Cancer Medicine

Open Access
ORIGINAL RESEARCH

Survival benefit of gastrectomy for gastric cancer with

peritoneal carcinomatosis: a propensity score-­matched
analysis
Xiuwen Geng1,2,a, Hao Liu1,a, Tian Lin1,a, Yanfeng Hu1, Hao Chen1, Liying Zhao1, Tingyu Mou1,
Xiaolong Qi1, Jiang Yu1 & Guoxin Li1
1Department of General Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
2Department of Gastrointestinal Surgery, The First People’s Hospital of Yueyang, Yueyang, China

Keywords Abstract
Chemotherapy, gastrectomy, gastric cancer,
peritoneal carcinomatosis Peritoneal carcinomatosis (PC) is the most frequent pattern of metastasis in
stage IV gastric cancer (GC). The study aims to investigate the efficacy of gas-
Correspondence trectomy in GC with PC. Clinicopathological data of 518 stage IV GC patients
Guoxin Li or Jiang Yu or Xiaolong Qi, were retrospectively collected in Nanfang Hospital. Among all cases, 312 GC
Department of General Surgery, Nanfang
patients with PC (without other site of metastasis) were eligible. Univariate and
Hospital, Southern Medical University, 1838
North Guangzhou Avenue, Guangzhou
multivariate analyses were performed to identify the independent prognostic fac-
510515, China. Tel: +86 20 6164 1681; tors. Propensity score matching analysis was performed to balance the charac-
Fax: +86 20 6164 1683; teristics and treatment-­related factors. There was a significantly improved overall
E-mail: [email protected] or [email protected] survival in gastrectomy group (148 patients) compared with nonresection group
or [email protected] (164 patients) (P  <  0.001). The 1-­year and 2-­year survival rates were 49.8% and
21.5% in gastrectomy group, whereas 28.8% and 9.7% in nonresection group,
Funding Information
respectively. Further analysis showed that gastrectomy had also improved survival
This work was supported by the Guangdong
Provincial Science and Technology Key Project
in P1 (P = 0.017) and P2 stage patients (P < 0.001), but not P3 stage (P = 0.495).
(No.2014A020215014), the Research Fund of The modality of gastrectomy plus chemotherapy plus hyperthermic intraperitoneal
Public Welfare in the Health Industry, the chemotherapy (HIPEC) showed an optimum survival. In addition, P3 disease,
National Health and Family Planning nongastrectomy, nonchemotherapy, non-­HIPEC, and age ≥ 60 years were inde-
Commission of China (No. 201402015), and pendently associated with poor survival. The gastrectomy plus chemotherapy
the Key Clinical Specialty Discipline plus HIPEC modality showed a significant survival benefit for gastric adenocar-
Construction Program.
cinoma patients, particularly in those with P1 and P2 diseases.
Received: 23 June 2016; Revised: 21 July
2016; Accepted: 28 July 2016

Cancer Medicine 2016; 5(10):2781–2791

doi: 10.1002/cam4.877

aThese authors contributed equally to this work.

Introduction highest among cancers in 2015 [3]. Besides, about 679,100

Chinese would be newly diagnosed gastric cancer and
Gastric cancer contributes significantly to the burden of about 498,000 Chinese would die from gastric cancer in
cancer-­
related death worldwide and in China, mainly 2015. Highly advanced gastric cancer characterized by local
because of most patients being presented with advanced invasion and/or metastasis has dismal prognosis, which,
disease at the time of diagnosis [1, 2]. According to the is particularly unfavorable for patients with peritoneal
most recent statistical data of National Central Cancer carcinomatosis (PC) and almost 60% of gastric cancer
Registry of China, both the incidence rate and mortality deaths are due to PC [4]. As the most frequent pattern
rate of gastric cancer are estimated to rank the second of metastasis [5, 6], PC was observed in 10–20% of patients

© 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd. 2781
This is an open access article under the terms of the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work is properly cited.
Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis X. Geng et al.

