Bjs 10881

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Original article

Detection of carcinoembryonic antigen in peritoneal fluid


of patients undergoing laparoscopic distal gastrectomy
with complete mesogastric excision

D. Xie1 , Y. Wang1 , J. Shen1 , J. Hu1 , P. Yin2 and J. Gong1


1
Department of Gastrointestinal Surgery, Tongji Hospital, Tongji Medical College, and 2 Department of Epidemiology and Biostatistics, School of Public
Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
Correspondence to: Professor J. Gong, Department of Gastrointestinal Surgery, Tongji Hospital, Tongji Medical College, Huazhong University of Science
and Technology, Wuhan, 430030, China (e-mail: [email protected])

Background: Surgery for gastric cancer may result in free intraperitoneal cancer cells. This study aimed
to determine whether laparoscopic gastrectomy with complete mesogastric excision (D2 + CME) reduces
the number of free intraperitoneal cancer cells.
Methods: Patients with gastric cancer who had a conventional D2 or D2 + CME laparoscopic distal
gastrectomy between April 2015 and February 2017 were included in the study. Intraoperative peritoneal
washings were collected before and after tumour resection. Reverse transcriptase–quantitative real-time
PCR for carcinoembryonic antigen (CEA) was used to assess the presence of gastric cancer cells.
Results: Eighty-five patients underwent conventional D2 lymphadenectomy and 76 had the D2 + CME
procedure. Of 161 peritoneal fluid samples obtained before gastrectomy, 137 (D2, 72; D2 + CME, 65) had
low CEA expression indicative of no cancer cells. After gastrectomy, high CEA expression was detected
in 23 of the 72 samples (32 per cent) from patients in the D2 group, and in ten of the 65 samples (15 per
cent) from the D2 + CME group. In the overall cohort, mean CEA expression level after gastrectomy was
lower in the D2 + CME group than in the D2 group (P = 0⋅0038). In patients with low CEA expression
before gastrectomy, disease-free survival in the D2 + CME group was better than that in the D2 group
(P = 0⋅033).
Conclusion: Laparoscopic distal gastrectomy with complete mesogastric excision reduces the number of
free intraperitoneal cancer cells and is associated with a better disease-free survival than conventional D2
gastrectomy.

Abstract presented to the 89th Annual Meeting of the Japanese Gastric Cancer Association,
Hiroshima, Japan, March 2017

Paper accepted 29 March 2018


Published online 2 July 2018 in Wiley Online Library (www.bjs.co.uk). DOI: 10.1002/bjs.10881

Introduction procedures, and these cells may be the main source of


tumour relapse and cancer-related death8 – 11 .
Advanced gastric cancer is a leading cause of death glob- Free intraperitoneal cancer cells can be detected with the
ally, particularly in East Asia1 . Radical surgery is still the reverse transcriptase–quantitative real-time (RT–qPCR)
main treatment modality2,3 . According to the Japanese method, by measuring cancer-specific antigens12 . Carcino-
gastric cancer treatment guidelines4 , the standard oper- embryonic antigen (CEA) is one of the most com-
ation for advanced gastric cancer is radical gastrectomy monly used markers for gastric cancer10,11,13,14 , and CEA
with D2 lymphadenectomy. However, up to 38 per cent RT–qPCR assay is an independent predictor of tumour
of patients develop recurrent disease 5 years after surgery5 . recurrence12 . It is estimated11 that cancer-specific antigens
The underlying cause of disease recurrence is unknown, can be detected after tumour dissection in more than 60
but it has been postulated that free intraperitoneal can- per cent of patients with gastric cancer.
cer cells may play an important role6,7 . Tumour cells may Previous studies have demonstrated the presence of
spill into the peritoneal cavity during conventional surgical cancer cells in the mesogastrium15,16 . As a consequence,

© 2018 BJS Society Ltd BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
1472 D. Xie, Y. Wang, J. Shen, J. Hu, P. Yin and J. Gong

