 (3) 0 0 0 0 d d
5 (5) d 39 (42) d 1 (1) 31 (33) 17 (18) 52 (56) 26 (28) 12 (13) 2 (2) d d 1 (1) 3 (3)
0 d 4 (4) d 0 1 (1) 0 2 (2) 2 (2) 1 (1) 0 d d 0 1 (1)
The station number is described according to the Japanese Gastric Cancer Association, Japanese Classification of Gastric Carcinoma, second English Edition [14]. No. 1 ¼ right paracardial lymph node (LN); No. 2 ¼ Left paracardial LN; No. 3 ¼ LN along the lesser curvature; No. 4sa ¼ LN along the short gastric vessels; No. 4sb ¼ LN along the left gastroepiploic vessels; No. 4d ¼ LN along the right gastroepiploic vessels; No. 5 ¼ suprapyloric LN; No. 6 ¼ infrapyloric LN; No. 7 ¼ LN along the left gastric artery; No. 8a ¼ LN along the common hepatic artery (anterosuperior group); No. 9 ¼ LN around the celiac artery; No. 11p ¼ LN along the proximal splenic artery; No. 11d ¼ LN along the distal splenic artery; No. 12a ¼ LN in the hepatoduodenal ligament (along the hepatic artery); No. 14v ¼ LN along the superior mesenteric vein. L ¼ lower third of the stomach; M ¼ middle third of the stomach; No. ¼ station number; U ¼ upper third of the stomach.
accuracy rate based on the SLN biopsy was markedly worse in pathologic T2 or deeper cases. There has been a report that the probability of falsenegative SLN increases with increasing tumor size in breast cancer [17]. In our study, the accuracy rate based on the SLN biopsy did not show a significant difference in gastric cancer patients with tumor size 4 cm. However, tumor size was a simple prognostic indicator. The gastric cancer patients with tumor 4 cm had a significantly worse prognosis as compared with tumor <4 cm [18]. Moreover, as the size of the tumor is >4 cm, it is more difficult to perform the standard limited gastrectomy. A size of <4 cm, approximately, might be appropriate for SLN mapping in gastric cancer. We should do accurate preoperative evaluations of tumor depth in particular. Several pathologic T2 or deeper tumors
might be contained within clinical T1 cases. As a result, the false-negative cases might increase. On detailed analysis of the false-negative cases, there was a relationship between SLNs and positive lymph nodes pathologically. The positive lymph nodes stayed within the sentinel basin in all but one case. If SLNs are detected in gastric cancer for virtually all cases, sentinel basin dissection guards against the possibility of leaving positive lymph nodes. There are many contrasting views about the sentinel basin. However, considering that tumor cells might flow to the distant LNs through the close SLNs and that the SLNs might be misidentified because of tracer being prevented from reaching the LNs for any reason, basin dissection should be regarded as a useful and important part of the safety net.
Table 8 e Results of multivariate logistic regression analysis of predictors for undetected or false-negative cases with SLN mapping. Variable
Odds ratio
95% CI
P value
Location of tumor Size of tumor Pathologic tumor depth Tumor differentiation Lymphatic invasion Venous invasion Operative methods (open versus laparoscopic) Gender Periods (the earlier versus the later)
1.96 0.97 3.42 0.51 0.95 3.91 6.32
0.76e5.08 0.93e1.02 0.67e17.42 0.16e1.65 0.24e3.79 0.53e28.81 0.95e42.11
0.166 0.269 0.140 0.258 0.942 0.181 0.057
0.99 0.056
0.28e3.42 0.01e0.42
0.981 0.005
Fig. 2 e Recurrence-free survival in SLN(L) versus SLN(D). (Color version of figure is available online.)
8
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Fig. 3 e Recurrence-free survival in SLN(D)/non-SLN(L) versus SLN(D)/non-SLN(D). (Color version of figure is available online.)
When comparing the results between the earlier period and the later period in the present study, we should point out the problem of the learning curve in the SLN mapping technique. Relatively few studies have reported the learning curve on SLN mapping on gastric cancer. The results of SLN mapping in gastric cancer in this study might improve with progress of the technique. In breast cancer, SLN mapping should be performed by surgeons with appropriate training and experience [19,20]. As a measure of whether to proceed with SLN biopsy, a detection rate of at least 95% has been recommended [21]. When SLN biopsy in gastric cancer would be used routinely in clinical care, we have to focus more attention on the stabilization or learning curve of the technique as with breast cancer. In our study, the percentage of the undetected and false-negative cases tended to be reduced gradually year by year (Fig. 1). The undetected cases in 2005 were increased for unaccountable reason. When SLN was not detected, the conventional lymphadenectomy had to be performed for these patients. The present results indicate that undetected or falsenegative cases with SLN mapping were significantly associated only with the time period (Table 8). This strongly suggests that stabilization of the procedure and experience with SLN mapping in gastric cancer might be associated with a decrease in undetected or false-negative cases. From another point of view, there was variability among the preoperative clinical features between the two periods. That essential differences as the diagnosis of tumor depth and lymph node metastases suggested improvements of the SLN detection rate. To explain the differences by time period, numerous factors may have been involved, such as the development of pretreatment diagnostic capability and clarification of the indications for SLN mapping in gastric cancer. In 226 patients with cT1 or cT2N0M0 in the present study, a single lesion, and no previous treatment in the later period, the detection rate was 98.7% (223 of 226), the accuracy rate based on the SLN biopsy was 99.6% (222 of 223), and there was only one false-negative patient. Although there is no standard of SLN biopsy in gastric cancer, surgeons should not abandon radical lymph node dissection until a certain level of results can be achieved consistently.
