Abstract
Background
It is unclear how preoperative therapy for gastric cancer affects the metastasis rate of lymph nodes (LNs) and whether the location of positive LNs affects survival after preoperative therapy. Therefore, we determined the association between positive central lymph nodes (CnLNs) and disease stage and overall survival (OS).
Methods
We reviewed a prospectively maintained database to identify patients who had undergone resection of gastric adenocarcinoma at our institution from 2005 to 2015. CnLNs were defined as common hepatic, celiac, and proximal splenic artery LNs (stations no. 8, 9, and 11p). The frequency of CnLN metastases and risk factors affecting OS were examined.
Results
We identified 356 patients. Preoperative therapy was administered to 66% of patients. D2 LN dissection was performed in 80% of patients, and the median number of LNs examined was 25 (IQR, 18–34). In 243 patients (68%), CnLNs had undergone separate pathologic examination; the CnLN-positive rate was 9.1% (22 of 243; station no. 8, 4.5%; no. 9, 2.1%; and no. 11p, 4.8%). CnLN metastasis was associated with shorter 3-year OS in patients with pN2/3 disease (33 vs. 62%; p = 0.004). Among patients who had undergone preoperative therapy, ypT3–4 stage (HR 2.44; p = 0.01) and positive CnLNs (HR 5.44; p < 0.001) were negatively associated with OS by multivariate analysis.
Conclusions
CnLN metastases are uncommon in gastric cancer and have an adverse effect on OS in patients who have undergone preoperative therapy. Larger multi-institutional studies are needed to determine whether CnLN positivity requires a separate staging category after preoperative therapy.
Keywords: Gastric cancer, Preoperative therapy, Lymph node station, Lymph node dissection, Survival
Introduction
The National Comprehensive Cancer Network (NCCN) guidelines for gastric cancer recommend D2 lymph node (LN) dissection on the basis of studies that have demonstrated an improvement in cancer-related mortality or survival in patients with positive LNs.1,2,3 However, while the recommended method of pathological staging, including LN assessment by station, is described in detail in the Japanese Classification of Gastric Carcinoma guidelines,4 it is not standardized among US institutions. The current American Joint Committee on Cancer (AJCC) staging system (7th edition) does not use the location of positive LNs to define N categories, and LNs are not examined separately by station at the majority of US institutions. Although the survival impact of location of positive LNs is well reported in the Japanese literature,5,6 no study has investigated the impact among patients who underwent preoperative therapy.
After the MAGIC trial demonstrated that preoperative therapy provided a survival benefit,7 it became a standard treatment strategy for advanced gastric cancer in the USA.8–12 However, it is unclear how preoperative therapy affects the metastasis rate of nodes obtained during D2 LN dissection and whether the location of positive nodes affects survival after preoperative therapy. We hypothesized that the positivity of LNs at a higher-order echelon is an important marker of treatment effect and is predictive of survival after preoperative therapy and surgery. The purpose of this study was to determine whether the positive LNs at a higher-order echelon are associated with survival in this era of preoperative therapy for gastric cancer.
Methods
After receiving institutional review board approval, we conducted a retrospective review of a prospectively maintained database of gastric and gastroesophageal carcinoma patients who had been treated at The University of Texas MD Anderson Cancer Center (Houston, Texas) from January 2005 to December 2015. For this study, central LNs (CnLNs) were defined as common hepatic, celiac, and proximal splenic artery LNs (stations no. 8, 9, and 11p; Fig. 1). In general, D2 gastrectomy, defined according to the National Cancer Center Network Guidelines, is the standard treatment for non-early gastric cancer at MD Anderson.13 CnLNs are typically dissected separately, after removal of stomach and perigastric LNs. Patients who had undergone potentially curative resection of gastric or gastroesophageal cancer were identified and included in the study.
Fig. 1.
Stations defined as central lymph nodes. No. 8: common hepatic artery, no. 9: celiac artery, and no. 11p: proximal splenic artery lymph nodes
The clinicopathologic variables collected included age, sex, self-reported race, diagnosis date, tumor location, histologic tumor grade, presence of signet ring cell features, pre-treatment TNM stage (AJCC, 7th edition; as determined by endoscopic ultrasonography [EUS]), use and types of preoperative treatment, date and type of surgical resection, extent of LN dissection and number of LNs examined, pathologic TNM stage (AJCC, 7th edition), post-operative complications, and last follow-up date and status. For patients in whom the pathological report included information on the presence or absence of metastasis at CnLNs, we obtained information on the rate of metastasis.
