Abstract
Background:
Previous studies have revealed that diabetes mellitus (DM) promotes disease progress of gastric cancer (GC). This study aimed to further investigating whether DM advanced lymph nodes (LNs) metastasis in GC.
Methods:
The clinicopathologic data of GC patients with >15 examined LN (ELN) between October 2004 and December 2019 from a prospectively maintained database were included. The observational outcomes included the number (N3b status) and anatomical distribution (N3 stations) of metastatic LN (MLN).
Results:
A total of 2142 eligible patients were included in the study between October 2004 and December 2019. N3 stations metastasis (26.8% in DM vs. 19.3% in non-DM, P = 0.026) and N3b status (18.8% in DM vs. 12.8% in non-DM, P = 0.039) were more advanced in the DM group, and multivariate logistic regression analyses confirmed that DM was an independent factor of developing N3 stations metastasis (odds ratio [OR] = 1.771, P = 0.011) and N3b status (OR = 1.752, P = 0.028). Also, multivariate analyses determined DM was independently associated with more MLN (β = 1.424, P = 0.047). The preponderance of N3 stations metastasis (DM vs. non-DM, T1–2: 2.2% vs. 4.9%, T3: 29.0% vs. 20.3%, T4a: 38.9% vs. 25.8%, T4b: 50.0% vs. 36.6%; ELN16–29: 8.6% vs. 10.4%, ELN30–44: 27.9% vs. 20.5%, ELN ≥ 45: 37.7% vs. 25.3%), N3b status (DM vs. non-DM, T1–2: 0% vs. 1.7%, T3: 16.1% vs. 5.1%, T4a: 27.8% vs. 19.1%, T4b: 44.0% vs. 28.0%; ELN16–29: 8.6% vs. 7.9%, ELN30–44: 18.0% vs. 11.8%, ELN ≥ 45: 26.4% vs. 17.3%), and the number of MLN (DM vs. non-DM, T1–2: 0.4 vs. 1.1, T3: 8.6 vs. 5.2, T4a: 9.7 vs. 8.6, T4b: 17.0 vs. 12.8; ELN16–29: 3.6 vs. 4.6, ELN30–44: 5.8 vs. 5.5, ELN ≥ 45: 12.0 vs. 7.7) of DM group increased with the advancement of primary tumor depth stage and raising of ELN.
Conclusions:
DM was an independent risk factor for promoting LN metastasis. The preponderance of LN involvement in the DM group was aggravated with the advancement of tumor depth.
Keywords: Diabetes mellitus, Gastric cancer, Lymph node, Metastasis
Introduction
As the most common metastasis mode and most important independent prognostic factor of gastric cancer (GC),[1–3] lymph node (LN) involvement contributes a lot to the status. GC remains one of the leading causes of cancer-related deaths worldwide.[4,5] Unfortunately, LN dissection in GC has reached a plateau in improving the prognosis since D2 lymphadenectomy been accepted as the standard surgical procedure.[1,6,7] Even more, new research progress has shown that LN is an effective gateway from lymphatics into the systemic circulation, and becoming a source of cancer cells for distant metastasis.[8,9] Thus, determining the risk of LN metastasis in GC is significant to provide guidance for investigating the mechanism of LN metastasis in basic research and for clinical diagnosis and treatments.
It has also been noted that diabetes mellitus (DM) was closely linked to cancer epidemiologically and biological-ly.[10,11] The association between DM and cancer has got more and more attention as a result of the alarming increases in the prevalence of DM globally. And many studies have indicated that DM is related to an increased risk in many cancer types.[11,12] Some even indicated that DM conferred an inferior prognosis for cancer patients.[13,14] It has been demonstrated that breast cancer patients with DM experienced an increased risk of death compared with the non-DM subgroup.[15,16] Similar findings have also been reported in colon cancer,[17] prostate cancer,[18] and pancreatic neuroendocrine tumor.[19] More importantly, our previous study has shown that DM may promote the disease progress of GC after gastrectomy.[20]
However, whether DM was associated with LN metastasis in solid tumors has not been investigated yet. Thus, we further investigated the relationship between LN metastasis and DM in GC based on 2142 GC patients in our institution to explore the potential mechanisms of LN metastasis.
Methods
Ethical approval
All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions. The data collection protocol was approved by the Ethics Committee of Nanfang Hospital, Southern Medical University (No. NFEC-2018–088). Informed consent to be included in the study, or the equivalent, was obtained from all patients.
Patients selection
In the period between October 2004 and December 2019, 3213 consecutive patients were pathologically diagnosed with GC and underwent surgery at Nanfang Hospital, Southern Medical University, Guangdong, China. The analyses were based on the prospectively maintained GC database, which includes information on GC derived from electronic medical records that have been maintained in the Nanfang Hospital since 2004.[21] Data monitoring was conducted by a quality medical recorder with about 10 years of relevant work experience. The recorded variables included demographic, clinical, pathological, and surgical characteristics. Two independent surgical oncologists were requested to appraise the pathological reports and medical records of the patients retrospectively. After professional reviewing, patients who did not receive gastrectomy or lymphadenectomy and age <18 years were excluded. To make sure that the assessment for the LN status was accurate, the patients with the examined LNs (ELNs) <16 or who receive preoperative chemotherapy/radiotherapy were also excluded. After the above exclusion criteria were carried out, 2142 patients were enrolled. According to having the comorbidity of DM or not at the time of diagnosis of GC, patients were classified into two groups: non-DM group (n = 1993) and DM group (n = 149) [Figure 1].
Figure 1.
Study flow diagram of the study. GC: Gastric cancer; ELN: Examined LN; LN: Lymph node.
LN involvement was determined by two N-stage systems: the Japanese Classification of Gastric Carcinoma (JCGC, 13th edition) N-classification system[22] and Union International Cancer Control (UICC-TNM, 8th edition) N-staging system.[23] JCGC N-classification system is based on the anatomic location-based involved LNs, whereas the UICC-TNM N-staging system is based on the number of metastatic LN (MLN).[22,23] N3 stations of JCGC N-classification system was defined as LN involvement reached station 2, station 10, station 11 (11p and 11d), station 12 (12a, 12b, and 12p), station 13, and station 14 (LN at the root of the mesenterium) for gastric antrum cancer. And for the lesions in the body or cardia of stomach, N3 stations include station 12, station 13, and station 14 whereas N3b stage of UICC-TNM was defined >15 MLNs.
The T stage and M stage were determined based on the 8th edition of the AJCC TNM staging system.[22,23] The resection approach (laparoscopic gastrectomy [LG] or open gastrectomy [OG]) methods followed standard guidelines and our previous reports.[24–26] The LN examination approaches followed the experiences we reported.[27] The tumor location of the upper (U) 1/3, middle (M) 1/3, and lower (L) 1/3 was determined by the cancer epicenter.
Diagnosis of diabetes
The diagnosis of DM was based on the record of the prospective GC database,[21] which defined diabetes at the time of diagnosis of GC according to the following criteria: (1) Fasting plasma glucose (FPG) ≥126 mg/dL (7.0 mmol/ L). Fasting is defined as no caloric intake for at least 8 hours; (2) 2 hours plasma glucose 200 mg/dL (11.1 mmol/ L) during an oral glucose tolerance test (OGTT); (3) Patients with classic symptoms of hyperglycemia or hyperglycemic crisis, a random plasma glucose 200 mg/ dL (11.1 mm/L); or (4) Patients do not meet above conditions (blood test was normal before surgery) but have a specific history of diabetes with well-controlled medication. The exact time of the patients receiving the test of DM was during the process of preparation for undergoing surgery in Nanfang Hospital.