There is a limitation of intraoperative rapid diagnosis. A discrepancy between the intraoperative SLN diagnosis and the final diagnosis was frequent, with eight cases (five micrometastasis and three ITC cases) in the present study. No cases had macrometastases. The most subtle histologic subclass of lymph node micrometastasis and macrometastasis has been defined as lymph nodes containing metastatic foci measuring number >2 mm in diameter [21]. These micrometastasis and/or ITC cases, furthermore, had no lymph node metastases beyond the sentinel basin. These results support the idea that sentinel basin dissection might be enough even in cases with intraoperatively negative SLNs to guarantee curability. The location of SLNs by tumor site is notable. The tumors in U area had few or no SLNs at the suprapyloric LN (#5), infrapyloric (#6) LN, and the LN along the common hepatic artery (#8) and had positive SLNs along the splenic artery (#11) with low frequency. When we select proximal or total gastrectomy for tumors in the upper gastric area, if the location of the SLN is limited in the paracardiac LN or along the left gastric and splenic arteries, the lower third of the stomach and LNs around #5, #6, and #8 might be safely preserved. Moreover, the LN at #11 should be considered for removal. Tumors in the middle third gastric area had lymphatic flow into right paracardial LN (#1) and along the lesser curvature (#3), the left gastric artery (#7), and the gastroepiploic artery (#4d, #4sb). Transected gastrectomy might be selected for a tumor in the middle third gastric area. The tumors in the lower third of the stomach had lymphatic flow to the hepatic artery (#8a, #12) or superior mesenteric vein (#14v), and one had metastasis at #14v. The LNs at #14v might not always be dissected for all patients with tumor in L area. However, if the tumors have SLNs at #14v, dissection is worthy of consideration. At any rate, we have to take into account the probable consequences with respect to curability and invasive potential. In survival analysis, 5y-RFS in SLN() cases was significantly better than that in SLN(þ) cases. This suggested that SLN status was a predictive marker for recurrence as in usual pathologic nodal status. Originally, we speculated that the poor prognosis of SLN(þ) cases was guided by the outcomes of SLN(þ)/ non-SLN(þ) cases. However, the prognosis of SLN(þ)/nonSLN(þ) cases was the same as that for SLN(þ)/non-SLN() cases by ensuring that SLN biopsy was followed by conventional lymphadenectomy. Therefore, if SLN was positive, conventional lymphadenectomy has to be performed. In the present study, SLN mapping for gastric cancer with cT1N0M0, diameter of primary tumor 4 cm, single lesion, and no previous treatment was considered acceptable in terms of the detection rate and the accuracy in determining SLN status. Furthermore, sentinel basin dissection is recommended for these patients to maintain safety [22]. The individualized, minimally invasive gastrectomy with selective and modified lymphadenectomy based on the SLN concept will lead to the improvement of long-term quality of life in the future.
Acknowledgment Y.K. made a contribution to conception and design. Y.S., T.T., N.W. made contributions to surgery and collecting data. M.N.
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
made a contribution to managing the radioactive isotope tracer. M.M. made a contribution to pathology diagnosis. H.T. made a contribution to approving the final version of the article.
Disclosure Professor Kitagawa has grants or payment for lectures from Ajinomoto Pharmaceuticals Co., ltd., Asahi Kasei Co., Ltd, Astellas Pharma Inc., Otsuka Pharmaceutical Factory, Inc., Otsuka Pharmaceutical Co., Ltd, ONO Pharmaceutical, Co., Ltd, Kaken Pharmaceutical, Co., Ltd, Kyowa Hakko Kirin Co., Ltd, GlaxoSmithKline Pharmaceuticals Ltd., LSE, Kureha Co., Kowa Pharmaceuticals, Co., Ltd, Shionogi & Co., Ltd, Sanofi K.K., CSL Behling KK, Taiho Pharmaceuticals, Co., Ltd, Chugai Pharmaceuticals, Co., Ltd, Takeda Pharmaceutical, Co., Ltd, Daiichi Sankyo Co., Ltd, TSUMURA & CO., Teijin Pharma Limited, Dainippon Sumitomo Pharma, Co., Ltd, Taisho Pharmaceuticals, Co., Ltd, Novartis Pharma, Co., Ltd, Nippon Kayaku Co., Ltd, Medtronic Japan Co., Ltd, Boehlinger Ingelheim GmbH, Pfizer Co., Ltd, Merck Serono Co., Ltd, Yakult Honsha Co., Ltd, Terumo Co., Torii Pharmaceutical, Co., Ltd, Astra Zeneca Co., Ltd, Ethicon Inc. (Part of Johnson & Johnson Family of Companies), and Olympus Co.
references
[8] [9]
[10] [11]
[12]
[13]
[14]
[15]
[16]
[17] [1] Maruyama K, Gunve0 n P, Okabayashi K, et al. Lymph node metastases of gastric cancer. Ann Surg 1989;210:596. [2] Songun I, Putter H, Kranenberg EM, et al. Surgical treatment of gastric cancer: 15-year follow-up results of the randomized nationwide Dutch D1D2 trial. Lancet Oncol 2010; 11:439. [3] Akagi T, Shiraishi N, Kitano S. Lymph node metastasis of gastric cancer. Cancer 2011;3:2141. [4] Folli S, Dente M, Dell’ Amore D, et al. Early gastric cancer: prognostic factors in 223 patients. Br J Surg 1995; 82:952. [5] Morton DL, Wen DR, Wong JH, et al. Technical details of intraoperatively lymphatic mapping for early stage melanoma. Arch Surg 1992;127:392. [6] Tonouchi H, Mohri Y, Tanaka K, et al. Laparoscopic lymphatic mapping and sentinel node biopsies for earlystage gastric cancer: the cause of false negativity. World J Surg 2005;29:418. [7] Uenosono Y, Natsugoe S, Ehi K, et al. Detection of sentinel nodes and micrometastases using radioisotope navigation
[18]
[19]
[20]
[21]
[22]
9
and immunohistochemistry in patients with gastric cancer. Br J Surg 2005;92:886. Park DJ, Lee HJ, Lee HS, et al. Sentinel node biopsy for cT1 and cT2a gastric cancer. Eur J Surg Oncol 2006;32:48. Kitagawa Y, Fujii H, Kitajima M, et al. The role of the sentinel lymph node in gastrointestinal cancer. Surg Clin North Am 2000;80:1799. Kitagawa Y, Kitajima M. Gastrointestinal cancer and sentinel node navigation surgery. J Surg Oncol 2002;79:188. Ichikura T, Sugasawa H, Ono S, et al. Limited gastrectomy with dissection of sentinel node stations for early gastric cancer with negative sentinel node biopsy. Ann Surg 2009; 249:942. Takeuchi H, Oyama T, Kitagawa Y, et al. Laparoscopyassisted proximal gastrectomy with sentinel node mapping for early gastric cancer. World J Surg 2011;35: 2463. Kitagawa Y, Takeuchi H, Takagi Y, et al. Sentinel node mapping for gastric Cancer: a prospective multicenter trial in Japan. J Clin Oncol 2013;31:3704. Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma -2nd English edition-. Gastric Cancer 1998;1:10. Ando N, Ogawa S, Kitagawa Y, et al. Improvement in the results of surgical treatment of advanced squamous esophageal carcinoma during 15 consecutive years. Ann Surg 2000;232:225. Takeuchi H, Fujii H, Kitagawa Y, et al. Validation study of radio-guided sentinel lymph node navigation in esophageal cancer. Ann Surg 2009;249:757. Coombs N, Chen W, Boyages J, et al. A decision tool for predicting sentinel node accuracy from breast tumor size and grade. Breast J 2007;13:593. Adachi Y, Oshiro T, Mori M, et al. Tumor size as a simple prognostic indicator for gastric carcinoma. Ann Surg Oncol 1997;4:137. Bass SS, Cox CE, Reintgen DS, et al. Learning curves and certification for breast cancer lymphatic mapping. Surg Oncol Clin N Am 1999;8:497. Schwartz GF, Guiliano AE, Veronesi U. Consensus Conference Committee. Proceeding of the consensus conference of the role of sentinel lymph node biopsy in carcinoma or the breast April 19-22, 2001, Philadelphia, PA, USA. Breast J 2002; 8:124. Huvos AG, Hutter RVP, Berg JW. Significance of axillary macrometastases and micrometastases in mammary cancer. Ann Surg 1971;173:44. Takeuchi H, Kitagawa Y. New sentinel node mapping technologies for early gastric cancer. Ann Surg Oncol 2013; 20:522.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Sentinel lymph node mapping for 385 gastric cancer patients Masahiro Niihara, MD,a Hiroya Takeuchi, MD, PhD,a,* Tadaki Nakahara, MD,b Yoshiro Saikawa, MD, PhD,a Tsunehiro Takahashi, MD, PhD,a Norihito Wada, MD, PhD,a Makio Mukai, MD, PhD,c and Yuko Kitagawa, MD, PhDa a
Department of Surgery, School of Medicine, Keio University, Tokyo, Japan Department of Radiology, School of Medicine, Keio University, Tokyo, Japan c Division of Diagnostic Pathology, School of Medicine, Keio University, Tokyo, Japan b
article info
abstract
Article history:
Background: The objectives were to investigate the accuracy of sentinel lymph node (SLN)
Received 19 February 2015
biopsy, detect the predictors for undetected or false-negative cases, evaluate the in-
Received in revised form
dications for SLN-navigated gastrectomy, and characterize the problems of SLN mapping in
25 June 2015
gastric cancer. The SLN concept may be applicable to early gastric cancer, particularly
Accepted 26 June 2015
clinical T1N0M0 or T2N0M0 with tumor diameter 4 cm.