Table 1 Clinicopathological characteristics of all patients Gastrointestinal Surgery, Tongji Hospital, Wuhan, China,
D2 + CME
between April 2015 and February 2017 were identified
D2 (n = 85) (n = 76) P† from an institutional database. The database included
Age (years)* 54⋅0(9⋅4) 53⋅3(9⋅3) 0⋅660‡
patients who had been entered into an ongoing RCT
Sex ratio (M : F) 54 : 31 47 : 29 0⋅825 (NCT01978444) comparing laparoscopic D2 + CME
Tumour size (cm2 )* 7⋅0(6⋅1) 8⋅0(8⋅6) 0⋅418‡ gastrectomy with conventional D2 laparoscopic gastrec-
Main tumour location 0⋅062 tomy. Patients who met following criteria were excluded:
Body 28 (33) 36 (47)
Antrum/pylorus 57 (67) 40 (53)
unresectable gastric cancer, age less than 18 years or more
Specific tumour location 0⋅558§ than 75 years, BMI above 30 kg/m2 , presence of other
Anterior 6 (7) 4 (5) malignant disease, previous upper abdominal surgery, and
Posterior 12 (14) 8 (11) preoperative neoadjuvant chemotherapy or radiotherapy.
Lesser curvature 51 (60) 52 (68)
Informed consent was obtained from all patients, and the
Greater curvature 4 (5) 6 (8)
Annular 12 (14) 6 (8) study was approved by the ethics committee of Tongji
pT category 0⋅977¶ Hospital.
pT1 9 (11) 6 (8)
pT2 19 (22) 16 (21)
pT3 44 (52) 47 (62) Surgical procedures and follow-up
pT4 13 (15) 7 (9)
pN category 0⋅148¶ Conventional laparoscopic D2 lymphadenectomy was per-
pN0 27 (32) 32 (42) formed according to the Japanese gastric cancer treatment
pN1 16 (19) 15 (20)
pN2 15 (18) 10 (13)
guidelines4,19 . Briefly, the stomach was mobilized by dis-
pN3 27 (32) 19 (25) section of the greater and lesser omentum, lymph node
pTNM stage 0⋅282¶ stations 1, 3, 4sb, 4d, 5, 6, 7, 8a, 9, 11p and 12a, and the
I 13 (15) 16 (21) gastrosplenic ligaments.
II 36 (42) 33 (43)
III 36 (42) 27 (36)
Laparoscopic D2 + CME was performed as described
Histological type 0⋅753 previously17,18 . Briefly, the CME procedure should meet
Well or moderately 15 (18) 12 (16) the following criteria: clear exposure of five mesenter-
differentiated ies of stomach; complete mobilization of the mesentery
Poorly differentiated 70 (82) 64 (84)
covered by an intact visceral fascial layer; dissection of
Values in parentheses are percentages unless indicated otherwise; *values the mesentery and ligation of blood vessels at the root.
are mean(s.d.). D2, conventional D2 lymphadenectomy; D2 + CME, D2
Details of the D2 + CME approach is provided in Fig. S1,
lymphadenectomy with complete mesogastric excision. †Pearson’s χ2 test,
except ‡Student’s t test, §Fisher’s exact test and ¶Mann–Whitney U test. Appendix S1 and Video S1 (supporting information). All
laparoscopic procedures were recorded for subsequent
laparoscopic D2 lymphadenectomy plus complete meso- evaluation. The reconstruction method (Billroth I, Billroth
gastric excision (D2 + CME) was proposed as a new II or Roux-en-Y) was left to the surgeon’s discretion.
concept in the surgical treatment of advanced gastric Patients were followed up every 3 months. Postopera-
cancer17,18 . By dissecting along the surgical planes and tive surveillance consisted of physical examination, blood
embryonic boundary of the mesogastrium, complete exci- chemistry (including CEA, carbohydrate antigen (CA) 19-9
sion of the proximal segment of dorsal mesogastrium and CA72-4), CT and gastroscopy.
can be achieved, with minimal blood loss and improved
short-term surgical outcomes18 .
Peritoneal lavage sample collection
The aim of the present study was to compare the presence
of CEA, representing free intraperitoneal cancer cells, in In brief, 150 ml of physiological saline was introduced
peritoneal lavage fluid from patients who had laparoscopic in the upper abdominal cavity and aspirated after gentle
D2 + CME distal gastrectomy with that in patients who stirring, both at the beginning of the operation, before
underwent conventional D2 lymphadenectomy. any manipulation, and immediately after tumour dis-
section. Fluids were aspirated and centrifuged at 717 g
Methods for 5 min to collect intact cells. After removing red
blood cells by Red Blood Cell Lysis Buffer (Beyotime,
All patients with advanced gastric cancer (pathological Haimen, China), the concentrated fluid samples were dis-
stage IIa–IIIc) who underwent curative laparoscopy- solved in RNAiso Plus (Takara, Tokyo, Japan) and stored
assisted distal gastrectomy at the Department of at −80 ∘ C.