Statistical analysis
Continuous variables are presented as mean ± standard deviation (SD) (with normal distributions) or medians with interquartile ranges (IQR) (with non-normal distributions), while categorical variables were presented as counts and percentages (%). Student's t-test and the Mann–Whitney U test were used to compare continuous variables, and the χ2 test and Fisher exact test were used to compare categorical variables, as appropriate. Risk factors for N3 stations metastasis and N3b status were evaluated by univariate analyses using the χ2 test or Fisher exact test or Mann–Whitney U test, and multivariate analyses using logistic regression models. Variables with statistical significance (P < 0.01) in univariate analysis were entered into the multivariable model and were analyzed by using an “Enter” method. Risk factors for the number of MLN were evaluated by uni- and multi-variate analyses using linear regression models. Variables with statistical significance (P < 0.10) in the univariate analysis as well as the critical factor in the present analysis, diabetes, and tumor location were entered into the multivariable models and were analyzed by using an “ Enter” method P < 0.05 (two-tailed) was considered statistically significant. The statistical software SPSS version 25.0 for Windows (IBM Corp., New York, USA) was used for all statistical analyses.
Results
Characteristics of patients
The clinical and pathological characteristics of the patients are shown in Table 1. There were more males (77.2% [115/149] vs. 65.4% [1304/1993], P = 0.003) and older patients (63.83 ± 8.50 years vs. 55.94 ± 12.10 years, P < 0.001) in the DM group than the non-DM group. Tumor location was higher in the DM group (P = 0.016). Besides, the DM group was more prone to develop lymphatic invasions (34.9% [52/149] vs. 25.7% [513/1993], P = 0.014) and venous invasions (32.9% [49/149] vs. 25.5% [508/1993], P = 0.047). Notably, the primary tumor depth (mean rank: 1065.28 vs. 1071.97, Z = −0.132, P = 0.895) and the median [IQR] number of ELN (40 (22) vs. 39 (27), P = 0.735), the most important factors affecting the detection of MLN, were very similar between the DM and non-DM groups. Other clinical and pathological characteristics were also comparable between the two groups.
Table 1.
Clinicopathologic characteristics of diabetes and non-diabetes groups
| Characteristic | Non-diabetes (n = 1993) | Diabetes (n = 149) | Statistics | P value |
| Age, years | 55.94 ± 12.10 | 63.83 ± 8.50 | 10.552∗ | <0.001 |
| ELN | 39 (27) | 40 (22) | -0.338† | 0.735 |
| Gender | 8.563‡ | 0.003 | ||
| Male | 1304 (65.4) | 115 (77.2) | ||
| Female | 689 (34.6) | 34 (22.8) | ||
| History of abdominal surgery | 0.481‡ | 0.488 | ||
| No | 1675 (84.0) | 122 (81.9) | ||
| Yes | 318 (16.0) | 27 (18.1) | ||
| Primary tumor depth | -0.132† | 0.895 | ||
| T1a | 233 (11.7) | 18 (12.1) | ||
| T1b | 222 (11.1) | 11 (7.4) | ||
| T2 | 205 (10.3) | 17 (11.4) | ||
| T3 | 296 (14.9) | 31 (20.8) | ||
| T4a | 805 (40.4) | 54 (36.2) | ||
| T4b | 232 (11.6) | 18 (12.1) | ||
| Lymphatic invasions | 5.989‡ | 0.014 | ||
| No | 1480 (74.3) | 97 (65.1) | ||
| Yes | 513 (25.7) | 52 (34.9) | ||
| Venous invasions | 3.942‡ | 0.047 | ||
| No | 1485 (74.5) | 100 (67.1) | ||
| Yes | 508 (25.5) | 49 (32.9) | ||
| Nerve invasions | 0.059‡ | 0.809 | ||
| No | 1130 (56.7) | 86 (57.7) | ||
| Yes | 863 (43.3) | 63 (42.3) | ||
| Histology | 2.319‡ | 0.128 | ||
| Signet ring | 499 (25.0) | 29 (19.5) | ||
| Others | 1494 (75.0) | 120 (80.5) | ||
| Grade | 6.091‡ | 0.107 | ||
| G1 | 102 (5.1) | 10 (6.7) | ||
| G2 | 322 (16.2) | 35 (23.5) | ||
| G3 | 1553 (77.9) | 103 (69.1) | ||
| G4 | 16 (0.8) | 1 (0.7) | ||
| Tumor size, cm | 5.154‡ | 0.161 | ||
| <2 | 413 (20.7) | 23 (15.4) | ||
| 2–5 | 698 (35.0) | 49 (32.9) | ||
| ≥5 | 858 (43.1) | 73 (49.0) | ||
| Unknown | 24 (1.2) | 4 (2.7) | ||
| Tumor location | 10.292‡ | 0.016 | ||
| Upper 1/3 | 424 (21.3) | 48 (32.2) | ||
| Middle 1/3 | 333 (16.7) | 18 (12.1) | ||
| Lower 1/3 | 1217 (61.1) | 82 (55.0) | ||
| Unknown | 19 (1.0) | 1 (0.7) |
Data are shown as n (%), mean ± standard deviation or median (interquartile).∗ Student's t-test. † Mann–Whitney U test. ‡ χ2 test. ELN: Examined lymph node; IQR: Interquartile ranges; SD: Standard deviation.
LN status of DM and non-DM groups
As shown in Supplementary Table 1, the DM group had more N3 stations metastasis than the non-DM group (26.8% [40/149] vs. 19.3% [385/1993], P = 0.026). Equally, the DM group remarkably has more N3b status patients (18.8% [28/149] vs. 12.8% [256/1993], P = 0.039). Consistently, the N stage was slightly more advanced in the DM group (mean rank: 1126.40 vs. 1067.40, Z = -1.187, P = 0.235), although it has not yet reached significance. The median (IQR) MLN number was 2 (10.5) in DM group whereas 2 (9.0) in non-DM group.
The independent risk factors of N3 stations metastasis
As shown in Table 2, the univariate analyses showed that DM status (P = 0.026), sex (P = 0.034), age (P = 0.028), ELN (P < 0.001), primary tumor depth (P < 0.001), lymphatic invasion (P < 0.001), venous invasion (P < 0.001), nerve invasion (P < 0.001), grade (P < 0.001), tumor location (P = 0.014), and tumor size (P < 0.001) were associated with the detection of N3 stations metastasis. Subsequently, the multivariate analyses confirmed that DM (odds ratio [OR] = 1.771, 95% confidence interval [CI] = 1.139–2.755, P = 0.011), female (OR = 1.314, 95% CI = 1.018 -1.695, P = 0.036), younger age (40–69 vs. < 40 years: OR = 0.710, 95% CI = 0.513–0.983, P = 0.039; >69 vs. < 40 years: OR = 0.505, 95% CI = 0.273–0.935, P = 0.030), ELN ≥ 45 (OR = 1.618, 95% CI = 1.255–2.085, P < 0.001), more advanced primary tumor depth (T2 vs. T1a: OR = 2.627, 95% CI = 1.013–6.815, P = 0.047; T3 vs. T1a: OR = 4.222, 95% CI = 1.733–10.285, P = 0.002; T4a vs. T1a: OR = 6.575, 95% CI = 2.802–15.429, P < 0.001; T4b vs. T1a: OR = 10.743, 95% CI = 4.444–25.969, P < 0.001), positive lymphatic invasion (OR = 2.408, 95% CI = 1.770–3.278, P < 0.001), positive venous invasion (OR = 1.542, 95% CI = 1.142–2.081, P = 0.014), tumor size ≥5 cm (OR = 1.957, 95% CI = 1.145–3.344, P = 0.005), and tumor location in lower 1/3 (OR = 2.034,95% CI = 1.494–2.770, P < 0.001) were independent risk factors of developing N3 stations metastasis.
