Comparison of Survival and Patterns of Recurrence in ...

Original Investigation | Gastroenterology and Hepatology

Comparison of Survival and Patterns of Recurrence in Gastric NeuroendocrineCarcinoma, Mixed Adenoneuroendocrine Carcinoma, and AdenocarcinomaJianxian Lin, MD; Yajun Zhao, MD; Yanbing Zhou, MD; Yantao Tian, MD; Qingliang He, MD; Junpeng Lin, MD; Hankun Hao, MD; Bingbing Zou, MD;Lixin Jiang, MD; Gang Zhao, MD; Wei Lin, MD; Yanchang Xu, MD; Zhi Li, MD; Fangqin Xue, MD; Shuliang Li, MD; Weihua Fu, MD; Yongxiang Li, MD;Zekuan Xu, MD; Yong Li, MD; Jinping Chen, MD; Xiaojun Zhou, MD; Zhenggang Zhu, MD; Lisheng Cai, MD; En Li, MD; Honglang Li, MD; Chaohui Zheng, PhD;Ping Li, PhD; Changming Huang, MD; Jianwei Xie, PhD

Abstract

IMPORTANCE Gastric neuroendocrine carcinoma and mixed adenoneuroendocrine carcinoma arerare pathological types of gastric cancer, and there is a lack of multicenter studies comparing theprognosis and recurrence patterns of gastric neuroendocrine carcinoma, gastric mixedadenoneuroendocrine carcinoma, and gastric adenocarcinoma.

OBJECTIVE To compare the differences in long-term survival and patterns of recurrence amonggastric neuroendocrine carcinoma, gastric mixed adenoneuroendocrine carcinoma, and gastricadenocarcinoma.

DESIGN, SETTING, AND PARTICIPANTS This cohort study included patients with resectable gastricneuroendocrine carcinoma and gastric mixed adenoneuroendocrine carcinoma at 23 hospitals inChina from January 2006 to December 2016. In addition, patients with gastric adenocarcinoma wereselected as controls. Propensity score–matched analysis was used to match pathological stageamong the different pathological types, and disease-free survival (DFS), postrecurrence survival(PRS), and patterns of recurrence were examined. Data analysis was conducted from July 15, 2020,to October 21, 2020.

EXPOSURES Curative resection for gastric neuroendocrine carcinoma, gastric mixedadenoneuroendocrine carcinoma, and gastric adenocarcinoma.

MAIN OUTCOMES AND MEASURES The main outcomes were DFS and patterns of recurrence.

RESULTS A total of 3689 patients were analyzed (median [interquartile range] age, 62 [55-69]years; 2748 [74.5%] men), including 503 patients (13.6%) with gastric neuroendocrine carcinoma,401 patients (10.9%) with gastric mixed adenoneuroendocrine carcinoma, and 2785 patients(75.5%) with gastric adenocarcinoma. After propensity score matching, 5-year DFS was 47.6% (95%CI, 42.7%-52.5%) for patients with gastric neuroendocrine carcinoma, compared with 57.6% (95%CI, 55.1%-60.1%) with gastric adenocarcinoma (P < .001) and 51.1% (95% CI, 46.0%-56.2%) forpatients with gastric mixed adenoneuroendocrine carcinoma, compared with 57.8% (95% CI, 55.1%-60.5%) patients with gastric adenocarcinoma (P = .02). Multivariable analyses found that, comparedwith gastric adenocarcinoma, gastric neuroendocrine carcinoma (hazard ratio [HR], 1.64; 95% CI,1.40-1.93) and gastric mixed adenoneuroendocrine carcinoma (HR, 1.25; 95% CI, 1.05-1.49) wereindependent risk factors associated with worse DFS. Compared with matched patients with gastricadenocarcinoma, patients with gastric neuroendocrine carcinoma were more likely to have distantrecurrence (268 patients [17.2%] vs 101 patients [23.7%]; P = .002), as were patients with gastricmixed adenoneuroendocrine carcinoma (232 patients [17.3%] vs 76 patients [22.8%]; P = .02). In

(continued)

Key PointsQuestion Are there any differences in

prognoses or recurrence patterns

associated with gastric neuroendocrine

carcinoma, mixed

adenoneuroendocrine carcinoma, or

adenocarcinoma?

Findings This cohort study included

3689 patients with resectable gastric

adenocarcinoma, gastric

neuroendocrine carcinoma, or gastric

mixed adenoneuroendocrine

carcinoma. Propensity score matching

analysis found that, compared with

gastric adenocarcinoma, gastric

neuroendocrine carcinoma and gastric

mixed adenoneuroendocrine carcinoma

were independent risk factors

associated with worse disease-free

survival, postrecurrence survival, and

distant recurrence outcomes.

Meaning These findings suggest that

different follow-up and treatment

strategies should be developed to

improve the long-term survival of

patients with gastric neuroendocrine

carcinoma or gastric mixed

adenoneuroendocrine carcinoma,

especially for patients with tumors

penetrating into the subserosa or

deeper layers and with lymph node

metastasis.

+ Supplemental content

Author affiliations and article information arelisted at the end of this article.

Open Access. This is an open access article distributed under the terms of the CC-BY License.

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 1/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

Abstract (continued)

multivariate analysis, gastric neuroendocrine carcinoma (HR, 2.22; 95% CI, 1.66-2.98) and gastricmixed adenoneuroendocrine carcinoma (HR, 1.70; 95% CI, 1.24-2.34) were independent risk factorsassociated with distant recurrence. Additionally, T3 to T4 stage (odds ratio, 2.84; 95% CI, 1.57-5.14;P = .001) and lymph node metastasis (odds ratio, 2.01; 95% CI, 1.31-3.10; P = .002) were independentrisk factors associated with distant recurrence of gastric neuroendocrine carcinoma and gastricmixed adenoneuroendocrine carcinoma.

CONCLUSIONS AND RELEVANCE This cohort study found that patients with gastricneuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma had worse prognosesand were more prone to distant recurrence than those with gastric adenocarcinoma. Thus, differentfollow-up and treatment strategies should be developed to improve the long-term survival ofpatients with gastric neuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma,especially patients with tumors penetrating into the subserosa or deeper layers or with lymph nodemetastasis.

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180

Introduction

Gastric cancer, as a common cancer, is one of the main causes of cancer-related deaths worldwide,1

and adenocarcinoma is a common pathological type. However, gastric neuroendocrine neoplasm is arare type of gastric cancer. In 2011, the World Health Organization (WHO)2 classified neuroendocrineneoplasms of the stomach into 3 categories: neuroendocrine tumor, neuroendocrine carcinoma, andmixed adenoneuroendocrine carcinoma, in which gastric neuroendocrine carcinoma consists ofpoorly differentiated, high-grade malignant neoplasms, while gastric mixed adenoneuroendocrinecarcinoma contains at least 30% each of epithelial and neuroendocrine components. As a highlyinvasive tumor, gastric neuroendocrine carcinoma tends to present in late stages and to beassociated with lymph node metastasis.3,4 Previous studies have shown that the prognoses forgastric neuroendocrine carcinoma and gastric mixed adenoneuroendocrine carcinoma are worsethan that of gastric adenocarcinoma, but most of those studies were single-center studies with smallsample sizes.5-7

Currently, follow-up strategies for gastric neuroendocrine carcinoma and gastric mixedadenoneuroendocrine carcinoma are similar to those for gastric adenocarcinoma.8 Understandingthe patterns of recurrence is important in designing follow-up and treatment strategies.9,10 However,previous studies on the recurrence pattern of gastric cancer have mainly focused on gastricadenocarcinoma.10-12 The patterns of recurrence and postrecurrence survival for patients with gastricneuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma are still unknown andhave not been compared with those of patients with gastric adenocarcinoma, to our knowledge. Thepurpose of this study was to compare the survival and recurrence patterns among patients withresectable gastric neuroendocrine carcinoma, gastric mixed adenoneuroendocrine carcinoma, orgastric adenocarcinoma by using a multicenter, large sample series.