Statistical Methods
The differences between continuous variables were compared using the Wilcoxon rank-sum test, and differences between categorical variables were compared using Fisher’s exact test or chi-square tests, as appropriate. Overall survival (OS) was defined as the time interval between the dates of diagnosis and death and was censored at the last follow-up date for patients who were alive. The probabilities of 3-year OS were estimated using Kaplan-Meier methods, and the differences in OS among subgroups of patients were assessed using two-sided log-rank tests. Univariate and multivariate Cox proportional hazards regression models were used to assess the associations between patient characteristics and OS. Patient characteristics that were significant in the univariate models at the 0.10 level were included in the multivariate model. Backward elimination was implemented until all remaining predictors had a p value less than 0.05. Pathological N stage and central lymph node positivity were included in the final model, since these were considered important variables for the study purpose. Statistical analyses were performed using SAS software version 9.3 (SAS Institute, Cary, NC) and S-plus (TIBCO Corporation, Palo Alto, CA).
Results
This study included 356 patients (median age, 64 years [range, 54–72 years]; 209 male and 147 female). Table 1 shows the patients’ characteristics, both overall and by CnLN examination status. Patients who had undergone CnLN examination were more likely to have undergone preoperative therapy (p < 0.001), and diagnostic laparoscopy (p < 0.001); these differences were likely because of the higher proportion of patients with a recent diagnosis year (p < 0.001) and more aggressive tumor characteristics (poorly differentiated histologic grade [p = 0.03], linitis plastica [p = 0.01], and advanced EUS T stage [p = 0.02]).
Table 1.
Patient characteristics, overall and by CnLN examination status (N = 356)
| Variable | CnLN examined | p value | ||
|---|---|---|---|---|
| All patients, N =356 | Yes, N =243 | No, N =113 | ||
| Age, median (IQR), years | 64 (54–72) | 62 (54–70) | 66 (54–74) | 0.10 |
| Sex, N (%) %s OK | ||||
| Female | 147 (41.3) | 103 (42.4) | 44 (38.9) | 0.54 |
| Male | 209 (58.7) | 140 (57.6) | 69 (61.1) | |
| Race, N (%)%s OK | ||||
| White | 183 (51.4) | 124 (51.0) | 59 (52.2) | 0.99 |
| Black | 36 (10.1) | 25 (10.3) | 11 (9.7) | |
| Asian | 55 (15.4) | 37 (15.2) | 18 (15.9) | |
| Hispanic/Latino | 82 (23.0) | 57 (23.5) | 25 (22.1) | |
| Diagnosis year, N (%) | ||||
| 2004–2009 | 160 (44.9) | 80 (32.9) | 80 (70.8) | < 0.001 |
| 2010–2015 | 196 (55.1) | 163 (67.1) | 33 (29.2) | |
| Preoperative therapy, N (%) | ||||
| Yes | 235 (66.0) | 183 (75.3) | 52 (46.0) | < 0.001 |
| Chemotherapy only | 74 (20.8) | 46 (18.9) | 28 (24.8) | |
| Chemoradiation therapy | 161 (45.2) | 137 (56.4) | 24 (21.2) | |
| No | 121 (34) | 60 (24.7) | 61 (54) | |
| Post-operative therapy, N (%) | ||||
| Yes | 69 (19.4) | 48 (19.8) | 21 (18.6) | 0.80 |
| Chemotherapy | 69 (19.4) | 48 (19.8) | 21 (18.6) | |
| Radiation therapy | 23 (6.5) | 14 (5.8) | 9 (8) | |
| No | 287 (80.6) | 195 (80.2) | 92 (81.4) | |
| Diagnostic laparoscopy, N (%) | ||||
| Yes | 180 (50.6) | 152 (62.6) | 28 (24.8) | < 0.001 |
| No | 176 (49.4) | 91 (37.4) | 85 (75.2) | |
| Type of resection, N (%) | ||||
| Total gastrectomy | 182 (51.1) | 123 (50.6) | 59 (52.2) | 0.09 |
| Subtotal gastrectomy | 160 (44.9) | 114 (46.9) | 46 (40.7) | |
| Proximal gastrectomy | 14 (3.9) | 6 (2.5) | 8 (7.1) | |
| Pancreatectomy, N (%) | ||||
| Yes | 8 (2.2) | 5 (2.1) | 3 (2.7) | 0.71 |
| No | 348 (97.8) | 238 (97.9) | 110 (97.3) | |