who were scheduled for potentially curative resection and PC from the results of a retrospective cohort in a single
40% of patients who were clinically staged as II–III before medical center in China.
the intraoperative abdominal examination [7]. In the
­literature, the median survival time (MST) of gastric cancer
Patients and Methods
patients with PC is 3–9 months [8, 9].
In recent years, the overall survival rate of gastric cancer
Cohort
patients increased with the availability of new medicine
and combined chemotherapy compared with previous sup- A total of 4135 patients were diagnosed with gastric cancer
portive treatment [8]. Despite the improvement of systemic in Nanfang Hospital, Southern Medical University,
chemotherapy, the long-­term survival rate of patients with Guangzhou, China, from January 2005 to September 2015.
PC is still very low [9, 10]. Additionally, the acquired The study was approved by Institutional Review Board
resistance and adverse effects limit chemotherapy. Surgical of Nanfang Hospital, Southern Medical University. Among
resection has been considered as the most efficient treat- these patients, 1115 (27.0%) were classified as stage IV
ment for early gastric cancer patients for a long time. adenocarcinoma according to the third English edition
For patients with advanced disease, it is generally agreed of the Japanese classification of gastric carcinoma [11].
that surgery provides palliation of the major symptoms After reviewing the medical records of these patients ret-
such as bleeding and/or obstruction, although the optimal rospectively by the two independent surgical oncologists,
surgical management for patients with minimal symptoms there were five categories of patients unqualified for this
is debated. The National Comprehensive Cancer Network study: 108 patients (9.7%) with incomplete medical records
recommends that patients with metastatic disease are not data (including patients without the detailed description
candidates for surgery unless they present with bleeding of the peritoneal metastasis in the operative reports); 214
and/or obstruction. However, in 2011, the Japanese Gastric patients (19.2%) without follow-­ up data; 198 patients
Cancer Association guidelines suggested that patients with (17.8%) who refused to accept further advanced therapy
metastases might be candidates for gastrectomy even with- when the disease was diagnosed; 15 patients (1.3%) who
out major symptoms [11, 12]. passed away during the first course of hospitalization;
As PC is currently regarded as a variant of systemic and 62 patients (5.6%) who were diagnosed and treated
spread disease of gastric cancer, systemic chemotherapy in other hospitals previously.
is considered as the main treatment modality [11, 12]. It After the above exclusion, 518 patients were enrolled.
is controversial if gastrectomy is beneficial for gastric cancer As our study was focused on the gastric cancer patients
patients with PC. In the past few years, a few groups of with PC only, 6 patients who combined with other tumors
surgeons investigated the feasibility of noncurative gas- (0.5%) and 200 patients with hepatic/distant lymph node
trectomy in patients with incurable factors [13, 14]. Studies metastasis or other distant metastases (17.9%) were further
suggested that gastrectomy could raise the survival rate excluded. At last, 312 patients were eligible for this study.
of the patients with PC without increasing the mortality These enrolled patients were divided into a gastrectomy
rate and might be beneficial to reduce symptoms and group and a nonresection group based on whether the
enhance life quality [15–17]. Other reports, however, had gastrectomy was performed or not. Patients who under-
indicated that palliative gastrectomy was associated with went gastrojejunostomy or only exploratory laparotomy
significant morbidity, longer hospital stays, poor quality were classified into the nonresection group. Each group
of life, and had no survival benefit in these patients [18, was further divided into three subgroups according to
19]. In 2014, the GYMSSA (NCT00941655) trial reported the treatment strategies: chemotherapy plus HIPEC, chemo-
that maximal cytoreductive surgery combined with hyper- therapy, and no chemotherapy (Fig. 1). Then, the 312
thermic intraperitoneal chemotherapy (HIPEC) and sys- patients were divided into three subgroups in accordance
temic chemotherapy could achieve more prolonged survival with the extent of PC. In these three groups, each group
in selected gastric cancer patients with PC than the chemo- was further divided into a gastrectomy subgroup and a
therapy group [20]. However, the recent results of nonresection subgroup. The patients’ clinicopathologic
REGATTA (UMIN000001012) trial presented that gastrec- features, treatment-­related factors, and survival curves were
tomy followed by chemotherapy did not show any survival compared between these groups and subgroups.
benefit compared with chemotherapy alone in the advanced
gastric cancer patients with a single noncurable factor [21].
The diagnosis of PC and definition of P
Hence, there is still no consensus on the value of gas-
stages
trectomy for late-­stage gastric cancer patients with PC.
This study aimed to evaluate the impact of gastrectomy All patients were histologically diagnosed as gastric adeno-
on the survival of stage IV gastric cancer patients with carcinoma after endoscopy and biopsy before initial

2782 © 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.
X. Geng et al. Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis

Figure 1. Selection and grouping of stage IV gastric adenocarcinoma patients with peritoneal carcinomatosis.

treatment and diagnosed histopathologically after lapa- More than two cycles of chemotherapy were defined as
rotomy at the department of pathology, Nanfang Hospital, chemotherapeutic intervention, and a 5-­fluorouracil-­based
Southern Medical University. (93.3%) or paclitaxel-­based (6.7%) regimen was adminis-
On the basis of the operative findings, PC was classified tered either preoperatively and/or postoperatively. HIPEC
according to the first English edition of Japanese classi- was performed by using a closed circuit of 120 mg doc-
fication of gastric carcinoma as follows: P0, no peritoneal etaxel in 3500 mL normal saline at 43 ± 0.5°C for 60 min.
seeding; P1, disseminating metastasis to the region directly The perfusion tubes were placed at appropriate sites during
adjacent the peritoneum of stomach (above the transver- the primary operation and the HIPEC was conducted on
secolon), including the greater omentum; P2, several scat- day 2, day 4, and day 6 postoperatively.
tered metastases to the distant peritoneum and ovarian
metastasis alone; and P3, numerous metastases to the
Follow-­up
distant peritoneum [22].
The primary endpoint was MST. The secondary endpoints
were the 1-­year overall survival (OS) rate and 2-­year OS
Gastrectomy, chemotherapy, and HIPEC
rate. As of September 7, 2015, the average follow-­up dura-
Noncurative gastrectomy for gastric adenocarcinoma is tion was 11.9 months. Twelve patients in the gastrectomy
defined as a gastrectomy either with a residual tumor of group and six patients in the nonresection group were lost
metastatic lesion or with an unresectable lesion [23]. In to follow-­up in the process of this study. Most complete
this study, total or distal gastrectomy was performed accord- data on prognoses were collected during the outpatient
ing to the location of the primary lesion. Consequently, visit. Telephone calls and letters were used to identify patients
primary tumors and greater omentum were all removed who could not attend regular hospital visiting. In this study,
regardless of lymphadenectomy or metastasectomy. OS time was defined as the period from initial treatment
Gastrectomy was performed in the following two condi- (surgery or chemotherapy) to the date of death. Patients
tions: first, feasibility was evaluated by operating surgeons who were still alive till the cutoff date, lost to follow-­up,
based on patient’s symptom, performance status, nutritional and died of any other cause were marked as censored data.
status, and technical feasibility; and second, the operation
informed consent was signed preoperatively or intraopera-
Propensity score matching analysis
tively. The circumstances below were considered to be
contraindications to resection: infiltration of locoregionally We selected five covariates (age, Eastern Cooperative
advanced disease to the root of the mesentery, invasion Oncology Group Performance Status [ECOG-­PS], P stage,
or encasement of major vascular structures or important chemotherapy, and HIPEC) for propensity score matching
organs, and extensive adhesion or fixation of the tumor. (PSM) analysis. The propensity score was calculated using

© 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd. 2783
Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis X. Geng et al.

a logistic regression model and a nearest neighbor match- chi-­square test was used to compare the categorical vari-
ing algorithm. Patients from the gastrectomy group were ables and two independent t-­tests were used to compare
one-­to-­
one matched with patients from the unresection the quantitative variables. Cumulative survival analysis was
group based on the top 148 scores. estimated using the Kaplan–Meier method and compared
by the log-­rank test. Univariate and multivariate survival
analyses were performed using Cox proportional hazards
Statistical analysis
regression model to produce a hazard ratio. Hazard ratios
Statistical analyses were performed with SPSS software and their 95% confidence intervals provided an evaluation
version 21.0 for Windows (SPSS, Inc., Chicago, IL). The of the relative rate of death between the two groups

Table 1. Clinical and pathologic characteristic of enrolled patients.