Available online xxx
Methods: A total of 385 consecutive patients diagnosed with cT1N0M0 or cT2aN0M0 operable gastric cancer from April 1999 to December 2007 underwent radical gastrectomy with
Keywords:
SLN mapping. SLNs were identified using radio-guided and dye-guided methods. Predictors
Learning curve
for undetected or false-negative cases on SLN mapping were examined by multivariate
Sentinel basin
regression analysis.
Sentinel navigation surgery
Results: The detection rate of hot and/or blue nodes was 96.6% (372 of 385). The accuracy of
Long-term survival
metastatic status based on SLNs was 98.9% (368 of 372) for all cases in whom SLNs could be
Stabilization
detected. Furthermore, the accuracy of metastatic status based on SLNs was 99.1% (344 of 347)
Predictor
in cT1 gastric cancer and 96.0% (24 of 25) in cT2 gastric cancer. Pathologically, the tumors invaded to the muscularis propria or deeper in three of four false-negative cases. All but one case had metastatic lymph nodes within the sentinel basins. In terms of 5-y recurrence free survival, positive SLN cases (SLN(þ)) had a worse prognosis than negative SLN cases (SLN(); P ¼ 0.008). Moreover, SLN(þ) and non-SLN() cases (SLN(þ)/non-SLN()) had a similar prognosis as SLN(þ) and non-SLN(þ) cases (SLN(þ)/non-SLN(þ)) (P ¼ 0.511). On multivariate regression analysis, undetected or false-negative cases were significantly associated with the time period. Conclusions: The present results appeared to validate the SLN concept for untreated cT1 gastric cancer with tumor diameter 4 cm. SLN mapping may provide an effective method of staging the lymph node status of patients undergoing minimized gastrectomy. Sentinel basin dissection guards against the possibility of leaving positive lymph nodes. Stabilization of the procedure and experience with SLN mapping in gastric cancer might decrease undetected or false-negative cases. ª 2015 Elsevier Inc. All rights reserved.
* Corresponding author. Department of Surgery, School of Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. Tel.: þ81 3 3353 1211x62334; fax: þ81 3 3355 4707. E-mail address: [email protected] (H. Takeuchi). 0022-4804/$ e see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2015.06.064
2
1.
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Introduction
In the past decades, radical resection with lymphadenectomy for gastric cancer improved locoregional disease control and also improved survival in Far East [1]. The necessary extent of lymph node dissection has been heavily debated and controversial. Recently in Western countries also, D2 lymphadenectomy is recommended as the appropriate surgical approach for patients with resectable gastric cancer because of its long-term survival benefit. However, the routine D2 procedure has also been associated with significantly higher postoperative mortality, morbidity, and reoperation rates [2]. Meanwhile, the popularization of medical examinations and improvements in diagnostic technology make it possible to detect gastric cancer at earlier stages. Many of these patients have been cured by surgical treatment. Recently, laparoscopic-assisted gastrectomy has been improved and standardized. Therefore, the short-term quality of life of these patients is now considered to be better than before. However, such patients have suffered from the sequelae of gastrectomy, such as dumping syndrome, postvagotomy symptoms, and so on, over extended periods, even though some cases had no pathological lymph node metastases. The rate of lymph node metastasis reaches 10% in pathological mucosal gastric carcinoma, 25% in submucosal tumor, 45% in muscularis propria tumor, and 70% in subserosa tumor [3,4]. In fact, many resectable early-stage gastric cancers have no lymph node metastases. Accurate assessment of node status is mandatory to avoid unnecessary extended lymphadenectomy in nodenegative patients who may not get the full benefit of the standard D2 dissection. Given these problems, an individualized surgical procedure is needed, taking into account both curability and after effects, instead of a uniform method. The sentinel lymph node (SLN) concept may make it possible to select the appropriate extent of lymph node dissection to avoid unnecessary lymphadenectomy in node-negative patients. The SLN first receives lymphatic drainage from the tumor site [5]. The SLN concept that the SLNs are the possible site of first metastasis via the lymphatic route and that the pathological status of the SLN can predict metastases to all other downstream lymph nodes, has been evaluated in various solid tumors. We previously developed SLN mapping in gastric cancer and reported the procedure for the first time [6e9]. Some investigators demonstrated that the SLN concept can be applied in patients with early gastric cancer, in particular clinical T1N0M0 or T2N0M0 with a tumor diameter of 4 cm [10e12]. The present study is the largest single institutional report of SLN mapping in gastric cancer, to the best of our knowledge. The primary aim of the present work was to investigate the accuracy of the SLN of gastric cancer in our institution and detect the predictors for undetected or false-negative cases, focusing on the indication for SLN-navigated gastrectomy. These results supported the fundamental direction of sentinel node navigation surgery in gastric cancer, which involves functional preservation and is a minimally invasive surgical technique.
2.
Materials and methods
2.1.