Methods

This study was performed in accordance with the Declaration of Helsinki13 and the Ethical Guidelinesfor Clinical Studies, and was approved by the institutional review boards of all participating centers.Written informed consent or a substitute (eg, assent from a caretaker) for it was obtained from allpatients for inclusion in the study. This study followed the Strengthening the Reporting ofObservational Studies in Epidemiology (STROBE) reporting guideline.

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 2/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

This multicenter retrospective cohort study was conducted in patients with gastricneuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma without metastaticdisease who underwent potentially curative resection in 23 centers of the China GastricNeuroendocrine Tumor Study Group from January 2006 to December 2016. All the centers includedin this study are large- and medium-sized teaching hospitals in China that have rich experience in thediagnosis and treatment of gastric cancer. The annual number of gastric cancer surgeries in eachcenter is more than 300 surgeries per year. The main exclusion criteria were patients who underwentcompletion gastrectomy, patients with incomplete data (including tumor size and follow-upinformation), and patients who experienced postoperative death within 30 days. Patients withremnant gastric cancer, unknown tumor differentiation, or unknown follow-up data and patientswho died within 30 days after surgery were also excluded. Patients who died within 1 month ofsurgery were excluded because death within 30 days is usually not from recurrent disease butrelated to complications.14

The diagnostic criteria for gastric neuroendocrine carcinoma and gastric mixedadenoneuroendocrine carcinoma were based on the WHO 2011 classification.2 In each center, thepostoperative pathological findings were confirmed by 2 experienced upper gastrointestinalpathologists. All surgical procedures, including D2 lymphadenectomy, were performed according tothe Japanese gastric cancer treatment guidelines.15 Adjuvant chemotherapy was recommended forpatients with advanced gastric neuroendocrine carcinoma, gastric mixed adenoneuroendocrinecarcinoma, or gastric adenocarcinoma.8,16

Patient Characteristics and Follow-upDemographic and clinicopathological characteristics and treatment of the study population weredetermined by review of the database and of medical records. The pathological stages werereevaluated by pathologists according to the American Joint Committee on Cancer (AJCC)guidelines.17 The median number of patients with gastric neuroendocrine carcinoma or gastric mixedadenoneuroendocrine carcinoma in each center was used as the cutoff value to define high-volume(�30 patients) or low-volume (<30 patients) centers.

The last follow-up time was December 2019. Postoperative follow-up was performed every 3months for 2 years and then every 6 months in years 3 to 5, and every year for after 5 years. Mostpatients routinely received physical examinations, laboratory tests, chest radiography, abdominalultrasonography or computed tomography, and an annual endoscopic examination.

Outcomes and Pattern of RecurrenceThe primary outcomes of the study were disease-free survival (DFS) and patterns of recurrence. DFSwas defined as the time interval from the date of the operation to the date of recurrence or deathwith evidence of recurrence. Overall survival (OS) was calculated from the date of surgery to the dateof last contact or death. Postrecurrence survival (PRS) was defined as the time from first recurrenceto either death or the last follow-up,12,18 and patients who experienced recurrence and died in thesame month were censored at the evaluation of PRS.

The recurrence was diagnosed with radiologic findings or biopsies of suspicious lesions duringsurveillance after surgery,19 which were recorded from the date of initial surgery to the date of thefirst recurrence, death of any cause, or last follow-up, whichever occurred first.20 As previouslydescribed,10,21 recurrences were categorized according to the site involved as locoregional,peritoneal, distant, or multiple. The presence of recurrent disease in 2 or more sites was categorizedas multiple recurrence. Multiple recurrence in the same area (eg, anastomotic and gastric bed) wascoded in a single category.

Statistical AnalysisPropensity score matching was used to adjust the imbalance of potential confounding factors amongdifferent pathological types. Two propensity score models were generated for pairwise comparisons

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 3/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

(including gastric neuroendocrine carcinoma vs gastric adenocarcinoma and gastric mixedadenoneuroendocrine carcinoma vs gastric adenocarcinoma). Tumor staging is an importantpathological factor associated with the prognosis of gastric neuroendocrine carcinoma, gastric mixedadenoneuroendocrine carcinoma, and gastric adenocarcinoma,22,23 while age and sex are commonclinical factors associated with the prognosis of a tumor.24,25 Thus, propensity scores were calculatedusing a logistic regression model, including pathological TNM stage, age, and sex, for all patients.Then, a 1 to 4 matching procedure according to the propensity score, without replacement, wasundertaken using the nearest-neighbor method within a caliper of 0.01. Propensity score matchingwas performed using the matching package in R version 3.5.2 (R Project for Statistical Computing).26

Continuous variables are expressed as the medians and interquartile ranges (IQRs), andcategorical variables are expressed as numbers and percentages. Differences among groups incategorical variables were analyzed using the χ2 test or Fisher exact test, whereas differences incontinuous variables were evaluated using t tests or Mann-Whitney U tests. Survival rates andmedian survival outcomes were estimated with Kaplan-Meier curves. The log-rank test wasperformed for comparisons between groups. The incidence of recurrence and the differentproportions of recurrent sites between 2 groups were evaluated by the χ2 test or Fisher exacttest.19,27 A Cox proportional hazards regression model was used to identify the independentprognostic factors associated with DFS. The proportional hazard was evaluated by proportionalhazards assumption test and Schoenfeld residual test.28 A logistic regression model was used toidentify the risk factors associated with distant recurrence. Variables with P < .05 in the univariableanalysis were included in the multivariable model, and forward likelihood ratio method is used foranalysis.29 All tests were 2-sided, and statistical significance was inferred at P < .05. Statisticalanalyses were performed using SPSS statistical software version 22.0 (IBM). Data were analyzedfrom July 15, 2020, to October 21, 2020.

Results

Baseline Characteristics and Survival Outcomes From the Unmatched DataA total of 3689 patients (median [IQR] age, 62 [55-69] years; 2748 [74.5%] men) were included inthe analysis, including 503 patients with gastric neuroendocrine carcinoma (13.6%), 401 patientswith gastric mixed adenoneuroendocrine carcinoma (10.9%), and 2785 patients with gastricadenocarcinoma (75.5%) (eFigure 1 in the Supplement). There were 12 high-volume centers, whichtreated a total of 846 patients (22.9%), and 11 low-volume centers, which treated a total of 58patients (0.2%) (eTable 1 in the Supplement). The remaining 2785 patients (75.5%) were patientswith gastric adenocarcinoma and all came from 1 center. Compared with patients with gastricadenocarcinoma, patients with gastric neuroendocrine carcinoma or gastric mixedadenoneuroendocrine carcinoma patients were older (median [IQR] age, 61 [54-68] years vs 64[58-69] and 64 [58-70] years), more likely to be men (2035 [73.1%] men vs 400 men [79.5%] and313 [78.1%] men), and more likely to have tumors that were located in the upper stomach (637patients [22.9%] vs 279 patients [55.5%] and 173 patients [43.1%]), were larger in size (median [IQR]size, 4.0 [2.5-6.0] cm vs 4.5 [3.0-6.0] cm and 4.5 [3.0-6.0] cm ), had more vascular invasion (798patients [28.7%] vs 203 patients [49.0%] and 128 patients [34.3%]), and have more advanceddisease stages (Table 1). In terms of treatment, compared with patients with gastricadenocarcinoma, the proportion of total gastrectomy in patients with gastric neuroendocrinecarcinoma or gastric mixed adenoneuroendocrine carcinoma was higher (1457 patients [52.3%] vs291 patients [57.9%] and 231 patients [57.6%]; P < .001), but the number of lymph nodes retrievedwas lower (median [IQR], 33 [26-43] nodes vs 21 [14-30] nodes and 22 [15-31] nodes; P < .001). Therewere statistically significant differences between patients with gastric neuroendocrine carcinomaand those with gastric mixed adenoneuroendocrine carcinoma in terms of tumor location, number oflymph node metastasis, N stage, and vascular and nerve invasion, but there were no significantdifferences in other clinicopathological factors (Table 1).