| Splenectomy, N (%) | ||||
| Yes | 9 (2.5) | 4 (1.6) | 5 (4.4) | 0.15 |
| No | 348 (97.8) | 238 (97.9) | 110 (97.3) | |
| Post-operative complication < 30 days, N (%) | ||||
| Yes | 101 (28.4) | 67 (27.6) | 34 (30.1) | 0.62 |
| No | 255 (71.6) | 176 (72.4) | 79 (69.9) | |
| Post-operative mortality < 90 days, N (%) | ||||
| Yes | 5 (1.4) | 2 (0.8) | 3 (2.7) | 0.33 |
| No | 351 (98.6) | 241 (99.2) | 110 (97.3) | |
| Hospital stay, median (IQR), days | 11.5 (10–15) | 11 (10–15) | 12 (10–18) | 0.03 |
| Location, N (%) | ||||
| GEJ/cardia | 89 (25) | 63 (25.9) | 26 (23) | 0.32 |
| Body/fundus | 185 (52) | 125 (51.4) | 60 (53.1) | |
| Antrum/pylorus | 69 (19.4) | 49 (20.2) | 20 (17.7) | |
| Total | 13 (3.7) | 6 (2.5) | 7 (6.2) | |
| Histology grade, N (%) (N = 337) | ||||
| Well differentiated | 2 (0.6) | 0 (0) | 2 (2) | 0.03 |
| Moderately differentiated | 89 (26.4) | 57 (24.1) | 32 (32) | |
| Poorly differentiated | 246 (73) | 180 (75.9) | 66 (66) | |
| Signet ring cell, N (%) | ||||
| Yes | 179 (50.3) | 127 (52.3) | 52 (46) | 0.27 |
| No | 177 (49.7) | 116 (47.7) | 61 (54) | |
| Linitis plastica, N (%) | ||||
| Yes | 8 (2.2) | 8 (3.3) | 0 (0) | 0.01 |
| No | 340 (95.5) | 227 (93.4) | 113 (100) | |
| Suspicious | 8 (2.2) | 8 (3.3) | 0 (0) | |
| EUS T stage, N (%) (N =310) | ||||
| 0 | 10 (3.2) | 3 (1.4) | 7 (7.8) | 0.02 |
| 1 | 45 (14.5) | 30 (13.6) | 15 (16.7) | |
| 2 | 50 (16.1) | 33 (15) | 17 (18.9) | |
| 3 | 177 (57.1) | 130 (59.1) | 47 (52.2) | |
| 4 | 28 (9) | 24 (10.9) | 4 (4.4) | |
| EUS N stage, N (%) (N = 304) | ||||
| 0 | 187 (61.5) | 132 (60.3) | 55 (64.7) | 0.48 |
| 1 | 117 (38.5) | 87 (39.7) | 30 (35.3) | |
| Pathological T stage, N (%) | ||||
| 0 | 43 (12.1) | 32 (13.2) | 11 (9.7) | 0.04 |
| 1 | 103 (28.9) | 59 (24.3) | 44 (38.9) | |
| 2 | 48 (13.5) | 36 (14.8) | 12 (10.6) | |
| 3 | 114 (32) | 85 (35) | 29 (25.7) | |
| 4 | 48 (13.5) | 31 (12.8) | 17 (15) | |
| Pathological N stage, N (%) | ||||
| 0 | 218 (61.2) | 149 (61.3) | 69 (61.1) | 0.96 |
| 1 | 73 (20.5) | 51 (21) | 22 (19.5) | |
| 2 | 31 (8.7) | 21 (8.6) | 10 (8.8) | |
| 3 | 34 (9.6) | 22 (9.1) | 12 (10.6) | |
| R margin, N (%) | ||||
| R0 | 342 (96.1) | 230 (94.7) | 112 (99.1) | 0.04 |
| R1 | 14 (3.9) | 13 (5.3) | 1 (0.9) | |
| Number of LN examined, median (IQR) | 25 (18–34) | 27 (21–35) | 19 (14–27) | < 0.001 |
| Number of positive LN, median (IQR) | 0 (0–2) | 0 (0–1) | 0 (0–2) | 0.75 |
IQR interquartile range
CnLN Metastasis
The median number of examined CnLNs was 5 (range, 2–8, with 2 [1–4] at station no. 8, 3 [1–5] at no. 9, and 2 [1–4] at no. 11p). Twenty-two (9.1%) of the 243 patients had CnLN metastasis (11 [4.5%] at station no. 8, 5 [2.1%] at no. 9, and 9 [3.7%] at no. 11p). By univariate analysis, the positive CnLN rates were higher in patients with advanced pathological T stage (0% in pT1, 5.6% [2/36] in pT2, 17.6% [15/85] in pT3, and 12.9% [4/31] in pT4; p = 0.001), advanced pathological N stage (19.6% [10/51] in pN1, 9.5% [2/21] in pN2, and 45.5% [10/22] in pN3; p < 0.001), and positive EUS N status (5.3% [7/132] in EUS N-negative, and 13.8% [12/87] in EUS N-positive status; p = 0.03). EUS T stage was not associated with positive CnLN status (p = 0.49). When an additional subset analysis was performed of only patients who had been treated with preoperative therapy, neither EUS T (p = 0.46) nor N stage (p = 0.10) was associated with CnLN metastasis, while advanced pathologic T (0% in ypT1, 3.4% [1/29] in ypT2, 16.2% [12/74] in ypT3, and 11.5% [3/26] in ypT4; p = 0.05) and N stages (20.5% [8/39] in ypN1, 11.1% [2/18] in ypN2, and 43.8% [7/16] in ypN3; p < 0.001) were associated with CnLN metastasis.