Patients
All patients diagnosed with cT1N0M0 or cT2N0M0 operable primary gastric cancer at Keio University Hospital (Tokyo, Japan) during the period from April 1999 to December 2007 were enrolled in this study. Criteria for eligibility were, in principle, clinical T1N0M0 or T2aN0M0 gastric cancer according to the Union for International Cancer Control (sixth edition, 2002). The patients who obviously had surgical T3 tumor and those for whom endoscopic treatment was indicated were excluded from this analysis. The part of this cohort has been previously reported [13]. All 385 patients underwent a radical gastrectomy and lymphadenectomy with SLN mapping with their informed consent. (Table 1) The average age of the patients was 60.0 y (range, 26e86 y). Pathologic tumor size was <4.0 cm in 286 (74.3%) patients and 4.0 cm in 99 (25.7%) patients. Sixty-one (15.8%) tumors were in the upper third of the stomach, 224 (58.2%) in the middle third, and 100 (26.0%) in the lower third. Thirty-four (8.8%) patients underwent total gastrectomy, 256 (66.5%) distal gastrectomy, 44 (11.4%) proximal gastrectomy, and 51 (13.2%) underwent transected gastrectomy or pyloruspreserving gastrectomy.
2.2.
SLN mapping procedures
A double tracer method with radioactive tracer and blue dye (1% isosulfan blue (Lymphazurin; U.S. Surgical (a division of Tyco Healthcare Group LP), Norwalk, CT or indocyanine green; Daiichi Sankyo, Tokyo, Japan)) was used, as previously described [9e11]. The day before surgery, a 2.0-mL volume of technetium-99m tin colloid solution (150 Bq) was injected at four points around the primary lesion into the submucosal layer endoscopically. At the beginning of surgery, the blue dye was injected in a similar way. Intraoperative radio-labeled SLN sampling was performed using a handheld gamma probe (GPS Navigator; Sheen Man Co, Ltd., Osaka, Japan). Radio-labeled SLNs were defined as hot nodes containing ten times more radioactivity than surrounding tissue. Blue-dyed SLN (blue node) sampling was based on visual evaluation; blue nodes were identified through the afferent lymphatic flow. Finally, the hot and/or blue nodes were defined as the SLNs. Not all SLNs were picked up. SLN biopsy was followed by conventional D1 to D2 gastrectomy based on the Japanese guideline [14]. A skilful surgical team has sequentially performed all gastrectomy and SLN mapping.
2.3.
Lymph node specimens
SLNs and the other lymph nodes were examined separately. In virtually all cases, the SLN was immediately bisected during surgery, and frozen sections were examined histologically. The remaining SLN was then processed routinely for H&E staining. All lymph nodes including SLNs were examined in multiple permanent sections.
3
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 1 e Patients’ characteristics. Characteristics
SLN biopsy (n ¼ 385) No. of Patients
Mean, y Mean Two standard deviations Sex Male Female cT 1 (M or SM) 2a (MP) Location U M L Tumor size (mm; pathologic) 40> 40 pT 1 (M) 1 (SM) 2a (MP) 2b (SS) or 3 (SE) pN Negative Positive Lymphatic invasion þ Vascular invasion þ Infiltration A B C Operative methods Open surgery Laparoscopy assisted Procedure TG DG PG Transected Histologic type Differentiated Undifferentiated
%
Earlier period (n ¼ 159) No. of Patients
60.0 10.7
%
Later period (n ¼ 226) No. of Patients
58.8 10.9
P value
%
60.9 10.4
262 123
68.1 31.9
107 52
67.3 32.7
155 71
68.6 31.4
0.82
359 26
93.2 6.8
137 22
86.2 13.8
222 4
98.2 1.8
<0.05
61 224 100
15.8 58.2 26.0
29 93 37
18.2 58.5 23.3
32 131 63
14.2 58.0 27.9
0.43
286 99
74.3 25.7
119 40
74.8 25.2
167 59
73.9 26.1
0.91
209 139 22 13
54.3 36.1 5.7 3.4
83 55 13 8
52.2 34.6 8.2 5.0
126 84 9 5
55.8 37.1 4.0 2.2
0.14
349 38
90.6 9.9
136 23
85.5 14.5
211 15
93.4 6.6
<0.05
303 79
78.7 20.5
119 40
74.8 25.2
184 39
81.4 17.3
0.07
362 16
94.0 4.2
153 4
96.2 2.5
209 12
92.5 5.3
0.20
80 222 53
20.8 57.7 13.8
34 98 21
21.4 61.6 13.2
46 124 32
20.4 54.9 14.2
0.85
250 135
64.9 35.1
155 4
97.5 2.5
95 131
42.0 58.0
<0.05
34 256 44 51
8.8 66.5 11.4 13.2
23 117 16 3
14.5 73.6 10.1 1.9
11 139 28 48
4.9 61.5 12.4 21.2
<0.05
194 189
50.4 49.0
85 74
53.5 46.5
109 115
48.2 50.9
0.41
DG ¼ distal gastrectomy; Infiltration A ¼ The tumor shows expanding growth and a distinct border with the surrounding tissue; Infiltration B ¼ This category is between Infiltration Alpha and Infiltration C; Infiltration C ¼ The tumor shows infiltrating growth and an indistinct border with the surrounding tissue; L ¼ lower third of the stomach; M ¼ middle third of the stomach; M ¼ mucosa; MP ¼ muscularis propria; PG ¼ proximal gastrectomy; SE ¼ Tumor penetration of serosa; SM ¼ submucosa; SS ¼ subserosa; TG ¼ total gastrectomy; U ¼ upper third of the stomach.
In this study, micro-metastasis was defined as a metastatic focus ranging from 0.2 to 2 mm in diameter, and ITC was defined as a single tumor cell or a cluster of tumor cells <0.2 mm in diameter.
2.4.
Follow-up
All patients were surveyed after gastrectomy by performing computed tomography (CT) every 6 mo and upper
gastrointestinal endoscopy annually. The median follow-up period was 58.9 mo.
2.5.
Statistical analysis
The results of SLN biopsy were quantified using the following definitions: SLN detection rate ¼ (number of patients with successful identification of one or more SLNs)/(total number of patients with SLN biopsy procedure), false-negative
4
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
tumors (6.4%) invaded into the muscularis propria or deeper in clinical T1 cases. In 5 (19.2%) of 26 cT2 cases, the tumor invaded into the subserosa or deeper.
Table 2 e Pathologic status of SLNs (n [ 372). SLN
Nodal status from all dissected LN
Positive Negative Total
Positive
Negative
Total
34 (9.1%) 4 (1.1%) 38
0 (0%) 334 (89.8%) 334
34 338 372
The gastric cancer patients with cT1 or cT2N0M0, single lesion, and no previous treatment. The false-negative cases: 4 (1.1%). The accuracy rate based on the SLN biopsy: 98.9%.
rate ¼ false-negative/total metastatic cases, accuracy rate ¼ (number of patients with true positive and true negative patients)/(number of patients with successful identification of one or more SLNs). The success of SLN identification, accuracy, sensitivity, and specificity was compared between the earlier period (1999e2002) and the later period (2003e2007). The data are reported with 95% binomial confidence intervals (CIs); the chisquared test and the Fisher exact test were performed for comparisons of qualitative variables, whereas the Manne Whitney U test was used for comparing quantitative variables. Multivariate logistic regression analyses were used to calculate odds ratios (ORs) and their 95% CIs for undetected or falsenegative cases in SLN mapping. In survival analysis, recurrence free survival was measured from the date of surgery to the date of the first evidence of recurrence or death. Survival probabilities were estimated by KaplaneMeier method, and group differences were assessed by log-rank test. All statistical analyses were carried out using SPSS 17.0 software (SPSS Inc, Chicago, IL).