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 4/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

The median (range) follow-up time was 58.0 (1.0-153.0) months for patients with gastricneuroendocrine carcinoma, 56 (1.2-152.0) for patients with gastric mixed adenoneuroendocrinecarcinoma, and 70.0 (1.5-121.0) months for patients with gastric adenocarcinoma. The incidence ofrecurrence was 239 patients with gastric neuroendocrine carcinoma (47.5%) and 184 patients withgastric mixed adenoneuroendocrine carcinoma (45.9%), which were significantly higher than theincidence of recurrence in patients with gastric adenocarcinoma (793 patients [28.5%]; P < .001)(eFigure 2 in the Supplement). Survival curves showed that the DFS and OS of patients with gastricneuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma were significantlyworse than those of patients with gastric adenocarcinoma, while the PRSs of patients with gastric

Table 1. Baseline Clinicopathologic Characteristics of Included Patients

Characteristic

No. (%) P value

AC (n = 2785) NEC (n = 503) MANEC (n = 401) NEC vs AC MANEC vs AC NEC vs MANECAge, median (IQR), ya 61 (54-68) 64 (58-69) 64 (58-70) <.001 <.001 .89

Sexb

Men 2035 (73.1) 400 (79.5) 313 (78.1).002 .03 .59

Women 750 (26.9) 103 (20.5) 88 (21.9)

Tumor locationb

Lower 1223 (43.9) 117 (23.3) 115 (28.7)

<.001 <.001 .002Middle 607 (21.8) 76 (15.1) 86 (21.4)

Upper 637 (22.9) 279 (55.5) 173 (43.1)

Mix 209 (8.1) 31 (6.2) 27(6.7)

Tumor size, median (IQR), cma 4.0 (2.5-6.0) 4.5 (3.0-6.0) 4.5 (3.0-6.0) <.001 <.001 .82

Lymph nodes examined, median (IQR), No.a 33 (26-43) 21 (14-30) 22 (15-31) <.001 <.001 .14

Metastatic lymph nodes, median (IQR), No.a 2 (0-8) 2 (0-6) 3 (0-7) .53 .006 .04

Invasionb,c

Vascular 798 (28.7) 203 (49.0) 128 (34.3) <.001 .02 <.001

Neural 761 (27.3) 145 (35.8) 103 (27.8) <.001 .86 .02

T stageb

T1 758 (27.2) 29 (5.8) 36 (9.0)

<.001 <.001 .05

T2 320 (11.5) 48 (9.5) 43 (10.7)

T3 917 (32.9) 88 (17.5) 61 (15.2)

T4a 760 (27.3) 328 (65.2) 260 (64.8)

T4b 30 (1.1) 10 (2.0) 1 (0.2)

N stageb

N0 1073 (38.5) 147 (29.2) 108 (26.9)

<.001 <.001 .03

N1 447 (16.1) 113 (22.5) 85 (21.2)

N2 490 (17.6) 139 (27.6) 98 (24.4)

N3a 495 (17.8) 87 (17.3) 78 (19.5)

N3b 280 (10.1) 17 (3.4) 32 (8.0)

TNM stageb

I 862 (31.0) 47 (9.3) 48 (12.0)

<.001 <.001 .38II 659 (23.7) 144 (28.6) 105 (26.2)

III 1264 (45.4) 312 (62.0) 248 (61.8)

Underwent gastrectomyb

Total 1457 (52.3) 291 (57.9) 231 (57.6).02 .047 .94

<Total 1328 (47.7) 212 (42.1) 170 (42.4)

Received chemotherapy

Neoadjuvant yb 61 (2.2) 18 (3.6) 11 (2.7) .06 .49 .48

Adjuvantb,d 1645 (59.1) 283 (63.2) 226 (61.2) .10 .42 .57

Abbreviations: AC, adenocarcinoma; IQR, interquartile range; MANEC, mixedadenoneuroendocrine carcinoma; NEC, neuroendocrine carcinoma.a Compared by using Mann-Whitney U tests.b Compared by using the χ2 test.

c Vascular invasion data were missing for 117 patients, and neural invasion data weremissing for 128 patients.

d Adjuvant chemotherapy of 87 patients was unknown.

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 5/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

neuroendocrine carcinoma and patients with gastric mixed adenoneuroendocrine carcinoma weresimilar to that of patients with gastric adenocarcinoma patients (eFigure 3 in the Supplement). Therecurrence pattern of gastric neuroendocrine carcinoma was similar to that of gastric mixedadenoneuroendocrine carcinoma, and both were more prone to distant recurrence than gastricadenocarcinoma (eFigure 4 in the Supplement).

Gastric Neuroendocrine Carcinoma vs Gastric Adenocarcinoma in the Matched DataAfter propensity score matching, matched data sets were developed for 500 patients with gastricneuroendocrine carcinoma and 1573 patients with gastric adenocarcinoma. There were no significantdifferences in stage, age, or sex between patients with gastric neuroendocrine carcinoma and thosewith gastric adenocarcinoma (eTable 2 in the Supplement).

SurvivalIn the matched data, the incidence of recurrence in gastric neuroendocrine carcinoma wassignificantly higher than that in gastric adenocarcinoma (237 patients [47.4%] vs 568 patients[36.1%]; P < .001) (eFigure 2 in the Supplement). Among patients with gastric neuroendocrinecarcinoma, 5-year DFS was 47.6% (95% CI, 42.7%-52.5%), and 5-year OS was 50.5% (95% CI,45.6%-55.4%); and for patients with gastric adenocarcinoma, 5-DFS was 57.6% (95% CI,55.1%-60.1%; P < .001), and 5-year OS was 60.5% (95% CI, 58.0%-63.0%; P < .001) (Figure, A and

Figure. Kaplan-Meier Survival Curves for Patients With Different Histologic Subtypes of Gastric Carcinoma in the Matched Data

1.0

0.8

0.6

0.4

0.2

0

No. at risk

0

1573500

843201

36

640148

48 72

490110

60

Prob

abili

ty o

f sur

viva

l

Time after operation, mo

1303361

12 24

1107283

ACNEC

Disease-free survival for gastric NEC vs ACA

P <.001

1.0

0.8

0.6

0.4

0.2

0

No. at risk

0

1573500

1051220

36

777162

48 72

581117

60

Prob

abili

ty o

f sur

viva

l

Time after operation, mo

1439413

12 24

1245325

ACNEC

Overall survival for gastric NEC vs ACB

P <.001

1.0

0.8

0.6

0.4

0.2

0

No. at risk

0

1355394

715181

36

550134

48 72

41993

60

Prob

abili

ty o

f sur

viva

l

Time after operation, mo

1122301

12 24

952245

ACNEC

Disease-free survival for gastric MANEC vs ACC

P =.02

1.0

0.8

0.6

0.4

0.2

0

No. at risk

0

1355394

900192

36

658142

48 72

50394

60

Prob

abili

ty o

f sur

viva

l

Time after operation, mo

1237339

12 24

1069270

ACNEC

Overall survival for gastric MANEC vs ACD

P =.002

AC

AC AC

MANEC MANEC

AC

NEC NEC

AC indicates adenocarcinoma; MANEC, mixed adenoneuroendocrine carcinoma; and NEC, neuroendocrine carcinoma.