Gastrectomy Nonresection
Total (n = 312) n = 148 (51.3%) n = 164 (48.7%)

Characteristic No.(%) Mean ± SD No.(%) Mean ± SD No.(%) Mean ± SD P value

Gender 0.539
Male 199 (63.8) 97 (65.5) 102 (62.2)
Female 113 (36.2) 51 (34.5) 62 (37.8)
Age, years 53.9 ± 13.5 55.3 ± 12.9 52.6 ± 13.8 0.081
ECOG-­PS 0.037
0, 1 274 (87.8) 136 (91.9) 138 (84.1)
2, 3 38 (12.2) 12 (8.1) 26 (15.9)
BMI, kg/m2 20.8 ± 3.0 20.9 ± 2.9 20.7 ± 3.0 0.503
Borrmann type 0.460
I 21 (6.7) 12 (8.1) 9 (5.5)
II 3 (1.0) 1 (0.7) 2 (1.2)
III 221 (70.8) 108 (73.0) 113 (68.9)
IV 67 (21.5) 27 (18.2) 40 (24.4)
Histology 0.489
Differentiated 22 (7.1) 12 (8.1) 10 (6.1)
Undifferentiated 290 (92.9) 136 (91.9) 154 (93.9)
Target chemotherapy 0.193
No 303 (97.1) 145 (98.0) 158 (96.3)
Trastuzumab 2 (0.6) 0 (0.0) 2 (1.2)
Cetuximab 2 (0.6) 2 (1.4) 0 (0.0)
Bevacizumab 3 (1.0) 1 (0.7) 2 (1.2)
Apatinib 2 (0.6) 0 (0.0) 2 (1.2)
P Stage 0.185
P1 115 (36.9) 62 (41.9) 53 (32.3)
P2 69 (22.1) 32 (21.6) 37 (22.6)
P3 128 (41.0) 54 (36.5) 74 (45.1)
Chemotherapy 0.115
Yes 154 (49.4) 80 (54.1) 74 (45.1)
No 158 (50.6) 68 (45.9) 90 (54.9)
HIPEC 0.018
Yes 40 (12.8) 12 (8.1) 28 (17.1)
No 272 (87.2) 136 (91.9) 136 (82.9)
Postoperative complication
Intraluminal hemorrhage 4 (1.3) 3 (2.0) 1 (0.6) 0.266
Small bowel obstruction 7 (2.2) 5 (3.4) 2 (1.2) 0.199
Wound problem 3 (1.0) 1 (0.7) 2 (1.2) 0.623
Pulmonary infections 9 (2.9) 3 (2.0) 6 (3.7) 0.390
Intraabdominal abscess 5 (1.6) 5 (3.4) 0 (0.0) 0.018
Thromboembolism 6 (1.9) 4 (2.7) 2 (1.2) 0.194
Other 8 (2.6) 4 (2.7) 4 (2.4) 0.474
Total 35 (11.2) 22 (14.9) 13 (7.9) 0.052

HIPEC, hyperthermic intraperitoneal chemotherapy.

2784 © 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.
X. Geng et al. Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis

compared. Additionally, PSM analysis was conducted to targeted therapy, and postoperative morbidity. However,
balance the baseline characteristics and treatment-­related the differences between these two groups on ECOG-­ PS
factors. Statistical significance was defined as P  <  0.05 and HIPEC were statistically significant (0.037 and 0.018,
(two sided). respectively).

Results Long-­term survival

Kaplan–Meier analysis showed the median survival for all
Baseline characteristics
312 patients is 10.0 (95% confidence interval [95% CI]:
A total of 312 patients were involved in this study and 8.82–11.18) months. The 1-­year survival and 2-­year sur-
the comparable baseline clinicopathologic features and vival were 39.6% and 16.1%, respectively (Fig. 2A). After
treatment-­ related characteristic were all summarized in the patients were divided based on the surgery situations
Table 1. Among these patients, the male and female ratio into gastrectomy group (n = 148) and nonresection group
was approximately 1.8 (119/113) and their average age (n = 164), further analysis demonstrated that the median
was 53.9 years old. More than 85% of patients were scored survival is 12.0 (95% CI: 10.39–13.62) months in the
0 and 1 of ECOG-­ PS. The average Body Mass Index gastrectomy group and 8.0 (95% CI: 6.90–9.10) months
(BMI) was 20.8 kg/m2. The peritoneal dissemination was in the nonresection group (P < 0.001, Fig. 2B). The 1-­year
classified as P1 in 115 patients (36.9%), P2 in 69 patients survival rate is 49.8% in the gastrectomy group and 28.8%
(22.1%), and P3 in 128 patients (41.0%). The maximal in nonresection group, whereas the 2-­ year survival rate
Borrmann type of the tumor is type III (>70%), but type is 21.5% in gastrectomy group and 9.7% in nonresection
II is the least (<2%). All 312 cases are histopathologically group (P  <  0.001, Table 2).
adenocarcinoma and more than 90% are undifferentiated. When the patients in gastrectomy group and nonresec-
Additionally, most patients (97.1%) did not receive tar- tion group were further stratified into three subgroups
geted therapy, whereas 158 patients (50.6%) did not by treatment modalities (chemotherapy plus HIPEC,
undergo palliative chemotherapy. Only 12.8% of patients chemotherapy, and no chemotherapy) (Table 2 and
experienced HIPEC. The overall postoperative morbidity Fig. 2C), the results manifested that patients in gastrec-
rate is 11.2% (35/312) and postoperative complications tomy group with either chemotherapy plus HIPEC or
including intraluminal hemorrhage, small bowel obstruc- chemotherapy had statistically significant better survival
tion, wound problem, pulmonary infections, intraabdomi- rates than those in nonresection group, including MST
nal abscess, thromboembolism, etc., were observed. As (17.0 vs. 13.0 months, P = 0.034; and 15.0 vs. 7.0 months,
shown in Table 1, the above characteristics were also P  <  0.001), 1-­year survival rate (66.7% vs. 50.0% and
compared between gastrectomy group (148 patients) and 64.3% vs. 27.8%), and 2-­ year survival rate (28.5% vs.
nonresection group (164 patients). There were no statisti- 3.6% and 30.9% vs. 10.6%). However, no differences were
cal differences between these two groups on patient’s age, observed in the subgroups of patients without chemo-
gender, BMI, Borrmannn type, tumor differentiation, the therapy between the gastrectomy group and the nonresec-
extent of peritoneal seeding (P stage), chemotherapy, tion group (MST: 7.0 vs. 7.0 months, P  =  0.691).