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
Carcinoembryonic antigen in peritoneal fluid of patients with gastric cancer 1473

*
100 100

10–1 10–1
CEA mRNA/GAPDH mRNA

CEA mRNA/GAPDH mRNA


10–2 10–2

10–3 10–3

10–4 10–4

10–5 10–5

10–6 10–6
D2 D2 + CME D2 D2 + CME

a All patients, CEA level before surgery b All patients, CEA level after surgery


100 100

10–1 10–1

CEA mRNA/GAPDH mRNA


CEA mRNA/GAPDH mRNA

10–2 10–2

10–3 10–3

10–4 10–4

10–5 10–5

10–6 10–6
D2 D2 + CME D2 D2 + CME

c pT3 status, CEA level before surgery d pT3 status, CEA level after surgery

Fig. 1Carcinoembryonic antigen (CEA) expression in peritoneal lavage samples. a CEA levels before surgery in all patients; CEA
mRNA expression was low in all 72 patients who had a conventional D2 lymphadenectomy (D2) and low in all 65 who had
lymphadenectomy with complete mesogastric excision (D2 + CME). b CEA levels after surgery in all patients; CEA mRNA was high in
23 of 72 patients (32 per cent) in the D2 group and high in ten of 65 patients (15 per cent) in the D2 + CME group. c CEA levels before
surgery in patients with pT3 disease; CEA mRNA expression was low in all 40 patients in the D2 group and in all 39 in the D2 + CME
group. d CEA levels after surgery in patients with pT3 disease; CEA mRNA expression was high in 13 of 40 patients (33 per cent) in
the D2 group and high in five of 39 patients (13 per cent) in the D2 + CME group. GAPDH, glyceraldehyde-3-phosphate
dehydrogenase. *P = 0⋅038, †P = 0⋅035 (Mann–Whitney U test)

Reverse transcriptase–quantitative real-time PCR using the ABI PRISM® 7300 Sequence Detection system
Total RNA was extracted from peritoneal lavage sample (Applied Biosystems, Foster City, California, USA) in
with RNAiso Plus, and cDNA was produced with the accordance with the manufacturer’s protocol. Expression
PrimeScript™ RT Master Mix Kit (Takara) in accordance of CEA mRNA was normalized by the internal control,
with manufacturer’s instructions. GAPDH.
Real-time quantitative PCR (qPCR) was performed
Determination of carcinoembryonic antigen
for CEA using the following primer sequences with
cut-off values
SYBR® Premix Ex Taq™ Kit (Takara): for CEA: forward
5′ -AGTCTATGCAGAGCCACCCAA-3′ , reverse 5′ -GG To determine the cut-off value, peritoneal washing buffer
GAGGCTCTGATTATTTACCCA-3′ ; for glyceralde was collected from 11 patients who underwent a dis-
hyde-3-phosphate dehydrogenase (GAPDH): forward 5′ - tal subtotal gastrectomy for benign gastric neoplasia20 .
CGAGATCCCTCCAAAATCAA-3′ , reverse 5′ -TTCA Cut-off values for preoperative and postoperative CEA
CACCCATGACGAACAT-3′ . qPCR was carried out levels were determined as the mean plus two standard

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
1474 D. Xie, Y. Wang, J. Shen, J. Hu, P. Yin and J. Gong

Table 2 Association between carcinoembryonic antigen Table 3 Comparison of clinicopathological characteristics of


expression in peritoneal fluid before and after surgery and pT patients with high and low postoperative carcinoembryonic
status in the two surgical groups of patients with low antigen expression in peritoneal fluid
preoperative expression
CEAL/L CEAL/H
pT status D2 (n = 72) D2 + CME (n = 65) P* (n = 104) (n = 33) P†