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 6/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

B). In the univariable analyses, gastric neuroendocrine carcinoma (vs gastric adenocarcinoma) wasassociated with reduced DFS (hazard ratio [HR], 1.36; 95% CI, 1.17-1.57; P < .001) (Table 2).Multivariable analysis confirmed that gastric neuroendocrine carcinoma (vs gastric adenocarcinoma)was an independent risk factor associated with worse DFS (HR, 1.64; 95% CI, 1.40-1.93; P < .001)(Table 2). Even compared with poorly differentiated gastric adenocarcinoma, gastric neuroendocrine

Table 2. Univariable and Multivariable Cox Regression Analyses of Factors Associated with Disease-Free Survival in the Matched Data Sets of PatientsWith AC vs Patients With NEC

Clinicopathological features

Univariable analysis Multivariable analysis Ia Multivariable analysis IIb

HR (95% CI) P value HR (95% CI)c P value HR (95% CI)c P valueAge, per 1-y increase 1.02 (1.01-1.03) <.001 1.02 (1.01-1.03) <.001 1.02 (1.01-1.03) <.001

Sex

Men 1 [Reference].03

NA.36

NA.36

Women 0.83 (0.71-0.99) NA NA

Center volume

Low 1 [Reference].26

NA NA NA NA

High 0.77 (0.49-1.21) NA NA NA NA

Tumor location

Lower 1 [Reference] NA NA NA NA NA

Middle 1.10 (0.91-1.33) .32 NA .14 NA .14

Upper 1.21 (1.04-1.42) .02 NA .50 NA .50

Mix 1.59 (1.29-1.96) <.001 NA .45 NA .45

Tumor size, per 1-cm increase 1.16 (1.14-1.19) <.001 1.06 (1.03-1.09) <.001 1.06 (1.03-1.09) <.001

Lymph nodes examined, per 1-unit increase 1.00 (0.99-1.00) .25 NA NA NA NA

No. of metastatic lymph nodes, per 1-unit increase 1.06 (1.06-1.07) <.001 1.05 (1.04-1.06) <.001 1.05 (1.04-1.06) <.001

Invasion

Vascular 1.72 (1.50-1.96) <.001 NA .93 NA .93

Neural 1.72 (1.51-1.97) <.001 1.24 (1.08-1.43) .002 1.24 (1.08-1.43) .002

T stage

T1-T2 1 [Reference]<.001

1 [Reference]<.001

1 [Reference]<.001

T3-T4 3.84 (3.02-4.90) 2.00 (1.54-2.59) 2.00 (1.54-2.59)

N stage

N0 1 [Reference]<.001

1 [Reference]<.001

1 [Reference]<.001

N1-N3 2.91 (2.40-3.54) 1.51 (1.22-1.86) 1.51 (1.22-1.86)

Type of gastrectomy

Total 1 [Reference]<.001

NA.61

NA.61

<Total 0.70 (0.61-0.80) NA NA

Received neoadjuvant chemotherapy

No 1 [Reference]<.001

1 [Reference].001

1 [Reference].001

Yes 1.92 (1.36-2.72) 1.78 (1.26-2.53) 1.78 (1.26-2.53)

Received adjuvant chemotherapy

No 1 [Reference] NA NA NA NA NA

Yes 1.12 (0.97-1.29) .13 NA NA NA NA

Unknown 0.96 (0.63-1.47) .86 NA NA NA NA

Pathological type I

AC 1 [Reference]<.001

1 [Reference]<.001

NA NA

NEC 1.36 (1.17-1.57) 1.64 (1.40-1.93) NA NA

Pathological type II

Poorly differentiated AC 1 [Reference] NA NA NA 1 [Reference] NA

Well or moderately differentiated AC 0.72 (0.62-0.84) <.001 NA NA 1.00 (0.85-1.17) .99

NEC 1.18 (1.01-1.38) .04 NA NA 1.64 (1.38-1.96) <.001

Abbreviations: AC, adenocarcinoma; HR, hazard ratio: NA, not applicable; NEC,neuroendocrine carcinoma.a Multivariable analysis I included pathological type I and excluded pathological type II.

b Multivariable analysis II included pathological type II, excluding pathological type I.c Because multivariate analyses used a forward likelihood ratio method, there are no

corresponding HR data for factors with P > .05.

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 7/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

carcinoma was significantly associated with worse DFS (HR, 1.64; 95%CI, 1.38-1.96; P < .001)(Table 2). The median (IQR) PRS of propensity score–matched patients with gastric neuroendocrinecarcinoma was 8 (4-15) months, compared with 6 (3-14) months among those with gastricadenocarcinoma (P = .08) (eFigure 5 in the Supplement).

Stratified analysis by center volumes showed that in high-volume centers, patients with gastricneuroendocrine carcinoma had lower DFS and OS than patients with gastric adenocarcinoma, whilethe PRS of patients with gastric neuroendocrine carcinoma was similar to that of patients with gastricadenocarcinoma (eFigure 6 in the Supplement). Owing to the small number of patients inlow-volume centers, we could not further analyze and compare the prognosis between groups.

Patterns of RecurrenceIn the matched data, 805 patients experienced recurrent disease, of whom 715 (88.8%) had a knownsite of recurrence. Most patients had initial recurrence involving only a single site (gastricneuroendocrine carcinoma: 143 patients [87.8%]; gastric adenocarcinoma: 446 patients [80.9%]),and 20 patients with gastric neuroendocrine carcinoma (12.2%) and 105 patients with gastricadenocarcinoma (19.1%) had multiple recurrence (eFigure 7 in the Supplement). The detailed sites ofrecurrence are shown in Table 3. The incidence of distant recurrence in gastric neuroendocrinecarcinoma was significantly higher than that in gastric adenocarcinoma (101 patients [23.7%] vs 268patients [17.2%]; P = .002), especially in liver metastasis (61 patients [14.3%] vs 97 patients [6.2%];P < .001) (Table 3). However, there was no significant difference between patients with gastricneuroendocrine carcinoma vs those with gastric adenocarcinoma in terms of locoregionalrecurrence, peritoneal recurrence, or multiple recurrence (Table 3). Multivariable analysis revealedthat gastric neuroendocrine carcinoma (vs gastric adenocarcinoma) was an independent risk factorassociated with distant recurrence (odds ratio [OR], 2.22; 95% CI, 1.66-2.98) (eTable 3 in theSupplement).

Table 3. Specific Sites of Recurrence in the Matched Data Sets

Variable

AC vs NEC AC vs MANECAC(n = 1556)

NEC(n = 427) P valuea

AC(n = 1340)

MANEC(n = 333) P valuea

Locoregional recurrence, No. (%) 88 (5.7) 22 (5.2) .69 74 (5.5) 12 (3.6) .16

Anastomosis 43 (2.8) 16 (3.7) .29 34 (2.5) 6 (1.8) .43

Regional node 31 (2.0) 4 (0.9) .14 31 (2.3) 2 (0.6) .07

Gastric or tumor bed 22 (1.4) 2 (0.5) .11 16 (1.2) 4 (1.2) >.99

Recurrence, No. (%)

Peritoneal 90 (5.8) 20 (4.7) .38 83 (6.2) 16 (4.8) .34

Distant 268 (17.2) 101(23.7) .002 232 (17.3) 76 (22.8) .02

Distant lymph node 120 (7.7) 21 (4.9) .05 105 (7.8) 32 (9.6) .29

Liver 97 (6.2) 61 (14.3) <.001 81 (6.0) 42 (12.6) <.001

Lung 44 (2.8) 13 (3.0) .81 42 (3.1) 5 (1.5) .11

Bone 39 (2.5) 8 (1.9) .45 29 (2.2) 4 (1.2) .26

Brain 6 (0.4) 2 (0.5) >.99 3 (0.2) 2 (0.6) .57

Pancreas 16 (1.0) 3 (0.7) .74 12 (0.9) 4 (1.2) .84

Spleen 7 (0.4) 0 .36 8 (0.6) 0 .37

Adrenal 11 (0.7) 3 (0.7) >.99 8 (0.6) 0 .37

Colorectum 9 (0.6) 0 .22 5 (0.4) 0 .59

Mediastinum and pleura 12 (0.8) 0 .14 10 (0.7) 0 .23

Bile duct 5 (0.3) 0 .59 4 (0.3) 0 >.99

Subcutaneous 9 (0.6) 2 (0.5) >.99 8 (0.6) 0 .37

Otherb 7 (0.4) 0 .36 7 (0.5) 0 .36

Multiple, No. (%) 105 (6.7) 20 (4.7) .12 85 (6.3) 15 (4.5) .21

Abbreviations: AC, adenocarcinoma; MANEC, mixedadenoneuroendocrine carcinoma; NEC,neuroendocrine carcinoma.a Calculated by using χ2 test or Fisher exact test.b Other sites included jejunum, ureter, esophagus,

testis, muscle, and wound.