3.
Results
3.1.
Clinicopathologic background factors
3.2.
The SLN was successfully identified in 372 (96.6%, CI: 94.8%e98.4%) of the 385 patients using the dye-guided and radio-guided methods (Table 2). The mean number of identified SLNs per patient was 4.7 (range 1e24). Only one SLN was found in 38 (10.2%) patients, 3 SLNs were found in 149 (40.1%), and 4 SLNs were found in 221 (59.4%) patients. Meanwhile, in patients who underwent laparoscopic-assisted gastrectomy, SLNs were identified in 131 (97.0%) of 135 patients. The mean number of SLNs was 5.0 in patients undergoing open surgery and 4.2 in patients undergoing laparoscopic-assisted surgery. The mean number of detected SLNs was not significantly different between open surgery and laparoscopy-assisted surgery. Thirty-four (89.5%) of 38 cases with lymph node metastases showed positive SLNs. Eleven patients had positive SLN and positive non-SLN (SLN(þ)/non-SLN(þ)), and 23 patients had positive SLN and negative non-SLN (SLN(þ)/non-SLN()). The accuracy of SLN biopsy for the detection of lymph node metastases was 98.9% (368 of 372).
3.3.
Clinicopathologic features of false-negative cases
In four of the 38 patients, SLN examination revealed a falsenegative diagnosis. The primary tumor was <4 cm in diameter in all four false-negative cases (Table 3). Three of these tumors invaded into the muscularis propria or deeper on pathologic examination. Three of these cases had metastases in non-SLN within the lymphatic area of the SLN (the sentinel basin). Only one case (case 3) had a metastatic lymph node in the node along the lesser curvature (#3), although the SLN was in the nodes along the common hepatic artery (#8a). This tumor invaded to the subserosa on pathologic examination.
3.4. A total of 385 consecutive patients were enrolled in this study. The clinicopathologic characteristics of the patients enrolled in this study are summarized in Table 1. On pathology, 23
SLN identification
Results in the earlier and later periods
The SLN detection rate was 93.7% in the earlier period (1999e2002) and 98.8% in the later period (2003e2007; Table 4).
Table 3 e False-negative cases. Case
Age
Sex
SLN#*
1
57
F
2 3
55 52
F M
3, 4d, 6, 7, 11p 1, 3 8a
4
63
M
3, 5, 7
Non-SLN met#y
cT
4d, 11p
1
3 3 3
Location
Location
Type
Size (mm)
Histology
Depth
Recurrence
M
Greater
0-IIc
30
Undiff.
MP
d
1 2
U M
Less Less
0-IIc 3
30 25
Undiff. Diff.
SM SS
1
M
Less
0-IIc
29
Diff.
MP
d Liver, po 1Y3M d
(U/M/L)
diff. ¼ differentiated type; L ¼ lower third of the stomach; M ¼ middle third of the stomach; MP ¼ muscularis propria; SM ¼ submucosa; SS ¼ subserosa; U ¼ upper third of the stomach; undiff. ¼ undifferentiated type. * SLN# ¼ station number of the SLN. y Non-SLN met# ¼ station number of the metastatic non-SLN (the Japanese Gastric Cancer Association, Japanese Classification of Gastric Carcinoma, second English Edition [14]).
5
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 4 e Results in the earlier period (1999e2002) and the later period (2003e2007).
SLN
Earlier period (159 patients)
Later period (226 patients)
The detection rate: 149 of159 (93.7%)
The detection rate: 223 of 226 (98.8%)
Nodal status from all dissected LN Positive
Negative
Positive 20 (13.4%) 0 (0%) Negative 3 (2.0%) 126 (84.6%) Total 23 126 False-negative cases: 3 (2.0%) The accuracy rate based on the SLN biopsy: 98.0%
20 129 149
Tumor depth
The accuracy rates based on the SLN biopsy by cT1 and cT2 are listed in Table 5. The false-negative rate was 10.7% and 10.0% in cT1 and cT2 cases, respectively. The accuracy rate based on the SLN biopsy was 99.2% in cases with cT1 and 96.0% in cases with cT2, which was not significantly different between the two groups (P ¼ 0.24). Meanwhile, the detection rate was 96.6% in 348 cases with pT1 and 97.1% in 35 cases with pT2 or deeper. The accuracy rate based on the SLN biopsy was 99.7% in cases with pT1 and 91.2% in cases with pT2 or deeper, and the accuracy rate was significantly better in pT1 cases than pT2 or deeper cases (P < 0.01).
3.6.
Nodal status from all dissected LN
Total
The SLN detection rates were significantly better in the later period than in the earlier period (P < 0.01; Table 4). Clinical tumor invasion (cT), pathological lymph node metastasis (pN), and operative method were significantly different between the two periods (Table 1). The histological type and tumor size were not significantly different. There were three falsenegative cases (2.0%) in the earlier period and only one case (0.4%) in the later period. The percentage of the undetected and false-negative cases each year was shown in Figure 1. In particular, most of the falsenegative cases were concentrated in the initial 2-y period.
3.5.
SLN
Intraoperative diagnosis of SLN metastases
In 5 (13.2%) of 38 node-positive cases, all SLNs had no metastases on intraoperative rapid diagnosis. However, seven
Positive
Positive 14 (6.3%) 0 (0%) Negative 1 (0.4%) 208 (93.3%) Total 15 208 False-negative cases: 1 (0.4%) The accuracy rate based on the SLN biopsy: 99.6%
Total 14 214 223
SLNs in the five cases had micrometastases, not macrometastases, on histopathologic diagnosis of the permanent sections. Three cases were ultimately evaluated for the presence of isolated tumor cells (ITCs). Remarkably, non-SLNs beyond the sentinel basin had no metastases in these cases with micrometastasis and/or ITCs.
3.7. The relationship between the location of the primary lesion and the SLNs The station numbers of the SLNs and the frequencies by tumor site are listed in Tables 6 and 7. The tumors in the upper third (U area) had SLNs at LNs along the left gastric artery (#3, #1, and #7) in many cases and few or no SLNs at the suprapyloric LN (#5), infrapyloric LN (#6), and the LN along the common hepatic artery (#8a). The tumors in the middle third (M area) had SLNs at the right paracardial LN and along the lesser curvature and left gastric and gastroepiploic arteries. The tumors in the lower third (L area) had SLNs at #6 and LNs along the right gastroepiploic artery (#4d) with high frequency, and SLNs along the superior mesenteric vein (#14v) with low frequency, but one had a positive SLN at #14v. The tumors in L area had no SLNs at the left paracardial LN (#2) and the LN along the short gastric artery (#4sa).