Overall 0⋅024 Age (years)* 54⋅3(9⋅2) 54⋅0(9⋅4) 0⋅906‡


CEAL/L 49 (68) 55 (85) Sex 0⋅308
CEAL/H 23 (32) 10 (15) M 67 (79) 18 (21)
pT1 1⋅000† F 37 (71) 15 (29)
CEAL/L 8 (89) 6 (100) Tumour size (cm2 )* 7⋅3(7⋅5) 6⋅5(3⋅6) 0⋅534‡
CEAL/H 1 (11) 0 (0) Main tumour location 0⋅991
pT2 0⋅330† Body 44 (76) 14 (24)
CEAL/L 11 (73) 14 (93) Antrum/pylorus 60 (76) 19 (24)
CEAL/H 4 (27) 1 (7) Specific tumour location 0⋅358§
pT3 0⋅037 Anterior 7 (78) 2 (22)
CEAL/L 27 (68) 34 (87) Posterior 14 (78) 4 (22)
CEAL/H 13 (33) 5 (13) Lesser curvature 67 (79) 18 (21)
pT4 1⋅000† Greater curvature 8 (80) 2 (20)
CEAL/L 3 (38) 1 (20) Annular 8 (53) 7 (47)
CEAL/H 5 (63) 4 (80) pT category 0⋅001¶
pT1 14 (93) 1 (7)
Values in parentheses are percentages. D2, conventional D2 pT2 25 (83) 5 (17)
lymphadenectomy; D2 + CME, D2 lymphadenectomy with complete pT3 61 (77) 18 (23)
mesogastric excision; CEAL/L , carcinoembryonic antigen expression low pT4 4 (31) 9 (69)
before and after surgery; CEAL/H , CEA expression low before and high pN category 0⋅060¶
after surgery. *Pearson’s χ2 test, except †Fisher’s exact test. pN0 43 (83) 9 (17)
pN1 21 (72) 8 (28)
deviations based on relative CEA expression levels of the pN2 21 (88) 3 (13)
pN3 19 (59) 13 (41)
negative controls (1⋅908 × 10−2 and 1⋅067 × 10−3 respec- pTNM stage 0⋅002¶
tively) (Fig. S2, supporting information). When the CEA I 23 (92) 2 (8)
level in the patient’s sample was higher than the cut-off II 52 (80) 13 (20)
value, the sample was defined as CEA-high (CEAH ); oth- III 29 (62) 18 (38)
Histological type 0⋅249
erwise it was defined as CEA-low (CEAL ). Well or moderately differentiated 22 (85) 4 (15)
Poorly differentiated 82 (74) 29 (26)
Statistical analysis Surgical procedure 0⋅024
D2 49 (68) 23 (32)
Statistical analyses of continuous variables were performed D2 + CME 55 (85) 10 (15)
with Student’s t test or the Mann–Whitney U test, and cat- Poorly differentiated 70 (82) 64 (84)

egorical variables were analysed using the χ2 or Fisher’s Values in parentheses are percentages unless indicated otherwise; *values
exact test, as appropriate. Ordinal variables were ana- are mean(s.d.). CEAL/L , carcinoembryonic antigen expression low before
and after surgery; CEAL/H , CEA expression low before and high after
lysed with the Mann–Whitney U test. The Kaplan–Meier
surgery; D2, conventional D2 lymphadenectomy; D2 + CME, D2
method, log rank test and two-stage test were used for sur- lymphadenectomy with complete mesogastric excision. †Pearson’s χ2 test,
vival analysis. Overall survival (OS) and disease-free sur- except ‡Student’s t test, §Fisher’s exact test and ¶Mann–Whitney U test.
vival (DFS) were calculated from the date of operation until
date of death/recurrence or last follow-up. Risk factors for patients participated in the RCT comparing laparoscopic
recurrence and survival were evaluated by univariable and D2 + CME (76 patients) and conventional D2 laparoscopic
multivariable analyses using logistic regression models (for- gastrectomy (85). Clinicopathological characteristics of the
ward likelihood ratio). P < 0⋅050 was considered statisti- patients are shown in Table 1. There was no statistically
cally significant. All statistical tests were performed with significant difference in characteristics between the two
SPSS® version 19.0 software (IBM, Armonk, New York groups.
USA) and R version 3.4.1 (http://www.r-project.org/).
Carcinoembryonic antigen expression in peritoneal
Results lavage
A total of 802 patients with advanced gastric cancer were RT–qPCR was performed on all the samples. Before
identified and 161 patients were included in the study. All gastrectomy, mean(s.d.) CEA mRNA expression

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
Carcinoembryonic antigen in peritoneal fluid of patients with gastric cancer 1475

Table 4 Univariable and multivariable analysis of patients with high and low postoperative carcinoembryonic antigen expression in the
peritoneal fluid

Univariable analysis Multivariable analysis

Odds ratio P Odds ratio P

Age 1⋅00 (0⋅96, 1⋅04) 0⋅905


Sex 0⋅310
M 1⋅00 (reference)
F 1⋅51 (0⋅68, 3⋅34)
Tumour size 0⋅98 (0⋅92, 1⋅05) 0⋅533
Main tumour location 0⋅991
Body 1⋅00 (reference)
Antrum/pylorus 1⋅01 (0⋅46, 2⋅22)
Specific tumour location 0⋅359
Anterior 1⋅00 (reference)
Posterior 1⋅00 (0⋅146, 6⋅85)
Lesser curvature 0⋅94 (0⋅18, 4⋅92)
Greater curvature 0⋅88 (0⋅95, 7⋅95)
Annular 3⋅06 (0⋅47, 19⋅88)
pT category 0⋅003 0⋅004
pT1 1⋅00 (reference) 1⋅00 (reference)
pT2 2⋅80 (0⋅30, 26⋅42) 3⋅14 (0⋅33, 30⋅26)
pT3 4⋅13 (0⋅51, 33⋅59) 4⋅68 (0⋅57, 38⋅73)
pT4 31⋅50 (3⋅02, 328⋅93) 35⋅81 (3⋅30, 388⋅32)
pN category 0⋅056
pN0 1⋅00 (reference)
pN1 1⋅82 (0⋅61, 5⋅39)
pN2 0⋅68 (0⋅17, 2⋅79)
pN3 3⋅27 (1⋅19, 8⋅95)
pTNM stage 0⋅015
I 1⋅00 (reference)
II 2⋅88 (0⋅60, 13⋅79)
III 7⋅14 (1⋅50, 33⋅97)
Histological type 0⋅255
Well or moderately differentiated 1⋅00 (reference)
Poorly differentiated 0⋅51 (0⋅16, 1⋅62)
Surgical procedure 0⋅026 0⋅027
D2 1⋅00 (reference) 1⋅00 (reference)
D2 + CME 0⋅39 (0⋅17, 0⋅90) 0⋅36 (0⋅15, 0⋅89)