Table 2.
Univariate and multivariate analyses for the detection of N3 stations metastasis
| Univariate analysis | Multivariate analysis | |||||
| Variables | Non-N3 stations metastasis (n = 1717) | N3 stations metastasis (n = 425) | Statistic | P value | OR (95% CI) | P value |
| Diabetes | 4.940 | 0.026 | 0.011 | |||
| No | 1608 (93.7) | 385 (90.6) | Ref | |||
| Yes | 109 (6.3) | 40 (9.4) | 1.771 (1.139–2.755) | |||
| Gender | 4.516 | 0.034 | 0.036 | |||
| Male | 1156 (67.3) | 263 (61.9) | Ref | |||
| Female | 561 (32.7) | 162 (38.1) | 1.314 (1.018 -1.695) | |||
| Age, years | −2.197 | 0.028 | 0.045 | |||
| <40 | 256 (14.9) | 82 (19.3) | Ref | |||
| 40–69 | 1365 (79.5) | 323 (76.0) | 0.710 (0.513–0.983) | 0.039 | ||
| >69 | 96 (5.6) | 20 (4.7) | 0.505 (0.273–0.935) | 0.030 | ||
| ELN | 34.516 | <0.001 | <0.001 | |||
| 16–44 | 1080 (62.9) | 201 (47.3) | Ref | |||
| ≥45 | 637 (37.1) | 224 (52.7) | 1.618 (1.255–2.085) | |||
| Primary tumor depth | −13.114 | <0.001 | <0.001 | |||
| T1a | 244 (14.2) | 7 (1.6) | Ref | |||
| T1b | 226 (13.2) | 7 (1.6) | 0.950 (0.324–2.783) | 0.925 | ||
| T2 | 203 (11.8) | 19 (4.5) | 2.627 (1.013–6.815) | 0.047 | ||
| T3 | 258 (15.0) | 69 (16.2) | 4.222 (1.733–10.285) | 0.002 | ||
| T4a | 630 (36.7) | 229 (53.9) | 6.575 (2.802–15.429) | <0.001 | ||
| T4b | 156 (9.1) | 94 (22.1) | 10.743 (4.440–25.969) | <0.001 | ||
| Lymphatic invasions | 196.084 | <0.001 | <0.001 | |||
| No | 1378 (80.3) | 199 (46.8) | Ref | |||
| Yes | 339 (19.7) | 226 (53.2) | 2.408 (1.770–3.278) | |||
| Venous invasions | 128.843 | <0.001 | 0.005 | |||
| No | 1370 (79.8) | 215 (50.6) | Ref | |||
| Yes | 347 (20.2) | 210 (49.4) | 1.542 (1.142–2.081) | |||
| Nerve invasions | 110.850 | <0.001 | 0.783 | |||
| No | 1071 (62.4) | 145 (34.1) | Ref | |||
| Yes | 646 (37.6) | 280 (65.9) | 0959 (0.713–1.291) | |||
| Histology | 2.367 | 0.124 | – | |||
| Signet ring | 411 (23.9) | 117 (27.5) | ||||
| Others | 1306 (76.1) | 308 (72.5) | ||||
| Grade | 33.865 | <0.001 | 0.035 | |||
| G1 | 105 (6.1) | 7 (1.6) | Ref | |||
| G2 | 314 (18.3) | 43 (10.1) | 1.135 (0.459–2.810) | 0.784 | ||
| G3 | 1284 (74.8) | 372 (87.5) | 1.838 (0.784–4.309) | 0.162 | ||
| G4 | 14 (0.8) | 3 (0.7) | 0.904 (0.191–4.287) | 0.899 | ||
| Tumor size, cm | 132.099 | <0.001 | 0.005 | |||
| <2 | 411 (23.9) | 25 (5.9) | Ref | |||
| 2–5 | 637 (37.1) | 110 (25.9) | 1.233 (0.720–2.112) | 0.445 | ||
| ≥5 | 649 (37.7) | 282 (66.4) | 1.957 (1.145–3.344) | 0.014 | ||
| Unknown | 20 (1.2) | 8 (1.9) | 1.597 (0.568–4.488) | 0.374 | ||
| Location | 10.633 | 0.014 | <0.001 | |||
| Upper 1/3 | 393 (22.9) | 79 (18.6) | Ref | |||
| Middle 1/3 | 295 (17.2) | 56 (13.2) | 0.992 (0.654–1.502) | 0.968 | ||
| Lower 1/3 | 1012 (58.9) | 287 (67.5) | 2.034 (1.494–2.770) | <0.001 | ||
| Unknown | 17 (1.0) | 3 (0.7) | 0.894 (0.235–3.403) | 0.870 | ||
Data are shown as n (%). CI: Confidence interval; ELN: Examined LN; OR: Odds ratio; –: Not applicable.
The independent risk factors of N3b status
As shown in Table 3, the univariate analyses showed that DM status (P = 0.039), age (P = 0.019), ELN (P < 0.001), primary tumor depth (P < 0.001), lymphatic invasions (P < 0.001), venous invasions (P < 0.001), nerve invasions (P < 0.001), grade (P < 0.001), and tumor size(P < 0.001) were associated with the detection of N3b status. Subsequently, the multivariate analyses confirmed that DM (OR = 1.752, 95% CI = 1.061–2.893, P = 0.028), younger age (40–69 vs. < 40 years: OR = 0.669, 95% CI = 0.466–0.959, P = 0.029; >69 vs. <40 years: OR = 0.362, 95% CI = 0.168–0.782, P = 0.010), ELN ≥45 (OR = 1.841, 95% CI = 1.369–2.475, P < 0.001), more advanced primary tumor depth (T3 vs. T1a: OR = 11.035, 95% CI = 1.411–86.297, P = 0.022; T4a vs. T1a: OR = 27.918, 95% CI = 3.687–211.372, P = 0.001; T4b vs. T1a: OR = 45.174, 95% CI = 5.888–346.601, P < 0.001), positive lymphatic invasion (OR = 1.870, 95% CI = 1.303–2.683, P = 0.001), positive venous invasion (OR = 1.993, 95% CI = 1.408–2.821, P < 0.001), negative nerve invasion (OR = 0.625, 95% CI = 1.408–2.821, P = 0.008), grade of G3 (G3 vs. G1: OR = 7.592, 95% CI: 1.021–56.425, P = 0.048), and tumor size ≥5 cm (≥5 cm vs. <2 cm: OR = 2.248, 95% CI = 1.152–4.784, P = 0.019) were independent risk factors of developing N3b status.
Table 3.