Figure 2. Overall survival and survivals of different treatment subgroups of stage IV gastric adenocarcinoma patients with peritoneal carcinomatosis.

© 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd. 2785
Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis X. Geng et al.

When all the six subgroups were compared with each nonresection group (nonresection plus chemotherapy plus
other (Table 2 and Fig. 2C), the results indicated that HIPEC, n = 28, MST: 13.0 months; nonresection plus
the treatment modality of gastrectomy plus chemotherapy chemotherapy, n = 66, MST: 7.0 months; and no chemo-
plus HIPEC (n = 12, MST: 17.0 months) had the opti- therapy, n = 77, MST: 7.0 months). In addition, within
mum survival when compared with the gastrectomy plus the nonresection group, when the three subgroups were
chemotherapy (n = 77, MST: 15.0 months) and gastrec- compared with each other, the patients with chemotherapy
tomy alone (n = 59, MST: 7.0 months) in the gastrectomy plus HIPEC showed a much significant survival benefit
group and all the three counterparts (P  <  0.001) in the than the patients with only chemotherapy (MST: 13.0 vs.

Table 2. Comparison between the subgroups of treatment modality and P stages.

Gastrectomy (n = 148) Nonresection (n = 164)

Survival rate (%) Survival rate (%)

Characteristic n MST (m) 95% CI 1 year 2 year n MST (m) 95% CI 1 year 2 year P value
Treatment modality
Chemotherapy + HIPEC 12 17.0 11.27–22.74 66.7 28.5 28 13.0 7.58–18.42 50.0 3.6 0.034
Chemotherapy 77 15.0 12.27–17.73 64.3 30.9 66 7.0 5.96–8.04 27.8 10.6 <0.001
No chemotherapy 59 7.0 4.66–9.34 24.9 6.6. 70 7.0 6.32–7.68 22.2 6.6 0.691
P stage
P1 62 16.0 12.29–19.71 69.0 31.5 53 12.0 8.49–15.51 47.0 25.7 0.017
P2 32 16.0 11.97–20.03 60.9 31.2 37 9.0 7.51–10.49 21.6 0.0 <0.001
P3 54 7.0 5.79–8.24 16.7 1.9 74 6.0 4.67–7.33 15.9 0.0 0.495
Overall 148 12.0 10.39–13.62 49.8 21.5 164 8.0 6.90–9.10 28.8 9.7 <0.001

MST, median survival time; HIPEC, hyperthermic intraperitoneal chemotherapy.

Figure 3. Overall survival and P stage subgroup survivals of stage IV gastric adenocarcinoma patients with peritoneal carcinomatosis.

2786 © 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.
X. Geng et al. Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis

7.0 months, P = 0.037) and patients with no chemotherapy survival benefit was still observed in the gastrectomy group
subgroups (MST: 13.0 vs. 7.0 months, P = 0.044) (Table 2). (MST: 12.0 vs. 7.0 months, P  <  0.001).
When the OS was analyzed according to the extent of
PC, it was unsurprising to find that there were significant
Multivariate analysis
heterogeneity of survival time among patients in P1, P2.
and P3 (MST: 15 months vs. 11 months vs. 6 months; To explore an optimization model for which patients could
P  <  0.001 by log-­rank test), as shown in Figure 3A and benefit from the treatment of gastric adenocarcinoma with
Table 2. For groups P1, P2, and P3, the 1-­year survival PC, analyses of univariate, multivariate, and prognostic
rates are 59.2%, 42.1%, and 17.6%, respectively (P < 0.001, factors were conducted in this study. Univariate analysis
respectively) and the 2-­year survival rates are 28.4%, 16.5%, of potential prognostic factors was conducted for the 312
and 0.8%, respectively (P < 0.001, respectively) (Table 2). patients. In the analysis, patient’s gender, age (<60
When patients in gastrectomy group and nonresection vs. > = 60 years old), ECOG-­PS (0, 1 vs. 2, 3), P stage
group were stratified by P stages and further analyzed, (P1, P2, vs. P3), gastrectomy (yes or no), chemotherapy
it was found that in P1 and P2 subgroups (Fig. 3B, C (yes or no), HIPEC (yes or no), and targeted therapy
and Table 2), patients with gastrectomy had statistically (yes or no) were enrolled as covariates. The univariate
significant longer survival time than those in the nonre- survival analysis revealed that, as shown in Table 3, age,
section subgroups including MST (16.0 vs. 12.0 months, P stage, gastrectomy, chemotherapy, and HIPEC were
P  =  0.017; and 16.0 vs. 9.0 months, P  <  0.001), 1-­year associated with survival. The following factors were con-
survival rate (69.0% vs. 47.0% and 60.9% vs. 21.6%), sidered to be irrelevant: gender (P  =  0.850), ECOG-­PS
and 2-­year survival rate (31.5% vs. 25.7% and 31.2% vs. (P  =  0.906), and targeted therapy (P  =  0.666). In mul-
0.0%). These differences were not observed between the tivariate analysis, the patient’s age ≥60 years old (HR: 1.561;
gastrectomy subgroup and the nonresection subgroup of 95% CI: 1.201–2.028; P  =  0.001), P3 disease (HR:
P3 patients, as shown in Figure 3D and detailed in Table 2. 2.698; 95% CI: 2.046–3.559; P < 0.001), nonresection (HR:
0.597; 95% CI: 0.456–0.781; P < 0.001), nonchemotherapy
(HR: 0.624; 95% CI: 0.479–0.814; P  <  0.001), and non-­
Propensity score matching analysis
HIPEC (HR: 0.539; 95% CI: 0.344–0.843; P  =  0.007) were
To obtain more reliable evidence, PSM was conducted identified as independent poor survival factors.
to compensate for selection bias and avoid potential con-
founding effects. The P values of ECOG-­PS and HIPEC
Discussion
between the gastrectomy group and nonresection group
in the initial baseline were 0.037 and 0.018. After PSM38, This study showed that median survival was significantly
the baseline and treatment-­ related characteristics were longer in the gastrectomy group compared with the non-
balanced and patients were one-­to-­one matched between resection group in gastric cancer patients with PC. The
the two groups. The P values were recalculated and there 1-­year and 2-­year survival rates were also significantly
were no significant differences between the gastrectomy higher in gastrectomy group compared with the nonresec-
group and nonresection group (for ECOG-­PS: P  =  0.328; tion group (28.8% and 9.7%, respectively). These results
and for HIPEC: P  =  0.427). The survival comparison was were supported by a report published by Xia et al., who
also performed again between these two groups, and the found that the OS of patients with gastrectomy was longer

Table 3. Independent prognostic factors on the univariate and multivariate analysis.

Univariate analysis Multivariate analysis

Characteristic n MST (m) HR 95% CI P value HR 95% CI P value

Gender, male/female 199/113 10.0/10.0 0.975 0.747–1.271 0.850

Age, <60/> = 60 years 196/116 11.0/8.0 1.450 1.121–1.874 0.005 1.561 1.201–2.028 0.001
ECOG-­PS, 0–1/2–3 274/38 10.0/10.0 1.023 0.701–1.492 0.906
P Stage, P1, P2/P3 184/128 13.0/6.0 2.816 2.148–3.691 <0.001 2.698 2.046–3.559 <0.001
Gastrectomy, Yes/No 148/164 12.0/8.0 0.596 0.460–0.773 <0.001 0.597 0.456–0.781 <0.001
Chemotherapy, Yes/No 154/158 13.0/8.0 0.544 0.420–0.705 <0.001 0.624 0.479–0.814 <0.001
HIPEC, Yes/No 40/272 14.0/9.0 0.638 0.411–0.992 0.046 0.539 0.344–0.843 0.007
Target chemotherapy, Yes/No 9/303 15.0/10.0 0.847 0.399–1.799 0.666

MST, median survival time, HIPEC, hyperthermic intraperitoneal chemotherapy.

© 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd. 2787
Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis X. Geng et al.