OS after R0 Resection
Among 342 patients who had undergone R0 resection, 99 (29%) had died and 243 (71%) were alive at the last follow-up. The median follow-up duration among survivors was 3.6 years (range, 0.2–12.1 years). The median OS from diagnosis was 11.6 years (95% confidence interval [CI], 7.8 years to not estimable).
Table 2 shows the comparison of 3-year OS rates between patients with positive CnLNs and negative CnLNs for each N category among the patients who had undergone R0 resection and pathological CnLN assessment, in all study patients (n = 230, Table 2 (A)) and in preoperative therapy patients (n = 170, Table 2 (B)). CnLN metastasis was associated with lower 3-year OS rates in patients with N2/3 (33 vs. 62%; p = 0.004), but it was not associated with OS in those with N1 disease (3-year OS 69 vs. 85%, median OS 6.1 vs. 5.7 years; p = 0.238) (Table 2 (A), Figure 2). Similar outcomes were observed in analyses of preoperative therapy patients (Table 2 (B)).
Table 2.
Comparison of positive and negative CnLNs across AJCC pN categories, in 230 patients who underwent R0 resection and CnLNs examined
| LN status | N | 3-year OS | Median OS, years | N | 3-year OS | Median OS, years | p value |
|---|---|---|---|---|---|---|---|
| A. All patients | |||||||
| Positive CnLN (N =21) | Negative CnLN (N =209) | ||||||
| pN0 (N =142) | 0 | N/A | N/A | 142 | 89% | Not estimable | NA |
| pN1 (N =47) | 10 | 69% | 6.1 | 37 | 85% | 5.7 | 0.238 |
| pN2/3 (N =41) | 11 | 33% | 1.2 | 30 | 62% | 3.2 | 0.004 |
| B. Preoperative therapy patients only | |||||||
| Positive CnLN (N =16) | Negative CnLN (N =154) | ||||||
| pN0 (N =103) | 0 | N/A | N/A | 103 | 85% | Not estimable | NA |
| pN1 (N =35) | 8 | 60% | Not estimable | 27 | 84% | 5.7 | 0.0852 |
| pN2/3 (N = 32) | 8 | 20% | 1.1 | 24 | 67% | 3.2 | < 0.001 |
p value less than 0.05 was shown in italics
Fig. 2.
Kaplan-Meier estimate of overall survival by pathological N stage and central lymph node positivity. CnLN, central lymph node
Analyses of risk factors that affect OS were performed in patients who had undergone R0 resection, by preoperative therapy status. A total of 221 patients underwent R0 resection after preoperative therapy; Table 3 (A) presents the results of univariate and multivariate Cox proportional hazards models for OS. Pathological T stage, pathological N stage, and CnLN positivity were included in the multivariate analysis, and pathological N stage and CnLN metastasis remained in the final model. CnLN metastasis was independently associated with shorter OS duration (HR, 3.73; 95% CI, 1.58–8.78; p = 0.003) after adjustment with pathological N stage.
Table 3.