3.8. Predictors for undetected or false-negative cases on SLN mapping Multivariate logistic regression analyses identified the factors associated with undetected or false-negative cases. Candidate variables included location, size, and depth of the tumor, tumor differentiation, lymphatic invasion, venous invasion, operative method (open surgery versus laparoscopic surgery), sex, and period (earlier period versus later period). The best predictor for undetected or false-negative cases on SLN mapping was the period (odds ratio ¼ 0.056; 95% CI ¼ 0.01e0.42), as listed in Table 8.
3.9.
Fig. 1 e Changes in percentage of the undetected cases and false-negative cases.
Negative
Survival
The 5-y recurrence free survival (RFS) rates were 96.2% in all patients. Moreover, 5y-RFS were 99.0% in negative SLN (SLN()) cases compared with 86.3% in positive SLN (SLN(þ)) cases. SLN(þ) cases had a worse prognosis than SLN() cases (P ¼ 0.008; Fig. 2). In SLN(þ) cases, RFS was not significantly different between the negative non-SLN cases
6
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 5 e Results as per tumor depth and size. cT1 (365 patients)
cT2 (26 patients)
The detection rate: 353/365 (96.7%)
The detection rate: 25/26 (96.1%)
SLN
Non-SLN Positive
Negative
Positive 8 (2.3%) 17 (4.8%) Negative 3 (0.8%) 325 (92.1%) Total 11 342 False-negative cases: 3 (0.8%) The accuracy rate based on SLN biopsy: 99.2% pT1 (354 patients)
SLN Total
Positive
25 328 353
Positive 3 (12.0%) 6 (24.0%) Negative 1 (4.0%) 15 (60.0%) Total 4 21 False-negative cases: 1 (4.0%) The accuracy rate based on SLN biopsy: 96.0% pT2 and pT3 (35 patients)
The detection rate: 342/354 (96.6%) SLN
Negative
Positive 6 (1.8%) 14 (4.1%) Negative 1 (0.3%) 321 (93.9%) Total 7 335 False-negative cases: 1 (0.3%) The accuracy rate based on SLN biopsy: 99.7%
Negative
Total 9 16 25
The detection rate: 34/35 (97.1%)
Non-SLN Positive
Non-SLN
SLN
Non-SLN
Total
Positive
Negative
20 322 342
Positive 5 (14.3%) 9 (26.5%) Negative 3 (8.8%) 17 (50.0%) Total 8 26 False-negative cases: 3 (8.8%) The accuracy rate based on SLN biopsy: 91.2%
Total 14 20 34
cT1 ¼ clinically tumor invaded lamina propria and submucosa; cT2 ¼ clinically invaded muscularis propria; pT1 ¼ pathologically tumor invaded lamina propria and submucosa; pT2 ¼ pathologically invaded muscularis propria; pT3 ¼ pathologically tumor penetrated serosa without invasion of adjacent structures.
(SLN(þ)/non-SLN()) and positive non-SLN cases (SLN(þ)/ non-SLN(þ)) (P ¼ 0.511; Fig. 3).
Table 6 e Results as per tumor site. Upper third area (61 patients) Detection rate: 58/61 (95.1%) SLN
4.
Discussion
Non-SLN Positive
Negative
Total
Positive 1 (1.7%) 5 (8.6%) 6 Negative 1 (1.7%) 51 (87.9%) 52 Total 2 56 58 False-negative cases: 1 (1.7%) The specificity true positive rate based on SLN biopsy: 98.3% Middle third area (230 patients) The detection rate: 227/230 (98.7%) SLN
Non-SLN Positive
Negative
Total
Positive 6 (2.6%) 14 (6.2%) 20 Negative 3 (1.3%) 204 (89.9%) 207 Total 9 218 227 False-negative cases: 3 (1.3%) The specificity true positive rate based on SLN biopsy: 98.7% Lower third area (100 patients) The detection rate: 93/100 (93.0%) SLN
Non-SLN Positive
Negative
Total
Positive 4 (4.3%) 4 (4.3%) 8 Negative 0 (0%) 85 (91.4%) 85 Total 4 89 93 False-negative cases: 0 (0%) The specificity true positive rate based on SLN biopsy: 100%
We previously developed SLN mapping using radio-guided and dye-guided methods in gastric cancer and reported the procedure for the first time [6e8]. Similar results concerning the feasibility of the SLN concept in gastric cancer have been reported in published studies [9e12,15]. The present report is the largest single institutional report of SLN in gastric cancer to the best of our knowledge. The SLN concept has been modified for clinical practice in several institutions [10e12,16]. The eligibility criteria for the SLN concept were cT1N0M0 or cT2N0M0, 4 cm tumor, and single tumor gastric cancer patients in most reports [9,12]. However, before introducing SLN biopsy in gastric cancer, we must be concerned about the learning curve of the SLN mapping technique and the indications with respect to tumor depth in particular. SLN biopsy was performed for 385 untreated cT1N0M0 or cT2N0M0 gastric cancer patients from 1999 to 2007. For all cases, the accuracy rate was 98.9%. The result was a match even for melanoma or breast cancer cases, in that the SLN concept has been shown to be generally applicable clinically. Furthermore, false-negative cases were seen in four patients in whom the SLNs were detected intraoperatively. From the perspective of tumor depth, the tumors penetrated the muscularis propria or deeper pathologically in three of the four false-negative cases. Given this, SLN mapping has to be adapted for gastric cancer patients whose tumor penetrates the mucosa or submucosa. The present study showed that the
7
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Table 7 e Location of the SLN by tumor site. U (n ¼ 58)
Station No.
1 2 3 4sa 4sb 4d 5 6 7 8a 9 11p 11d 12a 14v
M (n ¼ 221)
L (n ¼ 93)
SLN
Metastasis
SLN
Metastasis
SLN
Metastasis
19 (33) 5 (9) 44 (76) 3 (5) 3 (5) 2 (3) 2 (3) d 17 (29) d 1 (2) 2 (3) 2 (3) d d
1 (2) 1 (2) 1 (2) 0 0 0 0 d 1 (2) d 0 0 1 (2) d d
40 (18) 2 (1) 160 (72) 2 (1) 22 (10) 79 (36) 14 (6) 32 (14) 70 (32) 18 (8) 2 (1) 2 (1) 1 (0) d d
3 (1) 0 10 (5) 1 (0) 1 (0) 5 (2) 0 0 6 (3) 0 0 0 0 d d
5 (5) d 39 (42) d 1 (1) 31 (33) 17 (18) 52 (56) 26 (28) 12 (13) 2 (2) d d 1 (1) 3 (3)
0 d 4 (4) d 0 1 (1) 0 2 (2) 2 (2) 1 (1) 0 d d 0 1 (1)
The station number is described according to the Japanese Gastric Cancer Association, Japanese Classification of Gastric Carcinoma, second English Edition [14]. No. 1 ¼ right paracardial lymph node (LN); No. 2 ¼ Left paracardial LN; No. 3 ¼ LN along the lesser curvature; No. 4sa ¼ LN along the short gastric vessels; No. 4sb ¼ LN along the left gastroepiploic vessels; No. 4d ¼ LN along the right gastroepiploic vessels; No. 5 ¼ suprapyloric LN; No. 6 ¼ infrapyloric LN; No. 7 ¼ LN along the left gastric artery; No. 8a ¼ LN along the common hepatic artery (anterosuperior group); No. 9 ¼ LN around the celiac artery; No. 11p ¼ LN along the proximal splenic artery; No. 11d ¼ LN along the distal splenic artery; No. 12a ¼ LN in the hepatoduodenal ligament (along the hepatic artery); No. 14v ¼ LN along the superior mesenteric vein. L ¼ lower third of the stomach; M ¼ middle third of the stomach; No. ¼ station number; U ¼ upper third of the stomach.