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 8/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

Gastric Mixed Adenoneuroendocrine Carcinoma vs Gastric Adenocarcinomain the Matched DataAfter propensity score matching, 394 patients with gastric mixed adenoneuroendocrine carcinomaand 1355 patients with gastric mixed adenoneuroendocrine carcinoma formed the matched datasets. There were no significant differences in stage, age, or sex between patients with gastric mixedadenoneuroendocrine carcinoma vs those with gastric adenocarcinoma (eTable 2 in theSupplement).

SurvivalIn the matched data, the incidence of recurrence was 180 patients with gastric mixedadenoneuroendocrine carcinoma (45.7%) and 489 patients with gastric adenocarcinoma (36.1%)(P < .001) (eFigure 2 in the Supplement). Kaplan-Meier survival curves are shown in the Figure. TheDFS and OS of patients with gastric neuroendocrine carcinoma were significantly worse than those ofpatients with gastric mixed adenoneuroendocrine carcinoma (5-year DFS: 51.1% [95% CI, 46.0%-56.2%] vs 57.8% [95% CI, 55.1%-60.5%]; P = .02; 5-year OS: 52.5% [95% CI, 47.2%-57.8%] vs60.8% [95% CI, 57.3%-64.3%]; P = .002) (Figure, C and D). However, there was no significantdifference in median PRS between the matched gastric mixed adenoneuroendocrine carcinomagroup and gastric adenocarcinoma group (eFigure 5 in the Supplement). In the univariable analyses,gastric mixed adenoneuroendocrine carcinoma (vs gastric adenocarcinoma) was significantlyassociated with reduced DFS (Table 4). Multivariable analysis confirmed that gastric neuroendocrinecarcinoma (vs gastric adenocarcinoma) was an unfavorable risk factor associated with worse DFS(HR, 1.25; 95%CI, 1.05-1.49) (Table 4). Similarly, gastric mixed adenoneuroendocrine carcinoma wasstill associated with worse DFS compared with poorly differentiated gastric adenocarcinoma (HR,1.27; 95% CI, 1.06-1.54) (Table 4).

Stratified analysis confirmed that the DFS and OS of patients with gastric mixedadenoneuroendocrine carcinoma were significantly worse than those of patients with gastricadenocarcinoma in the high-volume centers. The PRS of patients with gastric mixedadenoneuroendocrine carcinoma was similar to that of patients with gastric adenocarcinoma(eFigure 8 in the Supplement).

Patterns of RecurrenceA total of 669 patients experienced recurrent disease in the matched data, of whom 593 (88.6%)had a known site of recurrence. The rate of recurrence involving only a single area was 104 patientswith gastric mixed adenoneuroendocrine carcinoma (87.3%) and 389 patients with gastricadenocarcinoma (82.0%) (eFigure 7 in the Supplement). The incidence of distant recurrence inpatients with gastric mixed adenoneuroendocrine carcinoma was significantly higher than that inpatients with gastric adenocarcinoma (76 patients [22.8%] vs 232 patients [17.3%]; P = .02),especially in liver metastasis (42 patients [12.6%] vs 81 patients [6.0%]; P < .001) (Table 3). However,there was no significant difference between gastric mixed adenoneuroendocrine carcinoma andgastric adenocarcinoma in other patterns of recurrence (Table 3). Multivariable analysis confirmedthat gastric mixed adenoneuroendocrine carcinoma (vs gastric adenocarcinoma) was anindependent risk factor associated with distant recurrence (OR, 1.70; 95% CI, 1.24-2.34) (eTable 4 inthe Supplement).

Pathological Factors Associated With Distant Recurrence of Gastric NeuroendocrineCarcinoma and Gastric Mixed Adenoneuroendocrine CarcinomaIn this study, 178 patients with gastric neuroendocrine carcinoma or gastric mixedadenoneuroendocrine carcinoma experienced distant recurrence. In univariable analyses, T3-T4stage, lymph node metastasis, Ki-67 expression greater than 20%, and tumors located in the middleof the stomach were associated with increased risk of distant recurrence (eTable 5 in theSupplement). However, compared with gastric neuroendocrine carcinoma, gastric mixed

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 9/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

adenoneuroendocrine carcinoma was not associated with increased distant recurrence (eTable 5 inthe Supplement). In multivariable analyses, only T3-T4 stage (OR, 2.84; 95% CI, 1.57-5.14; P = .001)and lymph node metastasis (OR, 2.01; 95% CI, 1.31-3.10; P = .002) were independent risk factorsassociated with distant recurrence of gastric neuroendocrine carcinoma and gastric mixedadenoneuroendocrine carcinoma (eTable 5 in the Supplement).

Table 4. Univariable and Multivariable Cox Regression Analyses of Factors Associated With Disease-Free Survival in the Matched Data Sets of PatientsWith AC vs Patients With MANEC

Clinicopathological features

Analysis

Univariable Multivariable Ia Multivariable IIb

HR (95% CI) P value HR (95% CI)c P value HR (95% CI)c P valueAge 1.02 (1.01-1.03) <.001 1.02 (1.01-1.03) <.001 1.02 (1.01-1.03) <.001

Sex

Men 1 [Reference].31

NA NA NA NA

Women 0.91 (0.76-1.09) NA NA NA NA

Center volume

Low 1 [Reference].049

NA.27

NA.26

High 0.56 (0.32-1.00) NA NA

Location

Lower 1 [Reference] NA NA NA NA NA

Middle 1.06 (0.87-1.30) .58 NA .12 NA .11

Upper 1.24 (1.04-1.47) .02 NA .45 NA .43

Mix 1.70 (1.36-2.12) <.001 NA .31 NA .49

Tumor size, per 1-cm increase 1.15 (1.12-1.18) <.001 NA .05 1.03 (1.00-1.06) .05

Lymph nodes examined, per 1-cm increase 1.00 (1.00-1.01) .62 NA NA NA NA

No. of metastatic lymph nodes, per 1-unit increase 1.06 (1.06-1.07) <.001 1.05 (1.04-1.06) <.001 1.05 (1.04-1.06) <.001

Invasion

Vascular 1.74 (1.51-2.01) <.001 NA .51 NA .52

Neural 1.83 (1.58-2.12) <.001 1.26 (1.09-1.47) .002 1.28 (1.10-1.49) .001

T stage

T1-T2 1 [Reference]<.001

1 [Reference]<.001

1 [Reference]<.001

T3-T4 4.47 (3.43-5.83) 2.62 (1.98-3.48) 2.48 (1.86-3.31)

N stage

N0 1 [Reference]<.001

1 [Reference]<.001

1 [Reference]<.001

N1-N3 3.33 (2.67-4.16) 1.58 (1.24-2.01) 1.56 (1.23-1.99)

Type of gastrectomy

Total 1 [Reference]<.001

NA.54

NA.32

Subtotal 0.75 (0.65-0.87) NA NA

Received neoadjuvant chemotherapy

No 1 [Reference].003

1 [Reference].003

NA NA

Yes 1.90 (1.25-2.87) 1.86 (1.23-2.83) NA NA

Received adjuvant chemotherapy

No 1 [Reference] NA NA NA NA NA

Yes 1.117 (0.968-1.286) .13 NA NA NA NA

Unknown 0.987 (0.529-1.244) .89 NA NA NA NA

Pathological type I

AC 1 [Reference].02

1 [Reference].01

NA NA

MANEC 1.21 (1.03-1.43) 1.25 (1.05-1.49) NA NA

Pathological type II

Poorly differentiated AC 1 [Reference] NA NA NA 1 [Reference] NA

Well or moderately differentiated AC 0.73 (0.62-0.86) <.001 NA NA 1.07 (0.90-1.27) .47

MANEC 1.05 (0.88-1.25) .59 NA NA 1.27 (1.06-1.54) .01

Abbreviations: AC, adenocarcinoma; HR, hazard ratio; MANEC, mixedadenoneuroendocrine carcinoma; NA, not applicable.a Multivariate Analysis I included pathological type I, excluding pathological type II.

b Multivariate Analysis II included pathological type II, excluding pathological type I.c Because multivariate analyses used a forward likelihood ratio method, there are no

corresponding HR data for factors with P > .05.