Univariate and multivariate Cox analysis of overall survival (OS) in 342 R0 patients
| Variable | HR (95% CI) | p value |
|---|---|---|
| A. Preoperative therapy patients (N =221) | ||
| Univariate analysis | ||
| Age at diagnosis, ≥ 65 vs. < 65 years | 1.18 (0.73–1.89) | 0.51 |
| Sex, female vs. male | 1.01 (0.62–1.65) | 0.97 |
| Race, white vs. non-white | 0.96 (0.59–1.56) | 0.87 |
| Location, GEJ or cardia vs. other | 1.14 (0.69–1.88) | 0.60 |
| Histological grade, poorly vs. well or moderately | 1.65 (0.86–3.16) | 0.13 |
| Signet ring cell, yes vs. no | 1.32 (0.81–2.14) | 0.27 |
| EUS T stage, 3–4 vs. 0–2 | 9.33 (1.29–67.47) | 0.03 |
| EUS N stage, 1 vs. 0 | 1.73 (0.97–3.07) | 0.06 |
| Pathological T stage, 3–4 vs. 0–2 | 2.86 (1.66–4.91) | < 0.001 |
| Pathological N stage (ref. N0) | ||
| N1 | 2.28 (1.24–4.18) | 0.008 |
| N2/3 | 4.37 (2.48–7.71) | < 0.001 |
| LN dissection, D1 vs. D1 and D2 | 1.31 (0.69–2.5) | 0.41 |
| CnLN examined, yes vs. no | 0.76 (0.45–1.26) | 0.29 |
| CnLN metastasis, positive vs. negative | 6.78 (2.97–15.46) | < 0.001 |
| Multivariate analysis | ||
| Pathological N stage (ref. N0) | ||
| N1 | 1.71 (0.81–3.58) | 0.157 |
| N2/3 | 3.71 (1.76–7.82) | 0.001 |
| CnLN metastasis, positive vs. negative | 3.73 (1.58–8.78) | 0.003 |
| B. Non-preoperative therapy patients (N = 121) | ||
| Univariate analysis | ||
| Age at diagnosis, ≥ 65 vs. < 65 years | 1.65 (0.77–3.51) | 0.20 |
| Sex, female vs. male | 1.07 (0.53–2.17) | 0.85 |
| Race, white vs. non-white | 0.97 (0.48–1.97) | 0.93 |
| Location, GEJ/cardia vs. other | 1.21 (0.42–3.46) | 0.73 |
| Histological grade, poorly vs. well/moderately | 0.54 (0.26–1.13) | 0.10 |
| Signet ring cell, yes vs. no | 0.8 (0.39–1.64) | 0.54 |
| EUS T stage, 3–4 vs. 0–2 | 1.37 (0.57–3.31) | 0.48 |
| EUS N stage, 1 vs. 0 | 2.82 (0.95–8.37) | 0.06 |
| Pathological T stage, 3–4 vs. 0–2 | 2.14 (1.05–4.39) | 0.04 |
| Pathological N stage (ref. N0) | ||
| N1 | 2.18 (0.95–5.00) | 0.065 |
| N2/3 | 3.03 (1.23–7.46) | 0.016 |
| LN dissection, D1 vs. D1 and D2 | 1.83 (0.9–3.7) | 0.09 |
| CnLN examined, yes vs. no | 0.61 (0.29–1.27) | 0.19 |
| Central LN, positive vs. negative | 1.82 (0.23–14.47) | 0.57 |
| Multivariate analysis | ||
| Pathological N stage (ref. N0) | ||
| N1 | 3.44 (0.81–14.6) | 0.093 |
| N2/3 | 5.93 (1.28–27.4) | 0.023 |
| CnLN metastasis, positive vs. negative | 0.90 (0.10–7.74) | 0.921 |
R0 resection without preoperative therapy was performed in 121 patients; Table 3 (B) presents the results of univariate and multivariate Cox proportional hazards models for OS. Pathological T stage, pathological N stage, CnLN metastasis, and LN dissection were included in the multivariate analysis, and pathologic N stage and CnLN metastasis remained in the final model. In contrast to the results in preoperative therapy patients, CnLN positivity was not predictive of survival (HR, 0.90; 95% CI, 0.10–7.74; p = 0.921) after adjustment with pathological N stage.
Discussion
In this study of gastric cancer patients who had undergone resection, we showed that metastasis in CnLNs was rare (9.1%) but was independently associated with OS among patients who had undergone preoperative therapy (HR, 3.73; 95% CI, 1.58–8.78). Sensitivity analysis showed that CnLN metastasis negatively affected survival in pN2/3 patients, but not in pN1 patients. The results indicate the importance of identifying the location (station) of positive LNs to predict survival in this era of preoperative treatment for gastric cancer.