accuracy rate based on the SLN biopsy was markedly worse in pathologic T2 or deeper cases. There has been a report that the probability of falsenegative SLN increases with increasing tumor size in breast cancer [17]. In our study, the accuracy rate based on the SLN biopsy did not show a significant difference in gastric cancer patients with tumor size 4 cm. However, tumor size was a simple prognostic indicator. The gastric cancer patients with tumor 4 cm had a significantly worse prognosis as compared with tumor <4 cm [18]. Moreover, as the size of the tumor is >4 cm, it is more difficult to perform the standard limited gastrectomy. A size of <4 cm, approximately, might be appropriate for SLN mapping in gastric cancer. We should do accurate preoperative evaluations of tumor depth in particular. Several pathologic T2 or deeper tumors
might be contained within clinical T1 cases. As a result, the false-negative cases might increase. On detailed analysis of the false-negative cases, there was a relationship between SLNs and positive lymph nodes pathologically. The positive lymph nodes stayed within the sentinel basin in all but one case. If SLNs are detected in gastric cancer for virtually all cases, sentinel basin dissection guards against the possibility of leaving positive lymph nodes. There are many contrasting views about the sentinel basin. However, considering that tumor cells might flow to the distant LNs through the close SLNs and that the SLNs might be misidentified because of tracer being prevented from reaching the LNs for any reason, basin dissection should be regarded as a useful and important part of the safety net.
Table 8 e Results of multivariate logistic regression analysis of predictors for undetected or false-negative cases with SLN mapping. Variable
Odds ratio
95% CI
P value
Location of tumor Size of tumor Pathologic tumor depth Tumor differentiation Lymphatic invasion Venous invasion Operative methods (open versus laparoscopic) Gender Periods (the earlier versus the later)
1.96 0.97 3.42 0.51 0.95 3.91 6.32
0.76e5.08 0.93e1.02 0.67e17.42 0.16e1.65 0.24e3.79 0.53e28.81 0.95e42.11
0.166 0.269 0.140 0.258 0.942 0.181 0.057
0.99 0.056
0.28e3.42 0.01e0.42
0.981 0.005
Fig. 2 e Recurrence-free survival in SLN(L) versus SLN(D). (Color version of figure is available online.)
8
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
Fig. 3 e Recurrence-free survival in SLN(D)/non-SLN(L) versus SLN(D)/non-SLN(D). (Color version of figure is available online.)
When comparing the results between the earlier period and the later period in the present study, we should point out the problem of the learning curve in the SLN mapping technique. Relatively few studies have reported the learning curve on SLN mapping on gastric cancer. The results of SLN mapping in gastric cancer in this study might improve with progress of the technique. In breast cancer, SLN mapping should be performed by surgeons with appropriate training and experience [19,20]. As a measure of whether to proceed with SLN biopsy, a detection rate of at least 95% has been recommended [21]. When SLN biopsy in gastric cancer would be used routinely in clinical care, we have to focus more attention on the stabilization or learning curve of the technique as with breast cancer. In our study, the percentage of the undetected and false-negative cases tended to be reduced gradually year by year (Fig. 1). The undetected cases in 2005 were increased for unaccountable reason. When SLN was not detected, the conventional lymphadenectomy had to be performed for these patients. The present results indicate that undetected or falsenegative cases with SLN mapping were significantly associated only with the time period (Table 8). This strongly suggests that stabilization of the procedure and experience with SLN mapping in gastric cancer might be associated with a decrease in undetected or false-negative cases. From another point of view, there was variability among the preoperative clinical features between the two periods. That essential differences as the diagnosis of tumor depth and lymph node metastases suggested improvements of the SLN detection rate. To explain the differences by time period, numerous factors may have been involved, such as the development of pretreatment diagnostic capability and clarification of the indications for SLN mapping in gastric cancer. In 226 patients with cT1 or cT2N0M0 in the present study, a single lesion, and no previous treatment in the later period, the detection rate was 98.7% (223 of 226), the accuracy rate based on the SLN biopsy was 99.6% (222 of 223), and there was only one false-negative patient. Although there is no standard of SLN biopsy in gastric cancer, surgeons should not abandon radical lymph node dissection until a certain level of results can be achieved consistently.
There is a limitation of intraoperative rapid diagnosis. A discrepancy between the intraoperative SLN diagnosis and the final diagnosis was frequent, with eight cases (five micrometastasis and three ITC cases) in the present study. No cases had macrometastases. The most subtle histologic subclass of lymph node micrometastasis and macrometastasis has been defined as lymph nodes containing metastatic foci measuring number >2 mm in diameter [21]. These micrometastasis and/or ITC cases, furthermore, had no lymph node metastases beyond the sentinel basin. These results support the idea that sentinel basin dissection might be enough even in cases with intraoperatively negative SLNs to guarantee curability. The location of SLNs by tumor site is notable. The tumors in U area had few or no SLNs at the suprapyloric LN (#5), infrapyloric (#6) LN, and the LN along the common hepatic artery (#8) and had positive SLNs along the splenic artery (#11) with low frequency. When we select proximal or total gastrectomy for tumors in the upper gastric area, if the location of the SLN is limited in the paracardiac LN or along the left gastric and splenic arteries, the lower third of the stomach and LNs around #5, #6, and #8 might be safely preserved. Moreover, the LN at #11 should be considered for removal. Tumors in the middle third gastric area had lymphatic flow into right paracardial LN (#1) and along the lesser curvature (#3), the left gastric artery (#7), and the gastroepiploic artery (#4d, #4sb). Transected gastrectomy might be selected for a tumor in the middle third gastric area. The tumors in the lower third of the stomach had lymphatic flow to the hepatic artery (#8a, #12) or superior mesenteric vein (#14v), and one had metastasis at #14v. The LNs at #14v might not always be dissected for all patients with tumor in L area. However, if the tumors have SLNs at #14v, dissection is worthy of consideration. At any rate, we have to take into account the probable consequences with respect to curability and invasive potential. In survival analysis, 5y-RFS in SLN() cases was significantly better than that in SLN(þ) cases. This suggested that SLN status was a predictive marker for recurrence as in usual pathologic nodal status. Originally, we speculated that the poor prognosis of SLN(þ) cases was guided by the outcomes of SLN(þ)/ non-SLN(þ) cases. However, the prognosis of SLN(þ)/nonSLN(þ) cases was the same as that for SLN(þ)/non-SLN() cases by ensuring that SLN biopsy was followed by conventional lymphadenectomy. Therefore, if SLN was positive, conventional lymphadenectomy has to be performed. In the present study, SLN mapping for gastric cancer with cT1N0M0, diameter of primary tumor 4 cm, single lesion, and no previous treatment was considered acceptable in terms of the detection rate and the accuracy in determining SLN status. Furthermore, sentinel basin dissection is recommended for these patients to maintain safety [22]. The individualized, minimally invasive gastrectomy with selective and modified lymphadenectomy based on the SLN concept will lead to the improvement of long-term quality of life in the future.