Univariate and multivariate analyses for the detection of N3b status
| Univariate analysis | Multivariate analysis | |||||
| Variables | Non-N3b status (n = 1858) | N3b status (n = 284) | Statistic | P value | OR (95% CI) | P value |
| Diabetes | 4.263 | 0.039 | 0.028 | |||
| No | 1737 (93.5) | 256 (90.1) | Ref | |||
| Yes | 121 (6.5) | 28 (9.9) | 1.752 (1.061 -2.893) | |||
| Gender | <0.001 | 0.985 | − | |||
| Male | 1231 (66.3) | 188 (66.2) | ||||
| Female | 627 (33.7) | 96 (33.8) | ||||
| Age, years | -2.343 | 0.019 | 0.016 | |||
| <40 | 282 (15.2) | 56 (19.7) | Ref | |||
| 40–69 | 1470 (79.1) | 218 (76.8) | 0.669 (0.466–0.959) | 0.029 | ||
| >69 | 106 (5.7) | 10 (3.5) | 0.362 (0.168–0.782) | 0.010 | ||
| ELN | 26.807 | <0.001 | <0.001 | |||
| 16–44 | 1151 (61.9) | 130 (45.8) | ||||
| ≥45 | 707 (38.1) | 154 (54.2) | 1.841 (1.369–2.475) | |||
| Primary tumor depth | -13.121 | <0.001 | <0.001 | |||
| T1a | 250 (13.5) | 1 (0.4) | Ref | |||
| T1b | 230 (12.4) | 3 (1.1) | 2.961 (0.304–28.849) | 0.350 | ||
| T2 | 215 (11.6) | 7 (2.5) | 5.916 (0.691–50.674) | 0.105 | ||
| T3 | 296 (15.9) | 31 (10.9) | 11.035 (1.411–86.297) | 0.022 | ||
| T4a | 690 (37.1) | 169 (59.5) | 27.918 (3.687–211.372) | 0.001 | ||
| T4b | 177 (9.5) | 73 (25.7) | 45.174 (5.888–346.601) | <0.001 | ||
| Lymphatic invasions | 114.741 | <0.001 | 0.001 | |||
| No | 1442 (77.6) | 135 (47.5) | Ref | |||
| Yes | 416 (22.4) | 149 (52.5) | 1.870 (1.303–2.683) | |||
| Venous invasions | 128.843 | <0.001 | <0.001 | |||
| No | 1453 (78.2) | 132 (46.5) | Ref | |||
| Yes | 405 (21.8) | 152 (53.5) | 1.993 (1.408–2.821) | |||
| Nerve invasions | 46.858 | <0.001 | 0.008 | |||
| No | 1108 (59.6) | 108 (38.0) | Ref | |||
| Yes | 750 (40.4) | 176 (62.0) | 0.625 (0.441–0.885) | |||
| Histology | 1.768 | 0.184 | – | |||
| Signet ring | 449 (24.2) | 79 (27.8) | ||||
| Others | 1409 (75.8) | 205 (72.2) | ||||
| Grade | 35.544 | <0.001 | 0.002 | |||
| G1 | 111 (6.0) | 1 (0.4) | Ref | |||
| G2 | 333 (17.9) | 24 (8.5) | 3.552 (0.459–27.463) | 0.224 | ||
| G3 | 1399 (75.3) | 257 (90.5) | 7.592 (1.021–56.425) | 0.048 | ||
| G4 | 15 (0.8) | 2 (0.7) | 3.490 (0.281–43.331) | 0.331 | ||
| Tumor size, cm | 112.016 | <0.001 | 0.010 | |||
| <2 | 425 (22.9) | 11 (3.9) | Ref | |||
| 2–5 | 680 (36.6) | 67 (23.6) | 1.462 (0.705–3.030) | 0.307 | ||
| ≥5 | 731 (39.3) | 200 (70.4) | 2.348 (1.152–4.784) | 0.019 | ||
| Unknown | 22 (1.2) | 6 (2.1) | 2.088 (0.641–6.798) | 0.222 | ||
| Location | 0.719 | 0.873 | – | |||
| Upper 1/3 | 413 (22.2) | 59 (20.8) | ||||
| Middle 1/3 | 306 (16.5) | 45 (15.8) | ||||
| Lower 1/3 | 1121 (60.3) | 178 (62.7) | ||||
| Unknown | 18 (1.0) | 2 (0.7) | ||||
Data are shown as n (%). CI: Confidence interval; ELN: Examined LN; OR: Odds ratio; –: Not applicable.
The independent risk factors of more number of MLN
As shown in Table 4, the univariate linear analyses showed that age (as continuous variable) (β = −0.042, P = 0.015), ELN (β = 2.899, P < 0.001), primary tumor depth (β = 2.548, P < 0.001), lymphatic invasion (β = 7.276, P < 0.001), venous invasions (β = 7.009, P < 0.001), nerve invasions (β = 4.365, P < 0.001), grade (β = 2.776, P < 0.001), and tumor size (β = 4.078, P < 0.001) were associated with the detecting more number of MLN. Subsequently, the multivariate linear analyses (Durbin–Watson = 1.924) confirmed that DM (β = 1.424, P = 0.047), younger age (scale) (β = −0.059, P < 0.001), ELN ≥45 (β = 2.212, P < 0.001), more advanced primary tumor depth (β = 1.951, P < 0.001), positive lymphatic invasion (β = 3.855, P < 0.001), positive venous invasion (β = 2.844, P < 0.001), negative nerve invasion (β = −1.975, P < 0.001), grade of G3 (β = 0.854, P = 0.011), and tumor size ≥5 cm (β = 1.246, < 0.001) were independent risk factors of detection of more number of MLN. The histogram of regression standardized residual [Supplementary Figure 1] and normal P-P plot of regression standardized residual [Supplementary Figure 2] for linear regression model showed the residual is an approximate normal distribution.
Table 4.
Univariate and multiple linear regression analyses for the detection of the number of MLNs
| Univariate linear regression analysis | Multiple linear regression analysis∗ | |||||||||
| 95% CI for β | 95% CI for β | |||||||||
| Variables | Durbin–Watson | β | Lower bound | Upper bound | P value | VIF | β | Lower bound | Upper bound | P value |
| Diabetes (yes vs. no) | 1.949 | 1.377 | -0.236 | 2.990 | 0.094 | 1.037 | 1.424 | 0.019 | 2.829 | 0.047 |
| Gender (female vs. male) | 1.952 | 0.031 | -0.838 | 0.899 | 0.945 | Not included | – | – | – | – |
| Age (scale) | 1.956 | -0.042 | -0.077 | -0.008 | 0.015 | 1.007 | -0.059 | -0.089 | -0.028 | <0.001 |
| No. of ELN (≥45 vs.16–44) | 1.939 | 2.899 | 2.070 | 3.727 | <0.001 | 1.110 | 2.212 | 1.457 | 2.966 | <0.001 |
| Primary tumor depth (T4b vs. T4a vs. T3 vs. T2 vs. T1b vs. T1a) | 1.946 | 2.548 | 2.309 | 2.786 | <0.001 | 1.824 | 1.951 | 1.648 | 2.253 | <0.001 |
| Lymphatic invasions (yes vs. no) | 1.862 | 7.276 | 6.396 | 8.155 | <0.001 | 1.753 | 3.855 | 2.800 | 4.910 | <0.001 |
| Venous invasions (yes vs. no) | 1.920 | 7.009 | 6.121 | 7.897 | <0.001 | 1.647 | 2.844 | 1.817 | 3.871 | <0.001 |
| Nerve invasions (yes vs. no) | 1.923 | 4.365 | 3.558 | 5.173 | <0.001 | 1.574 | −1.975 | −0.284 | −1.086 | <0.001 |
| Histology (others vs. signet ring) | 1.951 | 0.387 | -0.565 | 1.340 | 0.425 | Not included | – | – | – | – |
| Grade (G4 vs. G3 vs. G2 vs. G1) | 1.944 | 2.776 | 2.053 | 3.498 | <0.001 | 1.097 | 0.854 | 0.198 | 1.510 | 0.011 |
| Tumor size (≥5 cm vs. 2–5 cm vs. <2 cm) | 1.979 | 4.078 | 3.588 | 4.568 | <0.001 | 1.711 | 1.246 | 0.664 | 1.827 | <0.001 |
| Location (lower 1/3 vs. middle 1/3 vs. upper 1/3) | 1.951 | 0.256 | -0.234 | 0.747 | 0.306 | 1.057 | 1.038 | 0.607 | 1.469 | <0.001 |
Durbin–Watson: 1.924. CI: Confidence interval; ELN: Examined LN; LNs: Lymph nodes; MLNs: Metastatic LNs; VIF: Variance inflation factor.