Data on LN metastasis rates at each LN station of gastric cancer are very limited except for those from Japan. Sasako et al. reported detailed data on the incidence of metastasis and the 5-year survival rate of patients with metastasis, according to each specific LN station.6 They reported that the incidence rates of metastasis at stations no. 8, 9, and 11 were approximately 19, 13, and 11% (estimated from a histogram). In our study, we observed lower rates of CnLN metastasis than those reported by Sasako et al. However, we previously reported relatively higher T stage-specific LN metastasis rates in patients who did not undergo preoperative therapy than those reported in Eastern reports.14 With the similar T stage distribution of cohorts in Sasako’s study and our current study, the reduced rate of CnLN metastasis that we found is likely to have been caused by the use of preoperative therapy. According to the current standard treatment strategy at MD Anderson, we frequently give a total of 45 Gy of preoperative radiation therapy with concurrent chemotherapy after induction chemotherapy in most patients with advanced (≥ T2N0) gastric cancer.8,15,16 We have been using an intensity-modulated radiation therapy (IMRT) technique to target regional LNs, including CnLN. Because of the low CnLN metastasis rate associated with preoperative therapy, the survival benefit of extended LN dissection may be lower after preoperative therapy.
Historically, the AJCC (up to its 4th edition17) and Japanese Gastric Cancer Classification systems (up to its 2nd English edition18) both used the location or station of positive LNs, rather than the number of involved nodes, to define N categories. However, there is low compliance in the pathologic assessment of gastric cancer LNs in the USA, which limits reliable comparisons among institutions. As a result of an international effort made to create an accurate, reproducible, and comparable staging system, both staging systems now use the number of positive LNs to categorize N stages. Karpeh et al. reported that the location of positive nodes was not predictive of survival when adjusted by the number of positive nodes and concluded that the number-based N staging system served as a better discriminator of survival.19 Similarly, Hayashi et al. reported the superiority of a number-based N staging system in predicting the survival of gastric cancer patients in Japan.20
Although the association between survival and involved lymph node location for patients treated with upfront surgery is well described in the Japanese literature, no study has investigated the association for patients who underwent preoperative therapy. A significant knowledge gap exists in how to classify Western patients with involved lymph nodes after preoperative treatment and how to integrate existing lymph node classification systems from Eastern countries. We recently reported that left gastric artery LNs (station no. 7) were a common location of LN metastasis and it was not associated with shorter survival after adjustment for pN category, which is consistent with Japanese guidelines to include no. 7 LNs within the extent of D1 LN dissection.21 This is an important finding as current NCCN guidelines classify no. 7 LNs within D2 LN dissection.13 Building upon this line of investigation, we sought to investigate the impact of the next level (stations no. 8, 9, and 11p) of lymph node stations.
The association between CnLN positivity and survival shown in this study suggests that CnLN positivity can be a possible marker of treatment effect after preoperative therapy. Indeed, one of the benefits of preoperative therapy is that we can pathologically assess treatment response, which is predictive of survival.22–24 In the AJCC 8th edition,25 gastric cancer patients who are treated with preoperative therapy will be categorized by ypStage (postneoadjuvant pathologic stage). ypStage was created based on the concept that estimated survival is likely different by preoperative therapy status, since pathologic stage is significantly affected by treatment response. In this context, the survival impact of CnLN positivity can also be different by preoperative therapy status, and this hypothesis was supported by these study results. Larger multi-institutional studies are needed to determine whether CnLN positivity requires a separate staging category in ypStage. Given the small number of patients with positive CnLNs in this study, the combination of the CnLN positivity and the number of positive LNs may be the most effective method of defining ypN categories.
We observed that patients with pN1 disease can achieve long-term survival after LN dissection, even with CnLN metastasis; these results support the routine use of D2 LN dissection for gastric cancer, as recommended by both Eastern and Western cancer staging guidelines.8,26,27 In contrast, CnLN metastasis had a strong negative impact on survival in patients with pN2/3 disease. This may be related to lymphatic drainage patterns—either direct or indirect drainage to the CnLN. Ten patients with pN1 disease had CnLN metastasis in our study. These patients may be representative of a direct lymphatic drainage pattern; in other words, CnLNs were Bsentinel lymph nodeŝ in those patients. In patients with pN1 disease and CnLN metastasis, removal of the CnLNs by D2 LN dissection may be critical to improving survival or at least locoregional control. In contrast, patients with pN2 or pN3 disease and positive CnLNs have likely experienced lymphatic spread through other lymphatic channels (indirect drainage to the CnLN), frequently through left gastric artery LNs (station no. 7). Such a spread of LN metastases is indicative of an aggressive tumor and ineffective preoperative therapy; as the disease has likely metastasized beyond the extent of the D2 LN dissection, this procedure is unlikely to be curative. Future studies of sentinel LN biopsy (or identification of a lymphatic drainage pattern) may be useful for identifying patients who are more likely to benefit from extended LN dissection. The feasibility of sentinel LN biopsy has been reported in Japan,28,29 and further reports of its usefulness are expected.