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 10/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

Discussion

To our knowledge, this cohort study, including 3689 patients in 23 centers in China, is the first andlargest multicenter retrospective study comparing the outcomes and patterns of recurrence ingastric neuroendocrine carcinoma, gastric mixed adenoneuroendocrine carcinoma, and gastricadenocarcinoma. This study found that patients with gastric neuroendocrine carcinoma or gastricmixed adenoneuroendocrine carcinoma had worse prognoses than those with gastricadenocarcinoma and that gastric neuroendocrine carcinoma and gastric mixedadenoneuroendocrine carcinoma were independent risk factors associated with worse DFS and PRS.In addition, patients with gastric neuroendocrine carcinoma or gastric mixed adenoneuroendocrinecarcinoma were more likely to experience distant recurrence. Our results could provide importantmedical evidence for the development of follow-up and treatment strategies for patients with gastricneuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma.

Owing to the rarity of gastric neuroendocrine carcinoma and gastric mixedadenoneuroendocrine carcinoma, large reports with comparisons of prognosis among these patientsand those with gastric adenocarcinoma are lacking. A 2017 single-center study in South Korea6 foundthat patients with gastric neuroendocrine carcinoma had worse survival than those withconventional gastric cancer, but there were significant differences in stage between patients withgastric neuroendocrine carcinoma and those with conventional gastric cancer. In our study, patientswith gastric neuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma also hadlater stages and worse prognoses than those with gastric adenocarcinoma, which was consistentwith previous studies.5-7 Thus, propensity score matching was used to balance the differences instage. In the matched data sets, we found that the DFS and OS of patients with gastricneuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinoma were also worse thanthose of patients with gastric adenocarcinoma. Nearly half of patients with gastric neuroendocrinecarcinoma or gastric mixed adenoneuroendocrine carcinoma experienced recurrence, but theprognosis of patients with recurrent gastric neuroendocrine carcinoma or gastric mixedadenoneuroendocrine carcinoma is still unknown. Therefore, more active and effectivemultidisciplinary treatments should be undertaken for patients with gastric neuroendocrinecarcinoma or gastric mixed adenoneuroendocrine carcinoma after surgery to improve thesurvival time.

Poorly differentiated adenocarcinoma is a pathological type of gastric cancer associated withworse prognosis.30,31 To our knowledge, only 1 single-center study6 has reported that the recurrence-free survival rate of patients with poorly differentiated gastric adenocarcinoma was lower than thatof patients with gastric neuroendocrine carcinoma, but reports comparing survival between patientswith poorly differentiated gastric adenocarcinoma and those with gastric mixedadenoneuroendocrine carcinoma are still lacking. In our study, after adjusting for a number of factorsassociated with worse prognosis, multivariable analysis found that gastric neuroendocrine carcinomaand gastric mixed adenoneuroendocrine carcinoma were associated with a worse prognosis thanpoorly differentiated gastric adenocarcinoma, suggesting that gastric neuroendocrine carcinoma andgastric mixed adenoneuroendocrine carcinoma are more malignant pathological types than poorlydifferentiated gastric adenocarcinoma.

The analysis and understanding of tumor recurrence patterns are conducive to optimizingfollow-up and treatment strategies.9 However, to our knowledge, there have been no previousstudies on the recurrence patterns of gastric neuroendocrine carcinoma and gastric mixedadenoneuroendocrine carcinoma. In this study, the rates of locoregional recurrence and peritonealrecurrence of gastric neuroendocrine carcinoma and gastric mixed adenoneuroendocrine carcinomawere similar to those of gastric adenocarcinoma. However, nearly three-fourths of patients withrecurrent gastric neuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinomaexperienced distant recurrence, and the incidence of recurrence was significantly higher for thesepatients than that in patients with gastric adenocarcinoma. The difference in distant recurrence

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 11/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

mainly concerned liver metastasis. In the multivariate model, we further found that gastricneuroendocrine carcinoma and gastric mixed adenoneuroendocrine carcinoma were independentrisk factors associated with distant recurrence. For gastric neuroendocrine carcinoma and gastricmixed adenoneuroendocrine carcinoma, patients with T3 to T4 stage and lymph node metastasiswere at high risk of postoperative distant recurrence. Thus, in future follow-up work, appropriateexaminations should be performed for patients with high-risk gastric neuroendocrine carcinoma orgastric mixed adenoneuroendocrine carcinoma, such as contrast-enhanced computed tomographyor magnetic resonance imaging, to improve the detection rate of distant recurrence. Adjuvanttherapies focusing on distant recurrence should be used and developed to improve the long-termsurvival of patients with gastric neuroendocrine carcinoma or gastric mixed adenoneuroendocrinecarcinoma.

LimitationsThis study has several limitations. First, patients with gastric adenocarcinoma were only includedfrom a single center, which may have caused selection bias. Thus, propensity score matching wasused to reduce the potential interference associated with stage difference. Second, this study wasperformed in China, so it is not clear whether our findings are generalizable for Western populations.Third, nearly 10.0% of patients with recurrence did not have a documented recurrence pattern, andthese patients had to be excluded from the recurrence pattern analysis, which may have led toselection bias. Fourth, patients with gastric adenocarcinoma patients were only included from onecenter, which may have produced center-related effects. Thus, stratified analyses by center volumewere performed, and we found that even in the high-volume centers, the long-term survival ofpatients with gastric neuroendocrine carcinoma or gastric mixed adenoneuroendocrine carcinomawas significantly worse than that of patients with gastric adenocarcinoma. Nevertheless, to ourknowledge, this study represents the largest multicenter study focusing on recurrence patterns ofgastric neuroendocrine carcinoma and gastric mixed adenoneuroendocrine carcinoma, which couldsupport the development of guidelines and suggests the need for prospective studies.

Conclusions

This cohort study found that gastric neuroendocrine carcinoma and gastric mixedadenoneuroendocrine carcinoma were associated with a worse prognosis than gastricadenocarcinoma and even poorly differentiated gastric adenocarcinoma. In addition, patients withgastric neuroendocrine carcinoma and gastric mixed adenoneuroendocrine carcinoma were morelikely to experience distant recurrence, especially patients with tumors penetrating into thesubserosa or deeper layers and with lymph node metastasis. Thus, additional close follow-upstrategies and effective adjuvant therapies, such as chemotherapy combined with targeted therapyor immunotherapy, should be implemented for these patients.

ARTICLE INFORMATIONAccepted for Publication: April 6, 2021.

Published: July 27, 2021. doi:10.1001/jamanetworkopen.2021.14180

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Lin J et al.JAMA Network Open.

Corresponding Authors: Changming Huang, MD ([email protected]), and Jianwei Xie, PhD([email protected]), Department of Gastric Surgery, Fujian Medical University Union Hospital, No. 29 XinquanRd, Fuzhou 350001 Fujian Province, China.