Values in parentheses are 95 per cent confidence intervals. D2, conventional D2 lymphadenectomy; D2 + CME, D2 lymphadenectomy with complete
mesogastric excision.

was similar in the conventional D2 and D2 + CME level (CEAL/H ), compared with ten (15 per cent) of the 65
groups: 1⋅8(3⋅7) × 10−3 versus 2⋅0(3⋅3) × 10−3 respectively patients in the D2 + CME group (P = 0⋅024) (Table 2).
(P = 0⋅232) (Fig. 1a). In patients with pT3 disease there was In patients with pT3 tumours, the proportion with
also no significant difference in CEA expression (P = 0⋅486) CEAL/H in the D2 group was 33 per cent (13 of 40) and
(Fig. 1c). Mean CEA expression following gastrectomy in that in the D2 + CME group was 13 per cent (5 of 39)
the D2 group was higher than in the D2 + CME group: (P = 0⋅037) (Table 2).
7⋅4 × 10−3 (s.d. 2⋅0 × 10−2 ) versus 2⋅6 × 10−3 (1⋅2 × 10−2 )
(P = 0⋅038) (Fig. 1b). This was also the case for patients Relationship between clinicopathological
with pT3 tumours (P = 0⋅035) (Fig. 1d). characteristics and carcinoembryonic antigen
Based on the cut-off value, high CEA expression was seen expression
before surgery in 13 of the 85 patients (15 per cent) in Clinicopathological characteristics of patients with low
the D2 group, and in 11 of the 76 patients (14 per cent) CEA levels before and after surgery (CEAL/L ) and those
in the D2 + CME groups. After excluding these patients, with low expression before and high expression after
23 (32 per cent) of the 72 patients in the D2 group with surgery (CEAL/H ) are compared in Table 3. There was a sta-
low preoperative CEA expression had a high postoperative tistically significant association between CEAL/H and depth

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
1476 D. Xie, Y. Wang, J. Shen, J. Hu, P. Yin and J. Gong

100 100

80 80
Disease-free survival (%)

Overall survival (%)


60 60
CEAL/L CEAL/L
CEAL/H CEAL/H
CEAH/H CEAH/H
40 40

20 20

0 6 12 18 0 6 12 18
Time after surgery (months) Time after surgery (months)
No. at risk No. at risk
CEAL/L 104 92 64 30 CEAL/L 104 93 64 31
CEAL/H 33 30 22 14 CEAL/H 33 31 22 14
CEAH/H 24 23 11 8 CEAH/H 24 23 11 8

a All patients, disease-free survival b All patients, overall survival

100 100

80 80
Disease-free survival (%)

Overall survival (%)

60 60
D2 D2
D2 + CME D2 + CME
40 40

20 20

0 6 12 18 0 6 12 18
Time after surgery (months) Time after surgery (months)
No. at risk No. at risk
D2 72 64 46 24 D2 72 66 46 25
D2 + CME 65 58 40 20 D2 + CME 65 58 40 21

c Patients with low preoperative CEA, disease-free d Patients with low preoperative CEA, overall survival
survival

Fig. 2Kaplan–Meier curves for overall survival and disease-free survival according carcinoembryonic antigen (CEA) expression in
patients with gastric cancer. a Disease-free and b overall survival in all 161 patients according to CEA expression before and after
surgery. c Disease-free and d overall survival following conventional D2 lymphadenectomy (D2) and D2 lymphadenectomy with
complete mesogastric excision (D2 + CME) in 137 patients with low CEA expression before surgery. CEAL/L , CEA expression low
before and after surgery; CEAL/H , CEA expression low before and high after surgery; CEAH/H , CEA expression high before and after
surgery. a P = 0⋅033 (CEAL/L versus CEAL/H ), P = 0⋅591 (CEAL/L versus CEAH/H ), P = 0⋅532 (CEAL/H versus CEAH/H ); b P = 0⋅727
(CEAL/L versus CEAL/H ), P = 0⋅903 (CEAL/L versus CEAH/H ), P = 1⋅000 (CEAL/H versus CEAH/H ) (two-stage test). c P = 0⋅033,
d P = 0⋅134 (log rank test)