The limitations of this study include its inherent selection bias and the incompleteness of CnLN data due to its retrospective nature. However, extended LN dissection (D1+ or D2) has been historically performed at our facility, and the average number of LNs that underwent pathological examination was higher than that in previous Western reports, which increased the validity of our findings. A small number of patients with CnLN metastasis, likely due to our institutional preference towards preoperative therapy, limited the power of this study to detect a survival difference by CnLN positivity on stratified analysis. To our knowledge, this is the first US study to report metastasis rates for CnLNs in gastric cancer and the first study to report these rates in patients who have undergone preoperative therapy. Because of the complexity of the Japanese Gastric Cancer Classification system, particularly regarding definitions of LN stations, it is not commonly used in Western countries; the proposed definition of CnLNs integrates with current NCCN guidelines for D2 lymph node dissection and also allows for comparison with the Japanese classification system.13,27
In conclusion, this study reported that positive CnLNs in gastric cancer are a strong predictor of survival, especially in patients who have undergone preoperative therapy. Long-term survival is still possible in pN1 patients with positive CnLNs; however, patients with pN2 or pN3 disease and positive CnLNs have extremely poor survival. Larger multi-institutional studies are needed to determine whether CnLN positivity requires a separate staging category, particularly in patients who have undergone preoperative therapy.
Funding
This study is supported in part by the NIH/NCI under award number P30CA016672 and used the Clinical Trials Support Resource.
Footnotes
This paper was presented at 2017 American Society of Clinical Oncology Gastrointestinal Cancer Symposium (January 19, 2017, San Francisco, USA) and at Annual Cancer Symposium of Society of Surgical Oncology (March 16, 2017).
Conflict of Interest The authors declare that they have no conflicts of interest.
References
- 1.Songun I, Putter H, Kranenbarg EM, Sasako M, van de Velde CJ. Surgical treatment of gastric cancer: 15-year follow-up results of the randomised nationwide Dutch D1D2 trial. The Lancet Oncology. 2010;11(5):439–449. [DOI] [PubMed] [Google Scholar]
- 2.Degiuli M, Sasako M, Ponti A, et al. Randomized clinical trial comparing survival after D1 or D2 gastrectomy for gastric cancer. Br J Surg. 2014;101(2):23–31. [DOI] [PubMed] [Google Scholar]
- 3.Ajani JA, Bentrem DJ, Besh S, et al. Gastric cancer, version 2.2013: featured updates to the NCCN Guidelines. Journal of the National Comprehensive Cancer Network : JNCCN. 2013;11(5):531–546. [DOI] [PubMed] [Google Scholar]
- 4.Japanese Gastric Cancer A. Japanese gastric cancer treatment guidelines 2014 (ver. 4). Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association. 2016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Katai H, Yoshimura K, Maruyama K, Sasako M, Sano T. Evaluation of the New International Union Against Cancer TNM staging for gastric carcinoma. Cancer. 2000;88(8): 1796–1800. [PubMed] [Google Scholar]
- 6.Sasako M, McCulloch P, Kinoshita T, Maruyama K. New method to evaluate the therapeutic value of lymph node dissection for gastric cancer. Br J Surg. 1995;82(3):346–351. [DOI] [PubMed] [Google Scholar]
- 7.Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. The New England journal of medicine. 2006;355(1):11–20. [DOI] [PubMed] [Google Scholar]
- 8.Badgwell B Multimodality Therapy of Localized Gastric Adenocarcinoma. Journal of the National Comprehensive Cancer Network : JNCCN. 2016;14(10):1321–1327. [DOI] [PubMed] [Google Scholar]
- 9.Badgwell B, Blum M, Elimova E, et al. Frequency of Resection After Preoperative Chemotherapy or Chemoradiotherapy for Gastric Adenocarcinoma. Annals of surgical oncology. 2016;23(6):1948–1955. [DOI] [PubMed] [Google Scholar]
- 10.Shridhar R, Dombi GW, Finkelstein SE, Meredith KL, Hoffe SE. Improved survival in patients with lymph node-positive gastric cancer who received preoperative radiation: an analysis of the Surveillance, Epidemiology, and End Results database. Cancer. 2011;117(17):3908–3916. [DOI] [PubMed] [Google Scholar]