Author Affiliations: Department of Gastric Surgery, Fujian Medical University Union Hospital, Fuzhou, China(Jianxian Lin, Junpeng Lin, Zheng, P. Li, Huang, Xie); Key Laboratory of Ministry of Education of GastrointestinalCancer, Fujian Medical University, Fuzhou, China (Jianxian Lin, Junpeng Lin, Zheng, P. Li, Huang, Xie); Department

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 12/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

of Gastrointestinal Surgery, West District, First Affiliated Hospital of the University of Science and Technology ofChina, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei, China(Y. Zhao); Department of Gastrointestinal surgery, Affiliated Hospital of Qingdao University, Qingdao, China(Y. Zhou); Department of Pancreatic and Gastric Surgery, National Cancer Center, National Clinical Research Centerfor Cancer, Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing,China (Tian); Department of gastrointestinal surgery, the First Affiliated Hospital of Fujian Medical University,Fuzhou, Fujian Province, China (He); Department of General Surgery, Huashan Hospital, Fudan University,Shanghai, China (Hao); Department of Gastrointestinal Surgery, First Affiliated Hospital of Anhui MedicalUniversity, Hefei, China (Zou, Yongxiang Li); Department of Gastrointestinal Surgery, Yantai Yuhuangding Hospital,Yantai, China (Jiang); Department of Gastrointestinal Surgery, Renji Hospital, Shanghai Jiaotong University,Shanghai, China (G. Zhao); Department of Gastrointestinal Surgery, Gastrointestinal Surgery Research Institute,Affiliated Hospital of Putian University, Putian, China (W. Lin); Department of Gastrointestinal Surgery, FujianMedicine University Teaching Hospital, First Hospital of Putian, Putian, China (Y. Xu); Department of GeneralSurgery, Henan Cancer Hospital, Zhengzhou, China (Z. Li); Department of Gastrointestinal Surgery, ProvincialClinical Medical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou, China (Xue); Departmentof Gastrointestinal Surgery, Second People’s Hospital of Liaocheng, Liaocheng, China (S. Li); Department ofGeneral Surgery, Tianjin Medical University General Hospital, Tianjin, China (Fu); Department of General Surgery,First Affiliated Hospital of Nanjing Medical University, Nanjing, China (Z. Xu); Department of General Surgery,Guangdong Provincial People’s Hospital, Guangdong Academy of Medical Sciences, Guangzhou, China (Yong Li);Department of Gastrointestinal Surgery, Quanzhou First Hospital Affiliated to Fujian Medical University, Quanzhou,China (Chen); Department of Gastroenterology Surgery, First Affiliated Hospital of Soochow University, Suzhou,China (X. Zhou); Department of Gastrointestinal Surgery, Ruijin Hospital Affiliated to Shanghai Jiao Tong UniversitySchool of Medicine, Shanghai, China (Zhu); Department of General Surgery, Zhangzhou Affiliated Hospital of FujianMedical University, Zhangzhou, China (Cai); Department of Gastrointestinal Surgery, Meizhou People’s Hospital,Meizhou, China (E. Li); Department of Gastrointestinal Surgery, Second Affiliated Hospital, Nanchang University,Nanchang, China (H. Li).

Author Contributions: Dr Xie had full access to all of the data in the study and takes responsibility for the integrityof the data and the accuracy of the data analysis. Drs Jianxian Lin, Y. Zhao, Y. Zhou, Tian, He, and Junpeng Lincontributed equally to this work and should be considered co–first authors.

Concept and design: Y. Zhao, Y. Zhou, Tian, He, Junpeng Lin, Hao, Zou, Jiang, G. Zhao, W. Lin, Y. Xu, Z. Li, Xue, Fu,Yongxiang Li, Z. Xu, Yong Li, Chen, X. Zhou, Zhu, Cai, E. Li, H. Li, Zheng, Huang, Xie.

Acquisition, analysis, or interpretation of data: Jianxian Lin, Y. Zhao, Y. Zhou, Tian, He, Hao, Zou, Jiang, W. Lin, Y. Xu,Xue, S. Li, Fu, Z. Xu, Yong Li, Chen, Zhu, Cai, E. Li, H. Li, P. Li, Huang.

Drafting of the manuscript: Jianxian Lin, Y. Zhao, Y. Zhou, Tian, He, Junpeng Lin, Hao, Zou, Jiang, W. Lin, Y. Xu, Xue,Fu, Z. Xu, Yong Li, Chen, Zhu, Cai, E. Li, H. Li.

Critical revision of the manuscript for important intellectual content: Y. Zhao, Y. Zhou, Tian, He, Hao, Zou, Jiang, G.Zhao, W. Lin, Y. Xu, Z. Li, Xue, S. Li, Fu, Yongxiang Li, Z. Xu, Yong Li, Chen, X. Zhou, Zhu, Cai, E. Li, H. Li, Zheng, P. Li,Huang, Xie.

Statistical analysis: Jianxian Lin, Y. Zhao, Y. Zhou, Tian, He, Junpeng Lin, Hao, Zou, Jiang, W. Lin, Y. Xu, Xue, Fu, Z.Xu, Yong Li, Chen, Zhu, Cai, E. Li, H. Li.

Obtained funding: Zheng, P. Li, Huang, Xie.

Administrative, technical, or material support: Z. Li, S. Li, Yongxiang Li, X. Zhou, Zheng, P. Li, Huang, Xie.

Supervision: Y. Zhao, Y. Zhou, Tian, He, Junpeng Lin, Hao, Zou, Jiang, W. Lin, Y. Xu, Xue, Fu, Z. Xu, Yong Li, Chen,Zhu, Cai, E. Li, H. Li, Zheng, P. Li, Huang, Xie.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by the Scientific and Technological Innovation Joint Capital Projectsof Fujian Province (grant No. 2017Y9004, 2019Y9098), Construction Project of Fujian Province Minimally InvasiveMedical Center (grant No. [2017]171), and Fujian Provincial Health Technology Project (grant No. 2019-ZQN-37).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection,management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; anddecision to submit the manuscript for publication.

REFERENCES1. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer statistics, 2012. CA Cancer J Clin.2015;65(2):87-108. doi:10.3322/caac.21262

2. Fléjou JF. WHO Classification of digestive tumors: the fourth edition. Article in French. Ann Pathol. 2011;31(5)(suppl):S27-S31. doi:10.1016/j.annpat.2011.08.001

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 13/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

3. Kubota T, Ohyama S, Hiki N, Nunobe S, Yamamoto N, Yamaguchi T. Endocrine carcinoma of the stomach:clinicopathological analysis of 27 surgically treated cases in a single institute. Gastric Cancer. 2012;15(3):323-330.doi:10.1007/s10120-011-0122-5

4. Jiang SX, Mikami T, Umezawa A, Saegusa M, Kameya T, Okayasu I. Gastric large cell neuroendocrine carcinomas:a distinct clinicopathologic entity. Am J Surg Pathol. 2006;30(8):945-953. doi:10.1097/00000478-200608000-00003

5. Park JY, Ryu MH, Park YS, et al. Prognostic significance of neuroendocrine components in gastric carcinomas.Eur J Cancer. 2014;50(16):2802-2809. doi:10.1016/j.ejca.2014.08.004

6. Kim BS, Park YS, Yook JH, Kim BS. Comparison of relapse-free survival in gastric neuroendocrine carcinoma(WHO grade 3) and gastric carcinoma. Therap Adv Gastroenterol. 2017;10(5):407-415. doi:10.1177/1756283X17697870

7. Ishida M, Sekine S, Fukagawa T, et al. Neuroendocrine carcinoma of the stomach: morphologic andimmunohistochemical characteristics and prognosis. Am J Surg Pathol. 2013;37(7):949-959. doi:10.1097/PAS.0b013e31828ff59d

8. Garcia-Carbonero R, Sorbye H, Baudin E, et al; Vienna Consensus Conference participants. ENETS consensusguidelines for High-Grade gastroenteropancreatic neuroendocrine tumors and neuroendocrine carcinomas.Neuroendocrinology. 2016;103(2):186-194. doi:10.1159/000443172

9. Xu XF, Xing H, Han J, et al. Risk factors, patterns, and outcomes of late recurrence after liver resection forhepatocellular carcinoma: a multicenter study from china. JAMA Surg. 2019;154(3):209-217. doi:10.1001/jamasurg.2018.4334

10. D’Angelica M, Gonen M, Brennan MF, Turnbull AD, Bains M, Karpeh MS. Patterns of initial recurrence incompletely resected gastric adenocarcinoma. Ann Surg. 2004;240(5):808-816. doi:10.1097/01.sla.0000143245.28656.15

11. Spolverato G, Ejaz A, Kim Y, et al. Rates and patterns of recurrence after curative intent resection for gastriccancer: a United States multi-institutional analysis. J Am Coll Surg. 2014;219(4):664-675. doi:10.1016/j.jamcollsurg.2014.03.062

12. Lee JH, Chang KK, Yoon C, Tang LH, Strong VE, Yoon SS. Lauren histologic type is the most important factorassociated with pattern of recurrence following resection of gastric adenocarcinoma. Ann Surg. 2018;267(1):105-113. doi:10.1097/SLA.0000000000002040

13. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medicalresearch involving human subjects. JAMA. 2013;310(20):2191-2194. doi:10.1001/jama.2013.281053.