than patients with nonresection group [15]. Our findings the peritoneum before other distant metastasis. In our
were also consistent with the reported data of Sun et al., study, we selected 312 cases with complete clinical data
Hioki et al., Kim et al., and Li et al. [16, 24–26]. In from 518 initially diagnosed stage IV gastric adenocarci-
further analysis of the subgroups stratified by the extent noma patients. In these cases, the tumors were mostly
of PC, we found that patients with P1 and P2 diseases located in the middle and lower parts of the stomach
had better median survivals than patients with P3 disease. (about 88%) and more than 90% were Borrmann types
Besides, compared with nonresection, gastrectomy was III and IV. In approximately 95% of our cases, the adeno-
beneficial for patients with P1 and P2 diseases rather than carcinoma was undifferentiated. Because of relatively poor
P3 disease. These results were also similar to the previous blood circulation and low drug concentration in the peri-
reports [15, 27, 28]. toneum, tumor cells spreading to the peritoneum have
Previous studies reported that gastric resection not only less response to the general systemic chemotherapy than
could prolong the survival of gastric cancer patients with other organs, which makes it an obstacle of treatment.
PC but also reduced symptoms and improved life quality Recently, the implementation of a multimodality treatment
without increasing the mortality rate [17, 23]. Theoretically, including noncurative gastretomy combined with HIPEC
it is believed that cancer patients can benefit from the has led to promising results in selected gastric adenocar-
removal of tumor mass through multiple ways. First, it cinoma patients with PC [20, 31–34]. A systematic review
can reduce the local complications caused by the primary and meta-­analysis of 13 randomized controlled trials with
disease; second, it can reduce the resource of additional acceptable quality have been established. It was concluded
metastasis; third, when the tumor load is removed, the that HIPEC was associated with marked improvement in
residual cancer cells might be more sensitive to the chemo- survival of advanced gastric cancer, in comparison with
therapeutics, making the response to chemotherapy other current standard treatments [35].
increased; fourth, gastrectomy can relieve the metabolic In this study, we also analyzed the effects of different
demands from the tumor; and finally, patients may profit treatment modalities on patient’s survival. We found that
immunologically from decreased tumor burden as the patients in gastrectomy group with either chemotherapy
cancer cells can produce some immunosuppressive plus HIPEC or chemotherapy had statistically significant
cytokines [25, 29]. In spite of these theoretical advantages better survival rates than those in the nonresection group
and a few reports on the patients’ survival benefits, gas- (MST: 17.0 vs. 13.0 months, and 15.0 vs. 7.0 months).
trectomy for gastric adenocarcinoma patients with PC still However, these differences were not observed in patients
remains a controversy and there are debates on the treat- without chemotherapy in gastrectomy group and nonre-
ment modality in the literature. The REGATTA trial by section group (MST: 7.0 vs. 7.0 months). These results
Fujitani et al. [21] investigated the role of gastrectomy indicated that patients did not benefit from gastrectomy
in management of advanced gastric cancer with a single without chemotherapy. We also compared the survival
incurable factor. The results showed that gastrectomy fol- rates of patients who accepted the six different treatment
lowed by chemotherapy did not display any survival benefit. strategies (Table 2). The results showed that the treatment
They reported a surprisingly high median OS modality of gastrectomy plus chemotherapy plus HIPEC
(16.6 months) for patients assigned to chemotherapy alone (MST: 17.0 months) had the optimum survival when
group, which was much higher than all the previous compared with gastrectomy plus chemotherapy (MST:
reported eight-­phase III trials performed from 2005 to 15.0 months) and gastrectomy alone (MST: 7.0 months)
2014 for stage IV gastric cancer treatment [30]. Moreover, in the gastrectomy group and the three counterparts in
they reported a 14.3-­month median OS for those assigned the nonresection group. These results were similar to the
to gastrectomy plus chemotherapy. The REGATTA trial report of Rudloff et al. in the GYMSSA trial [20], in
was a prospective study on stage IV gastric cancer with which cytoreductive surgery combined with intraperitoneal
different metastatic sites, using a different chemotherapy and systemic chemotherapy in selected patients with PC
regimen. There were no results of subgroup analysis for achieved a prolonged survival. Another prospective ran-
PC in the REGATTA trial. We remain in doubt if the domized study conducted by Yang et al. [31] reported
different conclusions between REGATTA trial and all the that patients in the gastrectomy plus HIPEC group had
other studies were caused by the different chemotherapy a longer median survival than those in the gastrectomy
regimens or way of patients’ selection. For example, in alone group (11.0 vs. 6.5 months) and several cohort
our study the gastric cancer patients were with PC and studies also demonstrated that gastrectomy plus HIPEC
no other sites of metastasis. could improve outcome of gastric cancer patients with
Gastric cancer patients with PC only are a specific PC [32, 33, 36]. The combination of treatment modalities