- 11.Ikoma N, Blum M, Estrella JS, et al. Evaluation of the American Joint Committee on Cancer 8th edition staging system for gastric cancer patients after preoperative therapy. Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association. 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ikoma N, Chen HC, Wang X, et al. Patterns of Initial Recurrence in Gastric Adenocarcinoma in the Era of Preoperative Therapy. Annals of surgical oncology. 2017;24(9):2679–2687. [DOI] [PubMed] [Google Scholar]
- 13.Ajani JA, D’Amico TA, Almhanna K, Bentrem DJ, Chao J, Das P et al. Gastric Cancer, Version 1.2017, NCCN Clinical Practice Guidelines in Oncology National Comprehensive Cancer Network; https://www.nccn.org/professionals/physician_gls/PDF/gastric.pdf. Accessed July 24th, 2017. [Google Scholar]
- 14.Ikoma N, Blum M, Chiang YJ, et al. Race Is a Risk for Lymph Node Metastasis in Patients With Gastric Cancer. Annals of surgical oncology. 2016. [DOI] [PubMed] [Google Scholar]
- 15.Badgwell B, Das P, Ajani J. Treatment of localized gastric and gastroesophageal adenocarcinoma: the role of accurate staging and preoperative therapy. J Hematol Oncol. 2017;10(1):149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Badgwell B, Blum M, Estrella J, et al. Predictors of Survival in Patients with Resectable Gastric Cancer Treated with Preoperative Chemoradiation Therapy and Gastrectomy. Journal of the American College of Surgeons. 2015;221(1):83–90. [DOI] [PubMed] [Google Scholar]
- 17.Beahrs OHHD, Hutter RVP, Kennedy BJ, editors Manual for Staging of Cancer (4th ed.). American Joint Committee on Cancer (AJCC). Philadelphia: Lippincott; 1992. [Google Scholar]
- 18.Japanese Gastric Cancer A Japanese Classification of Gastric Carcinoma - 2nd English Edition. Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association. 1998;1(1):10–24. [DOI] [PubMed] [Google Scholar]
- 19.Karpeh MS, Leon L, Klimstra D, Brennan MF. Lymph node staging in gastric cancer: is location more important than Number? An analysis of 1,038 patients. Annals of surgery. 2000;232(3):362–371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hayashi H, Ochiai T, Suzuki T, et al. Superiority of a new UICC-TNM staging system for gastric carcinoma. Surgery. 2000;127(2): 129–135. [DOI] [PubMed] [Google Scholar]
- 21.Ikoma N, Blum M, Estrella JS, et al. Left Gastric Artery Lymph Nodes Should Be Included in D1 Lymph Node Dissection in Gastric Cancer. Journal of gastrointestinal surgery : official journal of the Society for Surgery of the Alimentary Tract. 2017;21(10): 1563–1570. [DOI] [PubMed] [Google Scholar]
- 22.Lowy AM, Mansfield PF, Leach SD, Pazdur R, Dumas P, Ajani JA. Response to neoadjuvant chemotherapy best predicts survival after curative resection of gastric cancer. Annals of surgery. 1999;229(3): 303–308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Becker K, Mueller JD, Schulmacher C, et al. Histomorphology and grading of regression in gastric carcinoma treated with neoadjuvant chemotherapy. Cancer. 2003;98(7):1521–1530. [DOI] [PubMed] [Google Scholar]
- 24.Mansour JC, Tang L, Shah M, et al. Does graded histologic response after neoadjuvant chemotherapy predict survival for completely resected gastric cancer? Annals of surgical oncology. 2007;14(12):3412–3418. [DOI] [PubMed] [Google Scholar]
- 25.Amin MBES, Greene F, et al. AJCC Cancer Staging Manual 8th edition New York: Springer; 2017. [Google Scholar]
- 26.Ajani JA, D’Amico TA, Almhanna K, et al. Gastric Cancer, Version 3.2016, NCCN Clinical Practice Guidelines in Oncology. Journal of the National Comprehensive Cancer Network : JNCCN. 2016;14(10):1286–1312. [DOI] [PubMed] [Google Scholar]
- 27.Japanese Gastric Cancer A. Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association. 2011;14(2):113–123. [DOI] [PubMed] [Google Scholar]
- 28.Takeuchi H, Kitagawa Y. Sentinel lymph node biopsy in gastric cancer. Cancer journal. 2015;21(1):21–24. [DOI] [PubMed] [Google Scholar]
- 29.Takeuchi H, Goto O, Yahagi N, Kitagawa Y. Function-preserving gastrectomy based on the sentinel node concept in early gastric cancer. Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association. 2016. [DOI] [PubMed] [Google Scholar]