The LN status of DM and non-DM in subgroup divided by primary tumor depth
The variations of N3 stations metastasis divided by primary tumor depth in the DM and the non-DM groups were shown in Figure 2A. The proportion of N3 stations metastasis in each primary tumor depth stage was as follows (DM vs. non-DM): T1–2: 2.2% vs. 4.9%, T3: 29.0% vs. 20.3%, T4a: 38.9% vs. 25.8%, T4b: 50.0% vs. 36.6%. The variations of N3b status divided by primary tumor depth in the DM and non-DM groups were shown in Figure 2B. The proportion of N3b status in each primary tumor depth stage was as follows (DM vs. non-DM): T1–2: 0% vs. 1.7%, T3: 16.1% vs. 5.1%, T4a: 27.8% vs. 19.1%, T4b: 44.0% vs. 28.0%. The variations of the number of MLN divided by primary tumor depth in the DM and non-DM groups were shown in Figure 2C. The mean number of MLN in each primary tumor depth stage was as follows (DM vs. non-DM): T1–2: 0.4 vs. 1.1, T3: 8.6 vs. 5.2, T4a: 9.7 vs. 8.6, T4b: 17.0 vs. 12.8. The above result indicated that the percentage of N3 stations metastasis, N3b status, and the number of MLN increased with the advanced primary tumor depth stage. Thus, the subgroup analysis of T3+T4 was conducted to show the role of DM on tumor cells and tumor microenvironment with enough time. As shown in Supplementary Table 2 in the subgroup of T3+T4, the DM group has more N3 stations metastasis (DM vs. non-DM: 37.9% [39/103] vs. 26.5% [353/1333], P = 0.012). Equally, the DM group has more N3b status patients (DM vs. non-DM: 27.2% [28/ 103] vs. 18.4% [245/1333], P = 0.028). Similarly, the N stage was more advanced in the DM group (mean rank [DM vs. non-DM]: 782.1 vs. 713.6, Z = −1.710, P = 0.087) but has not yet reached significance. The number of MLN in the DM group was more than that in the non-DM group (median [IQR], 7 (14) vs. 5 (12), P = 0.066). Furthermore, the multivariate analyses confirmed that DM was an independent risk factor of N3 stations metastasis (b = 1.951, 95% CI: 1.226–3.104, P = 0.005) [Supplementary Table 3], N3b status (b = 1.930, 95% CI: 1.155–3.224, P = 0.012) [Supplementary Table 4], and more number of MLN (b = 2.082, Durbin–Watson = 1.918, Variance Inflation Factor [VIF] = 1.050, P = 0.040) [Supplementary Table 5] in the subgroup of T3+T4.
Figure 2.
(A) The N3 stations metastasis is divided by primary tumor depth in diabetes and non-diabetes groups. (B) The N3b status in the diabetes and non-diabetes groups is divided by primary tumor depth. (C)The number of MLNs is divided by primary tumor depth in the diabetes and non-diabetes groups. (D) The N3 stations metastasis is divided by the number of ELNs in the diabetes and non-diabetes groups. (E) The N3b status is divided by the number of ELNs in diabetes and non-diabetes groups. (F) The number of MLNs is divided by the number of ELNs in the diabetes and non-diabetes groups. ELNs: Examined lymph nodes; MLNs: Metastatic lymph nodes.
The LN status of DM and non-DM in subgroup divided by ELN
The variations of N3 stations metastasis divided by ELN in the DM and the non-DM groups were shown in Figure 2D. The proportion of N3 stations metastasis in each ELN was as follows (DM vs. non-DM): ELN16–29: 8.6% vs. 10.4%, ELN30–44: 27.9% vs. 20.5%, ELN≥45: 37.7% vs. 25.3%. The variations of N3b status divided by ELN in the DM and non-DM groups were shown in Figure 2E. The proportion of N3b status in each ELN was as follows (DM vs. non-DM): ELN16–29: 8.6% vs. 7.9%, ELN30–44: 18.0% vs. 11.8%, ELN≥45: 26.4% vs. 17.3%. The mean number of MLN in each ELN was as follows (DM vs. non-DM): ELN16–29: 3.6 vs. 4.6, ELN30–44: 5.8 vs. 5.8, ELN≥45: 12.0 vs. 7.7. The above results indicated that the percentage of N3 stations metastasis, N3b status, and the number of MLN increased with the raising of ELN. Thus, the analyses for the subgroup of ELN ≥45 were conducted to reflect the work of DM on LN status with the most exact and optimal effect. As shown in Supplementary Table 6 in the subgroup of ELN≥45, the DM group has more N3 stations metastasis (DM vs. non-DM: 37.7% [20/53] vs. 25.2% [204/808], P = 0.045). Equally, the N stage was significantly more advanced in the diabetes group (mean rank [DM vs. non-DM]: 507.5 vs. 425.9, Z = –2.458, P = 0.014). Similarly, the diabetes group has more N3b status (DM vs. non-DM: 26.4% [14/53] vs. 17.3% [140/808], P = 0.094) patients but has not yet reached significance. The number of MLN in the DM group was significantly more than that in the non-DM group (median [IQR], 8 [18.50] vs. 2 [10.75], P = 0.017). Furthermore, the multivariate analyses confirmed that DM was an independent risk factor of N3 stations metastasis (OR = 2.115, 95% CI: 1.063–4.209, P = 0.033), N3b status (OR = 1.647, 95% CI: 0.783–3.464, P = 0.189), and more MLN number (b = 3.529, Durbin–Watson = 1.913, VIF = 1.050, P = 0.021) in the subgroup of ELN≥45 [Supplementary Tables 7–9].
Discussion
Our study implied that DM promotes the LN metastasis in GC based on the analysis of 2142 GC patients between October 2004 and December 2019 in our institution. On the ground that the most important factors affecting the LN status, the status of ELM (P = 0.735), and the primary tumor depth (P = 0.895) were roughly equal between two groups, DM was still significantly associated with detecting more N3 stations metastasis and N3b status and more number of MLN. More importantly, the subsequent multivariate analyses confirmed that DM was a dependent factor of N3 stations metastasis and N3b status and more number of MLN. Furthermore, since primary tumor depth and ELN have been suggested as the most important factors affecting the LN status, thus the LN status in subgroups divided by primary tumor depth and ELN were performed. And the subgroups analyses showed the preponderance of N3 stations metastasis, N3b status, and the number of MLN in the DM group aggravated with the advancement of the primary tumor depth and the increase of ELN. The finding in our study indicated that DM promoted LN metastasis, which was revealed rarely.