population. The tumor cells either directly perforate the (gastrectomy plus chemotherapy plus HIPEC) had showed
entire gastric wall or through other ways to spread to advantages and was currently favored by more oncological

2788 © 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.
X. Geng et al. Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis

surgeon while a panel of international experts strongly is still largely relied on the operator’s experience. There
recommended cytoreductive surgery plus HIPEC to be is no standard protocol for the treatment of these patients.
the current standard treatment for advanced gastric cancer According to the solid data provided in the results of
(GC) [37, 38]. our study, the treatment strategy should be decided based
In our study, interestingly, when we analyzed the sur- on the extent of peritoneal seeding (P stage) and the
vivals in different subgroups within the nonresection group findings during an exploratory laparotomy (suitable for
patients, we found that the chemotherapy plus HIPEC resection or not).
subgroup had a more significant survival benefit than the In conclusion, the gastrectomy plus chemotherapy plus
chemotherapy subgroup (MST: 13.0 vs. 7.0 months) and HIPEC modality showed a significant survival benefit for
no chemotherapy subgroup (MST: 13.0 vs. 7.0 months), gastric adenocarcinoma patients, particularly patients with
which suggested that HIPEC was beneficial to the non- P1 and P2 diseases, but not for P3 patients. HIPEC was
resection patients. In the gastrectomy group, we observed also beneficial for the nonresection patients. These results
a significant prolonged survival when the chemotherapy may shed light on the role of adjuvant surgery on the
plus HIPEC subgroup was compared with the no chemo- treatment of stage IV gastric cancer with PC.
therapy and no HIPEC subgroup (MST: 17.0 vs.
7.0 months). However, in the gastrectomy group, only a
2-­month difference in MST was observed when the chemo- Conflicts of Interest
therapy plus HIPEC subgroup was compared with the The authors have declared no conflicts of interest.
chemotherapy subgroup (MST: 17.0 vs. 15.0 months). We
believe that the effect of HIPEC observed in the nonre- References
section group was due to the tumor mass in the peritoneal
 1. Jemal, A., R. Siegel, J. Xu, and E. Ward. 2010. Cancer
cavity. When the mass was removed, the survival benefit
statistics, 2010. CA Cancer J. Clin. 60:277–300.
of HIPEC observed by Yang [31] was masked by the
 2. Chen, S., Y. F. Li, X. Y. Feng, Z. W. Zhou, X. H.
effects of systemic chemotherapy in the gastrectomy group
Yuan, and Y. B. Chen. 2012. Significance of palliative
in our study.
gastrectomy for late-­stage gastric cancer patients. J.
The limitations of our study are as follow: first, this
Surg. Oncol. 106:862–871.
study is a retrospective study and it is well known that
 3. Chen, W., R. Zheng, P. D. Baade, S. Zhang, H. Zeng,
possible detection bias and performance of analysis bias
F. Bray, et al. 2016. Cancer statistics in China, 2015.
might exist. Our patients might have received a variety
CA Cancer J. Clin. 66:115–132.
of treatments, including gastrojejunostomy or noncurative
 4. Yonemura, Y., Y. Endou, M. Shinbo, T. Sasaki, M.
resection with or without preoperative and/or palliative
Hirano, A. Mizumoto, et al. 2009. Safety and efficacy of
chemotherapy, which may be relevance to the survival,
bidirectional chemotherapy for treatment of patients with
although we did a very strict selection. The possibility of peritoneal dissemination from gastric cancer: selection for
selection bias was also existed, that meant, less severe cytoreductive surgery. J. Surg. Oncol. 100:311–316.
cases might have been selected for surgical resection, which  5. Bando, E., Y. Yonemura, Y. Takeshita, K. Taniguchi, T.
would have resulted in a better survival outcome. To Yasui, Y. Yoshimitsu, et al. 1999. Intraoperative lavage
overcome these problems and obtain more conclusive for cytological examination in 1297 patients with gastric
results, a well-­designed, large-­scale randomized controlled carcinoma. Am. J. Surg. 178:256–262.
trial is needed to explore the survival benefit of gastrec-  6. Boku, T., Y. Nakane, T. Minoura, H. Takada, M.
tomy plus chemotherapy plus HIPEC compared with Yamamura, K. Hioki, et al. 1990. Prognostic significance
chemotherapy alone for gastric adenocarcinoma patients of serosal invasion and free intraperitoneal cancer cells
with PC. in gastric cancer. Br. J. Surg. 77:436–439.
For the classifications of PC, we used the first edition  7. Roviello, F., S. Caruso, D. Marrelli, C. Pedrazzani, A.
of the General Rules for Gastric Cancer Study that was Neri, A. De Stefano, et al. 2011. Treatment of
considered too rough and not enough to quantify to a peritoneal carcinomatosis with cytoreductive surgery and
certain extent. The Peritoneal Cancer Index, described by hyperthermic intraperitoneal chemotherapy: state of the
Jacquet and Sugarbaker [7, 39, 40], is a system of staging art and future developments. Surg. Oncol. 20:e38–e54.
PC and able to assess the carcinomatosis quantification  8. Wagner, A. D., S. Unverzagt, W. Grothe, G. Kleber, A.
and distribution. However, the peritoneal cancer index Grothey, J. Haerting, et al. 2010. Chemotherapy for
classification might be too complicated to be widely used. advanced gastric cancer. Cochrane Database Syst. Rev.
Therefore, a more practical classification for PC is expected. 3:CD004064.
In China, up to now, when surgeons encounter a gastric  9. Koizumi, W., H. Narahara, T. Hara, A. Takagane, T.
cancer patient with PC, the decision of treatment strategy Akiya, M. Takagi, et al. 2008. S-­1 plus cisplatin versus