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
Carcinoembryonic antigen in peritoneal fluid of patients with gastric cancer 1477

of tumour invasion (P = 0⋅001), pTNM stage (P = 0⋅002) (CEAL/H ) was worse than that in patients with low postop-
and surgical procedure (P = 0⋅024). The percentage of erative expression (CEAL/L ), indicating that the presence
CEAL/H increased with the increase in depth of tumour of free intraperitoneal cancer cells is clinically relevant and
invasion and pTNM stage (Table 3). may indicate a higher risk of tumour recurrence.
Univariable analysis revealed pT status, pTNM stage and Some studies10,11 have also demonstrated an association
surgical procedure as statistically significant factors, and between surgery-induced free intraperitoneal cancer cells
these variables were entered in the multivariable model. and tumour relapse. However, the origin of these tumour
Multivariable analysis revealed that depth of tumour inva- cells is still unclear. The traditional point of view is that
sion (P = 0⋅004) and surgical procedure (P = 0⋅027) were cancer cells may be carried from the lymphatic vessels
factors independently associated with high postoperative or shed from the serosal surface of the tumour into the
CEA expression (CEAL/H ) in patients who demonstrated abdominal cavity8,9,11 . However, this cannot explain why
free intraperitoneal cancer cells after surgery (Table 4). free intraperitoneal cancer cells are detected in tumours
that do not invade the serosal surface of the stomach,
or in tumours with no lymph node involvement21 – 25 . In
Survival
addition, tumour recurrence in these patients cannot be
The median length of follow-up was 13 (range 2–26) fully explained by direct spillage from the serosal surface,
months for the D2 group and 13 (4–26) months for the blood vessels or lymphatics26,27 .
D2 + CME group. Of the 104 patients with CEAL/L , four Previous studies have reported on the anatomy of the
(3⋅8 per cent) developed tumour recurrence and six (5⋅8 per gastric mesentery17 and the presence of cancer cells in the
cent) died from a complication or recurrence. Of the 33 mesogastrium, the so-called ‘metastasis V’ route16 . Meta-
patients with CEAL/H , five (15 per cent) developed tumour stasis V refers to isolated tumour cells or tumour nod-
recurrence, and three (9 per cent) died from a compli- ules present in the mesogastrium that have no continuous
cation or recurrence. Of the 24 patients with CEAH/H , relationship with the primary tumour or metastatic lymph
two (8 per cent) developed tumour recurrence and both node16 . In the present study, six patients with a pT2 or pT3
died from cancer. The DFS of patients with CEAL/H tumour without lymph node metastasis had free intraperi-
was significantly worse than that in patients with CEAL/L toneal cancer cells after surgery, and one patient with a pT3
(P = 0⋅033) (Fig. 2a), whereas OS was not significantly dif- N0 M0 tumour developed peritoneal recurrence 12 months
ferent (P = 0⋅727) (Fig. 2b). after surgery. This supports the theory that free intraperi-
Of patients with low preoperative CEA expression, eight toneal cancer cells may originate or leak from the meso-
of 72 patients (11 per cent) in the D2 group developed gastrium when the integrity of the mesentery is disrupted
tumour recurrence, and seven (10 per cent) died from during surgery.
a complication or disease recurrence. In the D2 + CME Takebayashi and colleagues11 reported that free
group, only one of 65 patients (2 per cent) with low preop- intraperitoneal cancer cells after conventional D2 gas-
erative CEA expression developed tumour recurrence, and trectomy have cellular activity, proliferative potential and
two patients (3 per cent) died from a complication. DFS fol- tumorigenic capacity, and carry a poor prognosis. In their
lowing D2 + CME was significantly better than that after research, over 60 per cent of patients with gastric cancer
conventional D2 (P = 0⋅033) (Fig. 2c), although the dif- who had no CEA or cytokeratin 20 (CK20) amplifica-
ference in OS was not statistically significant (P = 0⋅134) tion expressed in the peritoneal fluid before gastrectomy
(Fig. 2d). showed high levels of CEA or CK20 following tumour
dissection, suggesting that preventing the dissemination of
Discussion
cancer cells during surgery may be of benefit11 .
Consistent with these results11 , the present study also
In this comparison of the relative expression of the tumour detected surgery-induced free intraperitoneal cancer
cell marker CEA in peritoneal washing before and after cells in 32 per cent of the patients after conventional
laparoscopic gastrectomy for cancer, mean levels of CEA laparoscopic D2 lymphadenectomy. Furthermore, depth
mRNA were lower in patients who had the D2 + CME of tumour invasion, pTNM stage and surgical approach
procedure than in those undergoing the conventional D2 were associated with the presence of free intraperitoneal
operation. This suggests that D2 + CME may reduce the cancer cells. In contrast, D2 + CME was associated with a
presence of free intraperitoneal cancer cells following gas- lower rate of high CEA expression in the peritoneal fluid
trectomy. Of patients with low preoperative CEA expres- (15 per cent) and better DFS. A previous study18 found
sion, DFS in those with high postoperative expression that mean(s.d.) intraoperative bleeding was 12⋅44(22⋅89)