As the most important independent prognostic factor for GC,[1–3] great emphasis was put on the LN involvement assessing, and currently, two N-stage systems were most recognized: the Japanese Classification of Gastric Carcinoma (JCGC, 13th edition) N-classification system and Union International Cancer Control (UICC-TNM, 8th edition) N-staging system. JCGC N-classification system is based on the anatomic location-based involved node stations, whereas the UICC-TNM N-staging system is based on the number of MLNs.[2,22,23] The JCGC defined N stage by the site of LN metastasis relative to the primary tumor and this classification of LN metastasis can reflect the metastatic pathways of cancer cells from the primary tumor. Although the TNM N-staging system is based on the number of MLN and could provide an accurate prognostic evaluation for GC patients. Our study evaluating not only the exact number of MLN but also the N3 stations metastasis with JCGC N-classification system and N3b status with UICC-TNM N-staging system. Thus, this study assessed the node status not only in LN burdens but also the metastatic extent, in other words, assessing both dimensions of LN burdens and involved distance and pathways. The multiple dimensions assessment in our study is much more comprehensively than the previous studies, which could only evaluate one dimension.
The novelty of evaluating the LN status by assessing the N3 stations metastasis, N3b status, and the number of MLN are as followed: (1) N3b status was first put forward by the 7th AJCC TNM staging system in 2014 and was incorporated into the 8th TNM stage in 2017 for the first time. According to the result of the International Gastric Cancer Association Project which reported the good prognostic discrimination of N3b from N3a, the 8th edition AJCC set a high value to the impact of N3b on the TNM stage. Even for early GC, the N3b node status (T1N3b) could upgrade patients into stage IIIB. There is no doubt that the N3b node status could have a great impact on the disease stage. Thus, the N3b status should be particularly evaluated to investigate the effect of DM on LN metastasis. (2) N3 stations metastasis assesses the LN status from another dimension that is different from N3b and the number of MLN and makes the evaluation more comprehensive. N3 stations metastasis reflects the extent of cancer cells from the primary tumor. Previous studies have demonstrated that insufficiency of ELN could result in inadequate detection of MLN in all stations. However, ELN in many centers was not satisfying that resulted in their relevant research could not present the exact LN status, especially the evaluation of each station and N3 stations was not accurate. Sano et al[3] collected data from 59 institutions present that the mean/median number of LNs examined in Japan, Korea, selected advantaged other Asian centers, and selected advantaged Western centers was 39.4/36, 33.0/31, 24.8/22, and 29.5/27, respectively. While the mean number of ELN of 2142 patients in our studies reached 42, which evaluate N3 stations metastasis more accurately and appropriately than most previous studies. (3) Previous studies assessing the risk factors of LN involvement almost roughly defined the LN metastasis as a bivariate variable (negative vs. positive) which may neglect important information of LN metastasis burden. Although our study defined one of the primary points as the number of MLN, which could comprehensively measure the burden of LN. The change was more accurate, specific, and reasonable to investigate the impact of DM on the progress of LNs metastasis in GC.
Another encouraging finding in this study is that the preponderance of N3 stations metastasis, N3b status, and the number of MLN in the DM group was aggravated with the advancement of the primary tumor depth. The mechanisms behind this phenomenon may include the following reasons. First, the longer process of developing a more advanced primary tumor stage allows the biological function of DM to act on tumor cells and tumor microenvironment with enough time. And some basic research has revealed some potential mechanisms that were consistent with the finding in this study. First, it has been found that multiple factors associated with DM, such as hyperglycemia, hyperinsulinemia and insulin-like growth factor I, dyslipidemia, adipokines and cytokines, and the gut microbiome, potentially give rise to the progression of cancer.[28] Hyperglycemia, the hallmark of DM, could provide tumor cells with more glucose and add to tumor proliferation and migration as a result that cancer cells are glucose avid and generate more energy from glycolysis and lactate production compared with oxidative phos[2p9h]orylation than noncancerous cells. Besides, Lin et al[29] suggested that hyperglycemia may account for the impairment of the immune system. Of course, the suppression of the immune system may result in the progress of LN metastasis. Furthermore, Wu et al[30] recently reveal the pathway linking diabetes to cancer via glucose-regulated phosphorylation of ten-eleven translocation-2 (TET2) by AMP-activated kinase (AMPK). They identify the tumor suppressor TET2 as a substrate of the AMPK and thereby stabilizing the tumor suppressor. Sustained hyperglycemia caused by DM destabilizes the tumor suppressor TET2 and deregulates levels of 5–hydroxymethylcytosine (5hmC), and ultimately diabetes to cancer. This study presented a new sight of how DM contributed to cancer, as well as offering the potential mechanism of how DM promoted LN metastasis of GC. Likewise, hyperinsulinemia and insulin receptor signaling could also drive tumor growth and thus may potentially activate the migration of GC cells to LNs. Although dyslipidemia, which is common in patients with DM, has also been found to affect the growth and progression of cancers.[30] It is also noted that adipose tissue is a vital organ for the production of adipokines, inflammatory cytokines, and enzymes that potentially lead to tumor promotion.[28] Thus, the changes in adipose tissue resulted from DM might contribute to the release of tumor progressing factors into the circulation. Also, changes in the gut microbiome caused by DM were indicated to have a link between cancer.[31] It is worth noting that the relationship between bacteria and cancer is also being revealed. It has been found that bacterial biofilms in the colon change the cancer metabolome to produce a regulator of cellular proliferation and colon cancer development potentially influencing cancer progression.[32] Likewise, recently, it has also been demonstrated that the tumor microbiome diversity, which cross-talks to the gut microbiome, significantly affected the prognosis of pancreatic cancer patients by modulating the tumor microbiome and affecting tumor development and host immune response.[33] Thus, the subgroup with T3–4 allows the greatest biological effect of DM on LN metastasis.
Another key issue is the superiority of LN involvement in the DM group was enhanced with the increase of ELN in our study. The leading superiority of N3 stations metastasis in the DM group increased with the adding of ELN. The consistent trend was discovered in the mean number of MLN. These amazing phenomenon may be attributed as following reasons: (1) as our previous study has shown that increase of ELN significantly improve the detection of more MLN and screen more patients with N3b node status; (2) for the subgroups with insufficient ELN, some N3 stations could even have not any ELN at all, not to mention investigating the LN involvement in N3 stations. Thus, deficient ELN hinders assessing the exact N3 stations metastasis to some extent. Therefore, the increase of ELN could reflect a more exact LN burden. Thus, the subgroup with ELN≥45 could show the effect of DM on LN metastasis more comprehensive and exact.
Considering the above reasons, the subgroups of T3–4and ELN≥45 were conducted to show the work of DM on MLN with minimal confounding factors. And consistent with the above hypothesis, the subgroup analysis showed DM patients have more heavily LN status burdens and widen LN involvement extent in the T3–4 and ELN≥45 subgroups than the whole group.
As for other variables that were also commonly assessed in other studies, the results in our study were consistent with those in other studies. A single institutional experience from China also shown that large tumor size, undifferentiated type, and lymphovascular invasion were independent risk factors for LNs metastasis in early GC.[34] Another multivariate logistic regression analysis also showed that tumor size, depth of invasion, and macroscopic types were independent factors that affect the LN metastasis.[35] Similarly, Li et al[36] recently showed that the depth of invasion, vascular tumor thrombus, and neural invasion were independent predictive risk factors for LN metastasis in GC patients. Except for the clinical characteristics, the number of ELN has also been confirmed to be associated with the detecting of the number of metastasis LN in GC.[27] The consistency of our results with that in the previous studies[37] indicated that the design of our study was reasonable, our data were reliable and the analysis of the general linear regression model was calculable. Thus, the results in our study that DM significantly promoted the LN metastasis in GC was also credible and worth subsequent exploring.