14. Byrne BE, Mamidanna R, Vincent CA, Faiz O. Population-based cohort study comparing 30- and 90-dayinstitutional mortality rates after colorectal surgery. Br J Surg. 2013;100(13):1810-1817. doi:10.1002/bjs.9318

15. Japanese Gastric Cancer Association. Japanese gastric cancer treatment guidelines 2010 (ver. 3). GastricCancer. 2011;14(2):113-123. doi:10.1007/s10120-011-0042-4

16. Bang YJ, Kim YW, Yang HK, et al; CLASSIC trial investigators. Adjuvant capecitabine and oxaliplatin for gastriccancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial. Lancet. 2012;379(9813):315-321. doi:10.1016/S0140-6736(11)61873-4

17. Amin MB, Edge SB, Greene FL, et al, eds. AJCC Cancer Staging Manual. 8th ed. Springer; 2017. doi:10.1007/978-3-319-40618-3

18. Kolbeinsson H, Hoppe A, Bayat A, et al. Recurrence patterns and postrecurrence survival after curative intentresection for pancreatic ductal adenocarcinoma. Surgery. 2021;169(3):649-654. doi:10.1016/j.surg.2020.06.042

19. Lee D, Son SY, Kim YB, Han SU, Hur H. Neural invasion is a significant contributor to peritoneal recurrence insignet ring cell gastric carcinoma. Ann Surg Oncol. 2018;25(5):1167-1175. doi:10.1245/s10434-018-6371-3

20. Mokadem I, Dijksterhuis WPM, van Putten M, et al. Recurrence after preoperative chemotherapy and surgeryfor gastric adenocarcinoma: a multicenter study. Gastric Cancer. 2019;22(6):1263-1273. doi:10.1007/s10120-019-00956-6

21. Lee HJ, Hyung WJ, Yang HK, et al; Korean Laparo-endoscopic Gastrointestinal Surgery Study (KLASS) Group.Short-term outcomes of a multicenter randomized controlled trial comparing laparoscopic distal gastrectomy withd2 lymphadenectomy to open distal gastrectomy for locally advanced gastric cancer (KLASS-02-RCT). Ann Surg.2019;270(6):983-991. doi:10.1097/SLA.0000000000003217

22. Sano T, Coit DG, Kim HH, et al. Proposal of a new stage grouping of gastric cancer for TNM classification:International Gastric Cancer Association staging project. Gastric Cancer. 2017;20(2):217-225. doi:10.1007/s10120-016-0601-9

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 14/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022

23. Lin J, Zhao Y, Zhou Y, et al. Which staging system is more suitable for gastric neuroendocrine cancer and mixedadenoneuroendocrine carcinomas: a multicenter cohort study. Neuroendocrinology. 2020. doi:10.1159/000505924

24. Woo Y, Son T, Song K, et al. A novel prediction model of prognosis after gastrectomy for gastric carcinoma:development and validation using Asian databases. Ann Surg. 2016;264(1):114-120. doi:10.1097/SLA.0000000000001523

25. Li H, Wang C, Wei Z, et al. Differences in the prognosis of gastric cancer patients of different sexes and racesand the molecular mechanisms involved. Int J Oncol. 2019;55(5):1049-1068. doi:10.3892/ijo.2019.4885

26. Mokdad AA, Yopp AC, Polanco PM, et al. Adjuvant chemotherapy vs postoperative observation followingpreoperative chemoradiotherapy and resection in gastroesophageal cancer: a propensity score–matched analysis.JAMA Oncol. 2018;4(1):31-38. doi:10.1001/jamaoncol.2017.2805

27. Lu J, Xu BB, Zheng ZF, et al. CRP/prealbumin, a novel inflammatory index for predicting recurrence after radicalresection in gastric cancer patients: post hoc analysis of a randomized phase III trial. Gastric Cancer. 2019;22(3):536-545. doi:10.1007/s10120-018-0892-0

28. Wang N, Bu Q, Yang J, et al. Insurance status is related to overall survival in patients with small intestineadenocarcinoma: a population-based study. Curr Probl Cancer. 2020;44(1):100505. doi:10.1016/j.currproblcancer.2019.100505

29. Chan JC, Chan DL, Diakos CI, et al. The lymphocyte-to-monocyte ratio is a superior predictor of overall survivalin comparison to established biomarkers of resectable colorectal cancer. Ann Surg. 2017;265(3):539-546. doi:10.1097/SLA.0000000000001743

30. Jung DH, Bae YS, Yoon SO, et al. Poorly differentiated carcinoma component in submucosal layer should beconsidered as an additional criterion for curative endoscopic resection of early gastric cancer. Ann Surg Oncol.2015;22(suppl 3):S772-S777. doi:10.1245/s10434-015-4794-7

31. Chon HJ, Hyung WJ, Kim C, et al. Differential prognostic implications of gastric signet ring cell carcinoma: stageadjusted analysis from a single high-volume center in Asia. Ann Surg. 2017;265(5):946-953. doi:10.1097/SLA.0000000000001793

SUPPLEMENT.eTable 1. Distribution of Multicenter PatientseTable 2. Comparison of Stages Between Different Pathological Types in the Original and the Matched Data SetseTable 3. Univariable and Multivariable Logistic Regression Analyses of Factors Potentially Associated with DistantRecurrence in the Matched Data Sets of AC and NECeTable 4. Univariate and Multivariable Logistic Regression Analyses of Factors Potentially Associated with DistantRecurrence in the Matched Data Sets of AC and MANECeTable 5. Univariable and Multivariable Logistic Regression Analyses of Pathological Factors Potentially Associatedwith Distant Recurrence in Patients with NEC and MANECeFigure 1. Diagram Representing the Selection of the Study PopulationeFigure 2. Incidence of Recurrence According to Histologic Subtype in Original Data and Matched DataeFigure 3. Kaplan-Meier Survival Curves for Patients With Gastric Neuroendocrine Carcinoma (NEC), MixedAdenoneuroendocrine Carcinoma (MANEC), or Adenocarcinoma (AC) in Original DataeFigure 4. Recurrence Patterns According to Histologic Subtype in the Original DataeFigure 5. Kaplan-Meier Survival Curves for Postrecurrence Survival for Patients With Different HistologicSubtypes in the Matched DataeFigure 6. Kaplan-Meier Survival Curves for Patients With Gastric Neuroendocrine Carcinoma (NEC) andAdenocarcinoma (AC) in High-Volume CenterseFigure 7. Recurrence Patterns According to Histologic Subtype in the Matched DataeFigure 8. Kaplan-Meier Survival Curves for Patients With Gastric Mixed Adenoneuroendocrine Carcinoma(MANEC) and Adenocarcinoma (AC) in High-Volume Centers

JAMA Network Open | Gastroenterology and Hepatology Comparison of Survival and Patterns of Recurrence in Gastric Carcinoma Subtypes

JAMA Network Open. 2021;4(7):e2114180. doi:10.1001/jamanetworkopen.2021.14180 (Reprinted) July 27, 2021 15/15

Downloaded From: https://jamanetwork.com/ on 01/28/2022