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
1478 D. Xie, Y. Wang, J. Shen, J. Hu, P. Yin and J. Gong

ml and duration of hospital stay was 11⋅09(4⋅28) days in 6 Schott A, Vogel I, Krueger U, Kalthoff H, Schreiber HW,
patients undergoing laparoscopic distal gastrectomy with Schmiegel W et al. Isolated tumor cells are frequently
CME. These data indicate a further benefit of laparoscopic detectable in the peritoneal cavity of gastric and colorectal
D2 + CME in the surgical treatment of gastric cancer. cancer patients and serve as a new prognostic marker.
Some patients with high CEA expression did not have Ann Surg 1998; 227: 372–379.
7 De Andrade JP, Mezhir JJ. The critical role of peritoneal
tumour recurrence. Although RT–qPCR has a high sensi-
cytology in the staging of gastric cancer: an evidence-based
tivity, false-positive results may occur as a result of DNA
review. J Surg Oncol 2014; 110: 291–297.
contamination, dead tumour cells, or epithelial cells from
8 Marutsuka T, Shimada S, Shiomori K, Hayashi N, Yagi Y,
the gastrointestinal tract11 . High CEA expression sug- Yamane T et al. Mechanisms of peritoneal metastasis after
gests the presence of cancer cells in the peritoneal cav- operation for non-serosa-invasive gastric carcinoma: an
ity, but these cells may not be viable or have the ability ultrarapid detection system for intraperitoneal free cancer
to metastasize11 . Median follow-up in the present study cells and a prophylactic strategy for peritoneal metastasis.
was only 13 (range 2–26) months, which may not be long Clin Cancer Res 2003; 9: 678–685.
enough for peritoneal recurrence to develop, especially 9 Hao YX, Zhong H, Yu PW, Qian F, Zhao YL, Shi Y et al.
given the small number of patients in the study. Although Influence of laparoscopic gastrectomy on the detection rate
the DFS of patients with CEAL/L and CEAH/H expression of free gastric cancer cells in the peritoneal cavity. Ann Surg
was not significantly different, there was a significant dif- Oncol 2010; 17: 65–72.
ference in DFS between the two surgical procedures. 10 Han TS, Kong SH, Lee HJ, Ahn HS, Hur K, Yu J et al.
Dissemination of free cancer cells from the gastric lumen and
from perigastric lymphovascular pedicles during radical
Acknowledgements gastric cancer surgery. Ann Surg Oncol 2011; 18: 2818–2825.
11 Takebayashi K, Murata S, Yamamoto H, Ishida M,
D.X. and Y.W. contributed equally to this work. Yamaguchi T, Kojima M et al. Surgery-induced peritoneal
The authors thank B. Cao, J. Qin, J. Wu, Q. Yan, Y. cancer cells in patients who have undergone curative
Hu, C. Yang and Z. Cao for their assistance with surgical gastrectomy for gastric cancer. Ann Surg Oncol 2014; 21:
preparation. They also thank C. Jin, L. Zhu, J. Shi, J. 1991–1997.
Wu, M. Pan, J. Yuan and F. Sheng for assistance with 12 Wang JY, Lin SR, Lu CY, Chen CC, Wu DC, Chai CY et al.
sample collection and follow-up; C. Yu, Y. Zhang and B. Gastric cancer cell detection in peritoneal lavage: RT–PCR
Cao for assistance in the preparation of this paper and for carcinoembryonic antigen transcripts versus the combined
video recording. This work was supported by grants from cytology with peritoneal carcinoembryonic antigen levels.
the National Natural Science Foundation of China (grant Cancer Lett 2005; 223: 129–135.
numbers 81572861, 81773053). 13 Pecqueux M, Fritzmann J, Adamu M, Thorlund K, Kahlert
C, Reißfelder C et al. Free intraperitoneal tumor cells and
Disclosure: The authors declare no conflict of interest.
outcome in gastric cancer patients: a systematic review and
meta-analysis. Oncotarget 2015; 6: 35 564–35 578.
References 14 Chae HD, Kim IH. Prognostic significance of CEA
1 Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, expression by RT–PCR in peritoneal wash from patients
Jemal A. Global cancer statistics, 2012. CA Cancer J Clin with gastric cancer: result of a 5-year follow-up after curative
2015; 65: 87–108. resection. Scand J Gastroenterol 2016; 51: 956–960.
2 Sasako M, Saka M, Fukagawa T, Katai H, Sano T. Surgical 15 Xie D, Osaiweran H, Liu L, Wang X, Yu C, Tong Y et al.
treatment of advanced gastric cancer: Japanese perspective. Mesogastrium: a fifth route of metastasis in gastric cancer?
Dig Surg 2007; 24: 101–107. Med Hypotheses 2013; 80: 498–500.
3 Strong VE, Song KY, Park CH, Jacks LM, Gonen M, Shah 16 Xie D, Liu L, Osaiweran H, Yu C, Sheng F, Gao C et al.
M et al. Comparison of gastric cancer survival following R0 Detection and characterization of metastatic cancer cells in
resection in the United States and Korea using an the mesogastrium of gastric cancer patients. PLoS One 2015;
internationally validated nomogram. Ann Surg 2010; 251: 10: e0142970.
640–646. 17 Xie D, Gao C, Lu A, Liu L, Yu C, Hu J et al. Proximal
4 Japanese Gastric Cancer Association. Japanese gastric cancer segmentation of the dorsal mesogastrium reveals new
treatment guidelines 2010 (ver. 3). Gastric Cancer 2011; 14: anatomical implications for laparoscopic surgery. Sci Rep
113–123. 2015; 5: 16287.
5 Sasako M, Sano T, Yamamoto S, Kurokawa Y, Nashimoto A, 18 Xie D, Yu C, Liu L, Osaiweran H, Gao C, Hu J et al.
Kurita A et al. D2 lymphadenectomy alone or with Short-term outcomes of laparoscopic D2 lymphadenectomy
para-aortic nodal dissection for gastric cancer. N Engl J Med with complete mesogastrium excision for advanced gastric
2008; 359: 453–462. cancer. Surg Endosc 2016; 30: 5138–5139.