© 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd. 2789
Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis X. Geng et al.

S-­1 alone for first-­line treatment of advanced gastric gastric cancer with a single non-­curable factor
cancer (SPIRITS trial): a phase III trial. Lancet Oncol. (REGATTA): a phase 3, randomised controlled trial.
9:215–221. Lancet Oncol. 17:309–318.
10. Huang, J., Y. Zhao, Y. Xu, Y. Zhu, J. Huang, Y. Liu, 22. Japanese research society for gastric cancer. 1995.
et al. 2016. Comparative effectiveness and safety between Macroscopic findings. Pp. 3–14 in M. Nishi, Y. Omori,
oxaliplatin-based and cisplatin-based therapy in advanced K. Miwa, eds. Japanese classification of gastric
gastric cancer: a meta-analysis of randomized controlled carcinoma. 1st Englished. Kanehara & Co. Ltd., Tokyo.
trials Oncotarget doi: 10.18632/oncotarget.9189. 23. Ko, K. J., J. H. Shim, H. M. Yoo, O. SI, H. M. Jeon,
11. Japanese Gastric Cancer Association. 2011. Japanese C. H. Park, et al. 2012. The clinical value of non-­
classification of gastric carcinoma: 3rd English edition. curative resection followed by chemotherapy for
Gastric Cancer 14:101–112. incurable gastric cancer. World J. Surg. 36:1800–1805.
12. Ajani, J. A., D. J. Bentrem, S. Besh, T. A. D’Amico, P. 24. Hioki, M., N. Gotohda, M. Konishi, T. Nakagohri, S.
Das, C. Denlinger, et al. 2013. Gastric cancer, version Takahashi, and T. Kinoshita. 2010. Predictive factors
2.2013: featured updates to the NCCN Guidelines. J. improving survival after gastrectomy in gastric cancer
Natl. Compr. Canc. Netw. 11:531–546. patients with peritoneal carcinomatosis. World J. Surg.
13. Sugarbaker, P. H. 2010. Surgical responsibilities in the 34:555–562.
management of peritoneal carcinomatosis. J. Surg. 25. Kim, S. W. 2014. The result of conversion surgery in
Oncol. 101:713–724. gastric cancer patients with peritoneal seeding. J.
14. Chang, Y. R., D. S. Han, S. H. Kong, H. J. Lee, S. H. Gastric. Cancer 14:266–270.
Kim, W. H. Kim, et al. 2012. The value of palliative 26. Li, C., M. Yan, J. Chen, M. Xiang, Z. G. Zhu, H. R.
gastrectomy in gastric cancer with distant metastasis. Yin, et al. 2010. Survival benefit of non-­curative
Ann. Surg. Oncol. 19:1231–1239. gastrectomy for gastric cancer patients with synchronous
15. Xia, X., C. Li, M. Yan, B. Liu, X. Yao, and Z. Zhu. distant metastasis. J. Gastrointest. Surg. 14:282–288.
2014. Who will benefit from noncurative resection in 27. Yang, K., K. Liu, W. H. Zhang, Z. H. Lu, X. Z. Chen,
patients with gastric cancer with single peritoneal X. L. Chen, et al. 2015. The value of palliative
metastasis? Am. Surg. 80:124–130. gastrectomy for gastric cancer patients with
16. Sun, J., Y. Song, Z. Wang, X. Chen, P. Gao, Y. Xu, intraoperatively proven peritoneal seeding. Medicine
et al. 2013. Clinical significance of palliative gastrectomy (Baltimore) 94:e1051.
on the survival of patients with incurable advanced 28. Gretschel, S., R. Siegel, L. Estevez-Schwarz, M.
gastric cancer: a systematic review and meta-­analysis. Hunerbein, U. Schneider, and P. M. Schlag. 2006.
BMC Cancer 13:577. Surgical strategies for gastric cancer with synchronous
17. Lasithiotakis, K., S. A. Antoniou, G. A. Antoniou, I. peritoneal carcinomatosis. Br. J. Surg. 93:1530–1535.
Kaklamanos, and O. Zoras. 2014. Gastrectomy for stage 29. McCarter, M. D., and Y. Fong. 2001. Role for surgical
IV gastric cancer. A systematic review and meta-­analysis. cytoreduction in multimodality treatments for cancer.
Anticancer Res. 34:2079–2085. Ann. Surg. Oncol. 8:38–43.
18. Gold, J. S., D. P. Jaques, D. J. Bentrem, M. A. Shah, L. 30. Izuishi, K., and H. Mori. 2016. Recent Strategies for
H. Tang, M. F. Brennan, et al. 2007. Outcome of treating stage IV gastric cancer: roles of palliative
patients with known metastatic gastric cancer gastrectomy, chemotherapy, and radiotherapy. J.
undergoing resection with therapeutic intent. Ann. Surg. Gastrointestin. Liver Dis. 25:87–94.
Oncol. 14:365–372. 31. Yang, X. J., C. Q. Huang, T. Suo, L. J. Mei, G. L.
19. Moehler, M., P. R. Galle, I. Gockel, T. Junginger, and Yang, F. L. Cheng, et al. 2011. Cytoreductive surgery
H. Schmidberger. 2007. The multidisciplinary and hyperthermic intraperitoneal chemotherapy improves
management of gastrointestinal cancer. Multimodal survival of patients with peritoneal carcinomatosis from
treatment of gastric cancer. Best Pract. Res. Clin. gastric cancer: final results of a phase III randomized
Gastroenterol. 21:965–981. clinical trial. Ann. Surg. Oncol. 18:1575–1581.
20. Rudloff, U., R. C. Langan, J. E. Mullinax, J. D. Beane, S. 32. Glehen, O., F. N. Gilly, C. Arvieux, E. Cotte, F.
M. Steinberg, T. Beresnev, et al. 2014. Impact of Boutitie, B. Mansvelt, et al. 2010. Peritoneal
maximal cytoreductive surgery plus regional heated carcinomatosis from gastric cancer: a multi-­institutional
intraperitoneal chemotherapy (HIPEC) on outcome of study of 159 patients treated by cytoreductive surgery
patients with peritoneal carcinomatosis of gastric origin: combined with perioperative intraperitoneal
results of the GYMSSA trial. J. Surg. Oncol. 110:275–284. chemotherapy. Ann. Surg. Oncol. 17:2370–2377.
21. Fujitani, K., H. K. Yang, J. Mizusawa, Y. W. Kim, M. 33. Scaringi, S., R. Kianmanesh, J. M. Sabate, E. Facchiano,
Terashima, S. U. Han, et al. 2016. Gastrectomy plus P. Jouet, B. Coffin, et al. 2008. Advanced gastric cancer
chemotherapy versus chemotherapy alone for advanced with or without peritoneal carcinomatosis treated with

2790 © 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd.
X. Geng et al. Gastrectomy for Gastric Cancer with Peritoneal Carcinomatosis

hyperthermic intraperitoneal chemotherapy: a single 37. Sugarbaker, P. H., W. Yu, and Y. Yonemura. 2003.
western center experience. Eur. J. Surg. Oncol. Gastrectomy, peritonectomy, and perioperative
34:1246–1252. intraperitoneal chemotherapy: the evolution of treatment
34. Yonemura, Y., T. Kawamura, E. Bandou, S. Takahashi, strategies for advanced gastric cancer. Semin. Surg.
T. Sawa, and N. Matsuki. 2005. Treatment of peritoneal Oncol. 21:233–248.
dissemination from gastric cancer by peritonectomy and 38. Boerner, T., A. Graichen, T. Jeiter, F. Zemann, P.
chemohyperthermic peritoneal perfusion. Br. J. Surg. Renner, L. März, et al. 2016. CRS-­HIPEC prolongs
92:370–375. survival but is not curative for patients with peritoneal
35. Yan, T. D., D. Black, P. H. Sugarbaker, J. Zhu, Y. carcinomatosis of gastric cancer. Ann. Surg. Oncol.
Yonemura, G. Petrou, et al. 2007. A systematic review doi:10.1245/s10434-016-5306-0.
and meta-­analysis of the randomized controlled trials 39. Yonemura, Y., E. Bandou, T. Kawamura, Y. Endou, and
on adjuvant intraperitoneal chemotherapy for resectable T. Sasaki. 2006. Quantitative prognostic indicators of
gastric cancer. Ann. Surg. Oncol. 14:2702–2713. peritoneal dissemination of gastric cancer. Eur. J. Surg.
36. Glehen, O., V. Schreiber, E. Cotte, A. C. Sayag- Oncol. 32:602–606.
Beaujard, D. Osinsky, G. Freyer, et al. 2004. 40. Sugarbaker, P. H. 2001. Cytoreductive surgery and
Cytoreductive surgery and intraperitoneal peri-­operative intraperitoneal chemotherapy as a curative
chemohyperthermia for peritoneal carcinomatosis arising approach to pseudomyxoma peritonei syndrome. Eur. J.
from gastric cancer. Arch. Surg. 139:20–26. Surg. Oncol. 27:239–243.

© 2016 The Authors. Cancer Medicine published by John Wiley & Sons Ltd. 2791