Acknowledgment Y.K. made a contribution to conception and design. Y.S., T.T., N.W. made contributions to surgery and collecting data. M.N.
j o u r n a l o f s u r g i c a l r e s e a r c h x x x ( 2 0 1 5 ) 1 e9
made a contribution to managing the radioactive isotope tracer. M.M. made a contribution to pathology diagnosis. H.T. made a contribution to approving the final version of the article.
Disclosure Professor Kitagawa has grants or payment for lectures from Ajinomoto Pharmaceuticals Co., ltd., Asahi Kasei Co., Ltd, Astellas Pharma Inc., Otsuka Pharmaceutical Factory, Inc., Otsuka Pharmaceutical Co., Ltd, ONO Pharmaceutical, Co., Ltd, Kaken Pharmaceutical, Co., Ltd, Kyowa Hakko Kirin Co., Ltd, GlaxoSmithKline Pharmaceuticals Ltd., LSE, Kureha Co., Kowa Pharmaceuticals, Co., Ltd, Shionogi & Co., Ltd, Sanofi K.K., CSL Behling KK, Taiho Pharmaceuticals, Co., Ltd, Chugai Pharmaceuticals, Co., Ltd, Takeda Pharmaceutical, Co., Ltd, Daiichi Sankyo Co., Ltd, TSUMURA & CO., Teijin Pharma Limited, Dainippon Sumitomo Pharma, Co., Ltd, Taisho Pharmaceuticals, Co., Ltd, Novartis Pharma, Co., Ltd, Nippon Kayaku Co., Ltd, Medtronic Japan Co., Ltd, Boehlinger Ingelheim GmbH, Pfizer Co., Ltd, Merck Serono Co., Ltd, Yakult Honsha Co., Ltd, Terumo Co., Torii Pharmaceutical, Co., Ltd, Astra Zeneca Co., Ltd, Ethicon Inc. (Part of Johnson & Johnson Family of Companies), and Olympus Co.
references
[8] [9]
[10] [11]
[12]
[13]
[14]
[15]
[16]
[17] [1] Maruyama K, Gunve0 n P, Okabayashi K, et al. Lymph node metastases of gastric cancer. Ann Surg 1989;210:596. [2] Songun I, Putter H, Kranenberg EM, et al. Surgical treatment of gastric cancer: 15-year follow-up results of the randomized nationwide Dutch D1D2 trial. Lancet Oncol 2010; 11:439. [3] Akagi T, Shiraishi N, Kitano S. Lymph node metastasis of gastric cancer. Cancer 2011;3:2141. [4] Folli S, Dente M, Dell’ Amore D, et al. Early gastric cancer: prognostic factors in 223 patients. Br J Surg 1995; 82:952. [5] Morton DL, Wen DR, Wong JH, et al. Technical details of intraoperatively lymphatic mapping for early stage melanoma. Arch Surg 1992;127:392. [6] Tonouchi H, Mohri Y, Tanaka K, et al. Laparoscopic lymphatic mapping and sentinel node biopsies for earlystage gastric cancer: the cause of false negativity. World J Surg 2005;29:418. [7] Uenosono Y, Natsugoe S, Ehi K, et al. Detection of sentinel nodes and micrometastases using radioisotope navigation
[18]
[19]
[20]
[21]
[22]
9
and immunohistochemistry in patients with gastric cancer. Br J Surg 2005;92:886. Park DJ, Lee HJ, Lee HS, et al. Sentinel node biopsy for cT1 and cT2a gastric cancer. Eur J Surg Oncol 2006;32:48. Kitagawa Y, Fujii H, Kitajima M, et al. The role of the sentinel lymph node in gastrointestinal cancer. Surg Clin North Am 2000;80:1799. Kitagawa Y, Kitajima M. Gastrointestinal cancer and sentinel node navigation surgery. J Surg Oncol 2002;79:188. Ichikura T, Sugasawa H, Ono S, et al. Limited gastrectomy with dissection of sentinel node stations for early gastric cancer with negative sentinel node biopsy. Ann Surg 2009; 249:942. Takeuchi H, Oyama T, Kitagawa Y, et al. Laparoscopyassisted proximal gastrectomy with sentinel node mapping for early gastric cancer. World J Surg 2011;35: 2463. Kitagawa Y, Takeuchi H, Takagi Y, et al. Sentinel node mapping for gastric Cancer: a prospective multicenter trial in Japan. J Clin Oncol 2013;31:3704. Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma -2nd English edition-. Gastric Cancer 1998;1:10. Ando N, Ogawa S, Kitagawa Y, et al. Improvement in the results of surgical treatment of advanced squamous esophageal carcinoma during 15 consecutive years. Ann Surg 2000;232:225. Takeuchi H, Fujii H, Kitagawa Y, et al. Validation study of radio-guided sentinel lymph node navigation in esophageal cancer. Ann Surg 2009;249:757. Coombs N, Chen W, Boyages J, et al. A decision tool for predicting sentinel node accuracy from breast tumor size and grade. Breast J 2007;13:593. Adachi Y, Oshiro T, Mori M, et al. Tumor size as a simple prognostic indicator for gastric carcinoma. Ann Surg Oncol 1997;4:137. Bass SS, Cox CE, Reintgen DS, et al. Learning curves and certification for breast cancer lymphatic mapping. Surg Oncol Clin N Am 1999;8:497. Schwartz GF, Guiliano AE, Veronesi U. Consensus Conference Committee. Proceeding of the consensus conference of the role of sentinel lymph node biopsy in carcinoma or the breast April 19-22, 2001, Philadelphia, PA, USA. Breast J 2002; 8:124. Huvos AG, Hutter RVP, Berg JW. Significance of axillary macrometastases and micrometastases in mammary cancer. Ann Surg 1971;173:44. Takeuchi H, Kitagawa Y. New sentinel node mapping technologies for early gastric cancer. Ann Surg Oncol 2013; 20:522.