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd
Carcinoembryonic antigen in peritoneal fluid of patients with gastric cancer 1479

19 Park YK, Yoon HM, Kim YW, Park JY, Ryu KW, Lee YJ 23 Schwarz RE, Zagala-Nevarez K. Recurrence patterns after
et al.; COACT Group. Laparoscopy-assisted versus open D2 radical gastrectomy for gastric cancer: prognostic factors and
distal gastrectomy for advanced gastric cancer: results from a implications for postoperative adjuvant therapy. Ann Surg
randomized phase II multicenter clinical trial (COACT Oncol 2002; 9: 394–400.
1001). Ann Surg 2018; 267: 638–645. 24 Lee HJ, Kim YH, Kim WH, Lee KU, Choe KJ, Kim JP et al.
20 Katsuragi K, Yashiro M, Sawada T, Osaka H, Ohira M, Clinicopathological analysis for recurrence of early gastric
Hirakawa K. Prognostic impact of PCR-based identification cancer. Jpn J Clin Oncol 2003; 33: 209–214.
of isolated tumour cells in the peritoneal lavage fluid of 25 Roviello F, Marrelli D, de Manzoni G, Morgagni P, Di Leo
gastric cancer patients who underwent a curative R0 A, Saragoni L et al. Prospective study of peritoneal
resection. Br J Cancer 2007; 97: 550–556. recurrence after curative surgery for gastric cancer. Br J Surg
21 Ichiyoshi Y, Toda T, Minamisono Y, Nagasaki S, Yakeishi Y, 2003; 90: 1113–1119.
Sugimachi K. Recurrence in early gastric cancer. Surgery 26 Noguchi Y. Blood vessel invasion in gastric carcinoma.
1990; 107: 489–495. Surgery 1990; 107: 140–148.
22 Maehara Y, Hasuda S, Koga T, Tokunaga E, Kakeji Y, 27 Maehara Y, Oshiro T, Baba H, Ohno S, Kohnoe S,
Sugimachi K. Postoperative outcome and sites of recurrence Sugimachi K. Lymphatic invasion and potential for tumor
in patients following curative resection of gastric cancer. growth and metastasis in patients with gastric cancer. Surgery
Br J Surg 2000; 87: 353–357. 1995; 117: 380–385.

Supporting information
Additional supporting information can be found online in the Supporting Information section at the end of the
article.

© 2018 BJS Society Ltd www.bjs.co.uk BJS 2018; 105: 1471–1479


Published by John Wiley & Sons Ltd

You might also like