Encouragingly, the finding that DM significantly promoted LN metastasis in GC may provide a reference for subsequent clinical researches and treatments for GC. As we knew, postoperative adjuvant chemotherapy based on histological and biological characteristics to determine tailored treatment for specific subgroups could further improve the prognosis of GC patients.[38] Although LN metastasis has been demonstrated to be the strongest predictor of disease recurrence for GC[1–3] and even might be the active hubs for systemic tumor cell spread.[8,9] At the same time, it has been demonstrated that more intensive adjuvant chemotherapy was safe and could significantly improve prognosis for the subgroup GC patients with an indication for intense adjuvant chemotherapy in serious trials in Japan, including the Adjuvant Chemotherapy Trial of S-1 for Gastric Cancer study,[39] the Japan Clinical Oncology Group 1104 trial,[40] and the Japan Clinical Cancer Research Organization GC-07 trial.[41] Thus, since GC patients with DM are more prone to develop high-LN-burden and involve widen extent, the tailored perioperative management for them specifically may be worth being explored.
Also, the specific mechanisms that DM could promote LN metastasis should be specific and exactly studied so that corresponding target drugs could be used to relieved LN metastasis. Furthermore, the finding in this study gives clues for the orientation of the mechanisms of LN metastasis.
There are still some limitations to our study. Although the data in our study were prospectively collected,[21] our study was not prospectively designed but retrospectively analyzed. And as a result of the nonprospective design, the clinical characteristic was not comparable, such as age and sex. Also, to make sure the assessment for the LN status is more accurate to compensate for the inherent limitations of retrospective analysis, our study excluded the patients with the number of ELN < 16 or receive preoperative chemotherapy/radiotherapy. In addition, the size of the metastasis LN in each group was not registered in our database, so we could not investigate whether DM affects the size of metastasis LN, which may be a variable to reflect the metastasis LN burden. Therefore, the size of the metastasis LNs should be taken into consideration in the design of subsequent randomized control trials.
Conclusions
DM was an independent risk factor for promoting LN metastasis, especially for subgroups of T3–4 and ELN≥45, which allow the biological function of DM to act on tumor cell and tumor microenvironment with enough period and reflect the exact LN status. The preponderance of N3 stations metastasis, N3b status, and MLN in the DM group was aggravated with the advancement of primary tumor depth and the increase of ELN. This finding provides a reference for subsequent clinical researches and treatments for GC and gives clues for the orientation of the mechanisms of LN metastasis in basic research.
Acknowledgment
We would like to thank Xia Cheng for prospectively maintaining our GC database.
Funding
The current study was supported by the Guangdong Provincial Key Laboratory of Precision Medicine for Gastrointestinal Cancer (No.2020B121201004), Guangdong Provincial Major Talents Project (No.2019JC05 Y361), the Science and Technology Planning Project of Guangdong Province (No. 2017B020226005), the “Climbing Program”, Special Fund of Guangdong Province (No. pdjh2021b0098), National Natural Science Foundation of China (No. 81902444) and Science and Technology Program of Guangzhou (No. 201903010072).
Conflicts of interests
None.
Supplementary Material
Supplementary Material
Footnotes
How to cite this article: Chen X, Chen Y, Li T, Liang W, Huang H, Su H, Sui C, Hu Y, Chen H, Lin T, Chen T, Zhao L, Liu H, Li G, Yu J. Diabetes mellitus promoted lymph node metastasis in gastric cancer: a 15–year single-institution experience. Chin Med J 2022;135:950–961. doi: 10.1097/CM9.0000000000001795
Supplemental digital content is available for this article.
References
- 1.Ke B, Liang H. Current status of lymph node dissection in gastric cancer. Chin J Cancer Res 2021; 33:193–202. doi: 10.21147/j.issn.1000-9604.2021.02.07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Deng JY, Liang H. Clinical significance of lymph node metastasis in gastric cancer. World J Gastroenterol 2014; 20:3967–3975. doi: 10.3748/wjg.v20.i14.3967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Sano T, Coit DG, Kim HH, Roviello F, Kassab P, Wittekind C, et al. Proposal of a new stage grouping of gastric cancer for TNM classification: International Gastric Cancer Association staging project. Gastric Cancer 2017; 20:217–225. doi: 10.1007/s10120-016-0601-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Yang L, Ying X, Liu S, Lyu G, Xu Z, Zhang X, et al. Gastric cancer: epidemiology, risk factors and prevention strategies. Chin J Cancer Res 2020; 32:695–704. doi: 10.21147/j.issn.1000-9604.2020.06.03. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zhang T, Chen H, Yin X, He Q, Man J, Yang X, et al. Changing trends of disease burden of gastric cancer in China from 1990 to 2019 and itspredictions: findings from global burden of disease study. Chin J Cancer Res 2021; 33:11–26. doi: 10.21147/j.issn.1000-9604.2021.01.02. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Chen J, Bu Z, Ji J. Surgical treatment of gastric cancer: current status and future directions. Chin J Cancer Res 2021; 33:159–167. doi: 10.21147/j.issn.1000-9604.2021.02.04. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.National Health Commission of the People's Republic of China. Chinese guidelines for diagnosis and treatment of gastric cancer 2018. Chin J Cancer Res 2019; 31:707–737. doi: 10.21147/j.issn.1000-9604.2019.05.01. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Brown M, Assen FP, Leithner A, Abe J, Schachner H, Asfour G, et al. Lymph node blood vessels provide exit routes for metastatic tumor cell dissemination in mice. Science 2018; 359:1408–1411. doi: 10.1126/science.aal3662. [DOI] [PubMed] [Google Scholar]
- 9.Pereira ER, Kedrin D, Seano G, Gautier O, Meijer EFJ, Jones D, et al. Lymph node metastases can invade local blood vessels, exit the node, and colonize distant organs in mice. Science 2018; 359:1403–1407. doi: 10.1126/science.aal3622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Giovannucci E, Harlan DM, Archer MC, Bergenstal RM, Gapstur SM, Habel LA, et al. Diabetes and cancer: a consensus report. Diabetes Care 2010; 33:1674–1685. doi: 10.2337/dc10-0666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Shi Y, Hu FB. The global implications of diabetes and cancer. Lancet 2014; 383:1947–1948. doi: 10.1016/S0140-6736(14)60886-2. [DOI] [PubMed] [Google Scholar]
- 12.Seshasai SRK, Kaptoge S, Thompson A, Di Angelantonio E, Gao P, Sarwar N, et al. Diabetes mellitus, fasting glucose, and risk of cause-specific death. N Engl J Med 2011; 364:829–841. doi: 10.1056/NEJMoa1008862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Barone BB, Yeh HC, Snyder CF, Peairs KS, Stein KB, Derr RL, et al. Long-term all-cause mortality in cancer patients with preexisting diabetes mellitus: a systematic review and meta-analysis. JAMA 2008; 300:2754–2764. doi: 10.1001/jama.2008.824. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Campbell PT, Newton CC, Patel AV, Jacobs EJ, Gapstur SM. Diabetes and cause-specific mortality in a prospective cohort of one million U.S. adults. Diabetes Care 2012; 35:1835–1844. doi: 10.2337/dc12-0002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Srokowski TP, Fang S, Hortobagyi GN, Giordano SH. Impact of diabetes mellitus on complications and outcomes of adjuvant chemotherapy in older patients with breast cancer. J Clin Oncol 2009; 27:2170–2176. doi: 10.1200/JCO.2008.17.5935. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Peairs KS, Barone BB, Snyder CF, Yeh HC, Stein KB, Derr RL, et al. Diabetes mellitus and breast cancer outcomes: a systematic review and meta-analysis. J Clin Oncol 2011; 29:40–46. doi: 10.1200/JCO.2009.27.3011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Meyerhardt JA, Catalano PJ, Haller DG, Mayer RJ, Macdonald JS, Benson AR, et al. Impact of diabetes mellitus on outcomes in patients with colon cancer. J Clin Oncol 2003; 21:433–440. doi: 10.1200/JCO.2003.07.125. [DOI] [PubMed] [Google Scholar]
- 18.Bensimon L, Yin H, Suissa S, Pollak MN, Azoulay L. Type 2 diabetes and the risk of mortality among patients with prostate cancer. Cancer Causes Control 2014; 25:329–338. doi: 10.1007/s10552-013-0334-6. [DOI] [PubMed] [Google Scholar]
- 19.Vernieri C, Pusceddu S, de Braud F. Impact of metformin on systemic metabolism and survival of patients with advanced pancreatic neuroendocrine tumors. Fron Oncol 2019; 9:902.doi: 10.3389/fonc.2019.00902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chen X, Chen Y, Li T, Jun L, Lin T, Hu Y, et al. Impact of diabetes on prognosis of gastric cancer patients performed with gastrectomy. Chin J Cancer Res 2020; 32:631–644. doi: 10.21147/j.issn.1000-9604.2020.05.08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hu YF, Yu J, Zhang C, Wang YN, Cheng X, Huang F, et al. Development and implementation of a clinical data mining system for gastric cancer surgery (in Chinese). Chin J Gastrointest Surg 2010; 13:510–515. doi: 10.3760/cma.j.issn.1671-0274.2010.07.012. [PubMed] [Google Scholar]
- 22.Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma, 2nd English edition. Gastric Cancer 1998; 1:10–24. doi: 10.1007/s101209800016. [DOI] [PubMed] [Google Scholar]
- 23.Amin MB, Greene FL, Edge SB, Compton CC, Gershenwald JE, Brookland RK, et al. The eighth edition AJCC cancer staging manual: continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin 2017; 67:93–99. doi: 10.3322/caac.21388. [DOI] [PubMed] [Google Scholar]
- 24.Japanese Gastric Cancer Association. Japanese gastric cancer treatment guidelines 2014 (ver. 4). Gastric Cancer 2017; 20:1–19. doi: 10.1007/s10120-016-0622-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Smyth EC, Verheij M, Allum W, Cunningham D, Cervantes A, Arnold D, et al. Gastric cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol 2016; 27:v38–v49. doi: 10.1093/annonc/mdw350. [DOI] [PubMed] [Google Scholar]
- 26.Ettinger DS, Kuettel M, Malin J, McClure JS, Smith ML, Zelenetz AD, et al. NCCN roundtable: what are the characteristics of an optimal clinical practice guideline? J Natl Compr Canc Netw 2015; 13: (5 suppl): 640–642. doi: 10.6004/jnccn.2015.0190. [DOI] [PubMed] [Google Scholar]
- 27.Chen X, Chen Y, Hu Y, Lin T, Luo J, Li T, et al. The methods of lymph node examination make a difference to node staging and detection of N3b node status for gastric cancer. Front Oncol 2020; 10:123.doi: 10.3389/fonc.2020.00123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Gallagher EJ, LeRoith D. Obesity and diabetes: the increased risk of cancer and cancer-related mortality. Physiol Rev 2015; 95:727–748. doi: 10.1152/physrev.00030.2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lin SW, Freedman ND, Hollenbeck AR, Schatzkin A, Abnet CC. Prospective study of self-reported diabetes and risk of upper gastrointestinal cancers. Cancer Epidemiol Biomarkers Prev 2011; 20:954–961. doi: 10.1158/1055-9965.EPI-10-1244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wu D, Hu D, Chen H, Shi G, Fetahu IS, Wu F, et al. Glucose-regulated phosphorylation of TET2 by AMPK reveals a pathway linking diabetes to cancer. Nature 2018; 559:637–641. doi: 10.1038/s41586-018-0350-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Sheflin AM, Whitney AK, Weir TL. Cancer-promoting effects of microbial dysbiosis. Curr Oncol Rep 2014; 16:406.doi: 10.1007/s11912-014-0406-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Johnson CH, Dejea CM, Edler D, Hoang LT, Santidrian AF, Felding BH, et al. Metabolism links bacterial biofilms and colon carcinogenesis. Cell Metab 2015; 21:891–897. doi: 10.1016/j.cmet.2015.04.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Riquelme E, Zhang Y, Zhang L, Montiel M, Zoltan M, Dong W, et al. Tumor microbiome diversity and composition influence pancreatic cancer outcomes. Cell 2019; 178:795–806. doi: 10.1016/j.cell.2019.07.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Chen J, Zhao G, Wang Y. Analysis of lymph node metastasis in early gastric cancer: a single institutional experience from China. World J Surg Oncol 2020; 18:57.doi: 10.1186/s12957-020-01834-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Zhang Y, Zhu Z, Sun Z, Wang Z, Zheng X, Xu H. Preoperative predicting score of lymph node metastasis for gastric cancer. Tumour Biol 2014; 35:10437–10442. doi: 10.1007/s13277-014-2363-5. [DOI] [PubMed] [Google Scholar]
- 36.Li L, Liu P, Wang J, Niu X, He C. Clinicopathologic characteristics and risk factors of lymph node metastasis in patients with early gastric cancer in the Wannan region. Med Sci Monit 2020; 26:e923525.doi: 10.12659/MSM.923525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Nie Z, Zhu H, Gu M. Reduced colorectal cancer incidence in type 2 diabetic patients treated with metformin: a meta-analysis. Pharm Biol 2016; 54:2636–2642. doi: 10.1080/13880209.2016.1176057. [DOI] [PubMed] [Google Scholar]
- 38.Chen X, Liu H, Li G, Yu J. Implications of clinical research on adjuvant chemotherapy for gastric cancer: where to go next? Chin J Cancer Res 2019; 31:892–900. doi: 10.21147/j.issn.1000-9604.2019.06.05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Sasako M, Sakuramoto S, Katai H, Kinoshita T, Furukawa H, Yamaguchi T, et al. Five-year outcomes of a randomized phase III trial comparing adjuvant chemotherapy with S-1 versus surgery alone in stage II or III gastric cancer. J Clin Oncol 2011; 29:4387–4393. doi: 10.1200/JCO.2011.36.5908. [DOI] [PubMed] [Google Scholar]
- 40.Yoshikawa T, Terashima M, Mizusawa J, Nunobe S, Nishida Y, Yamada T, et al. Four courses versus eight courses of adjuvant S-1 for patients with stage II gastric cancer (JCOG1104 [OPAS-1]): an open-label, phase 3, non-inferiority, randomised trial. Lancet Gastroenterol Hepatol 2019; 4:208–216. doi: 10.1016/S2468-1253(18)30383-2. [DOI] [PubMed] [Google Scholar]
- 41.Yoshida K, Kodera Y, Kochi M, Ichikawa W, Kakeji Y, Sano T, et al. Addition of docetaxel to oral fluoropyrimidine improves efficacy in patients with stage III gastric cancer: interim analysis of JACCRO GC-07, a randomized controlled trial. J Clin Oncol 2019; 37:1296–1304. doi: 10.1200/JCO.18.01138. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.