Correlation between laparoscopic radical resection and tumor markers in peritoneal irrigation fluid

Abstract

BACKGROUND

Gastric cancer (GC) is one of the most common malignancies and types of cancer worldwide.

AIM

To compare the differences in tumor markers of GC with GC dissection, we evaluated the efficacy of recent tumor removal.

METHODS

A prospective cohort study was conducted to analyze the clinical data of patients with GC. Patients were divided into two groups based on the surgical approach: The membrane dissection (MD) group, which underwent membrane-guided laparoscopic radical gastrectomy with D2 lymph node dissection plus complete mesocolic excision, and the D2 group, which underwent traditional laparoscopic radical gastrectomy with D2 lymph node dissection. Abdominal lavage fluid was collected pre- and postoperatively from patients in both groups. The expression of carcinoembryonic antigen (CEA) and cytokeratin-19 (CK-19) message RNAs in the abdominal lavage fluid was detected using reverse transcription polymerase chain reaction. The factors influencing the increase of the tumor markers were analyzed, and the short-term efficacy of the two surgery types was compared.

RESULTS

In total, 135 eligible patients were included in this study, with 69 and 66 cases in the MD and D2 groups, respectively. Fourteen patients with benign gastric lesions were selected to detect tumor marker expression. After excluding patients positive for preoperative cancer leakage, we found that 9.52% and 26.67% of patients in the MD and D2 groups developed postoperative CEA positivity, respectively. Multivariate analysis revealed that the degree of differentiation and surgical approach were independent risk factors for postoperative CEA positivity. The surgical approach was an independent risk factor affecting postoperative CK-19 positivity and postoperative CEA and CK-19 positivity. Surgical time, intraoperative blood loss, number of lymph nodes dissected, time to first postoperative flatus, and time to first liquid intake were all significantly different between the two surgical approaches. There were no significant differences in the incision length, duration of postoperative hospital stays, or postoperative complications.

CONCLUSION

MD is a better radical surgical treatment than traditional D2 surgery and is worthy of further clinical promotion and application.

Key Words: Gastric cancer; Membrane anatomy; Abdominal irrigation fluid; Tumor markers; Cancer leakage

Core Tip: This study demonstrates that membrane dissection (MD) guided laparoscopic radical gastrectomy significantly reduces postoperative cancer cell leakage compared to traditional D2 lymph adenectomy. MD showed lower rates of postoperative carcinoembryonic antigen positivity (9.52% vs 26.67%) and cytokeratin-19 positivity, with surgical approach being an independent risk factor for marker positivity. While maintaining similar safety profiles, MD achieved better short-term outcomes including reduced blood loss and faster recovery, suggesting its superiority as a radical surgical approach for gastric cancer. These findings support MD as a promising technique for improving oncological outcomes in gastric cancer surgery.

INTRODUCTION

Gastric cancer (GC) is one of the most common malignancies and one of the main types of cancer worldwide. According to the 2020 Global Cancer Statistics[1], GC is the fifth most common malignancy (5.6%) and fourth most common cancer type (7.7%). Although chemoradiotherapy, immunotherapy, and biological agents have developed rapidly in recent years, the most important means to treat GC is comprehensive and individualized treatment centered on surgery[2,3].

According to the latest version of the Japanese Stomach Cancer Association protocol[4], radical GC resection combined with D2 lymph node dissection remains the standard surgical procedure for the treatment of advanced GC. However, between 38% and 50% of patients with GC develop recurrence after radical surgery, of which recurrence in and metastasis of the peritoneum are the predominant forms. The underlying reason for the high recurrence rate of GC is currently unknown; however, previous studies have speculated that, free cancer cells in the abdominal cavity may play an important role[5,6]. The traditional perspective is that intraperitoneal free cancer cells (IFCCs) can continuously develop to the serous surface through the primary tumor, soaking the serous layer of the gastric wall and shedding from the surface into the peritoneal cavity, before surgery; however, during surgery it may release iatrogenic IFCCs from the removed blood vessels, lymphatic vessels, and tissues surrounding the tumor, and spread in the peritoneal cavity[7]. However, IFCCs will still appear in the abdominal cavity of patients whose tumor does not invade the serosa layer and those without lymph node metastasis[8-11]. Notably, these patients still experience recurrence and metastasis after thorough radical resection of GC[12,13].

The metastatic pathway of GC is different from traditional direct invasion, lymphatic metastasis, blood metastasis, or peritoneal implant metastasis; however, isolated cancer cells or cancer nodules wrapped in the gastric mesangia are not connected with the primary focus and metastatic lymph nodes. During surgery, if the gastric mesangium’s structure is damaged, isolated cancer cells or cancer nodules may fall into the abdominal cavity, causing the cancer to “leak”. The free cancer cells eventually settle in the peritoneum, forming peritoneal metastasis. Therefore, peritoneal metastasis is an important risk factor for the prognosis of GC[14,15]. The American Joint Committee on Cancer tumor node metastasis (TNM) classified cancer cells as M1. In the 13^th Japanese edition of GC classification[16,17], the presence of free cancer cells in the abdominal cavity was classified as stage IV cancer. To avoid tumor residue or cancer cell spread, the primary tumor and adjacent gastric mesenteries must be removed completely. Over the past 40 years, the appearance of total mesorectal excision and complete mesocolic excision (CME) has reduced the possibility of tumor cell shedding during colorectal cancer surgery, reduced the chance of local recurrence, and increased 5-year survival rate[18,19].

Because of the development of mesenteric resection, to reduce the high local recurrence rate after D2 radical resection of GC, scholars can now explore the entire mesangial resection of advanced GC and identify mesangial structures with similar anatomical morphology and oncological effects. However, compared to the colorectum, the stomach undergoes rotational development in the embryo, resulting in complex fusion and folding of the gastric membrane and its adjacent organs, increasing the complexity of the membrane anatomy of the stomach. To date, there is no unified consensus on the definition of gastric mesangial excision, and there is great disagreement regarding the steps of a complete gastric mesangial excision.

In China, Poorolajal et al[20] and Poorolajal et al[21] found that the blood vessels, lymph nodes, adipose tissue, and free cancer cells in the stomach and adjacent structures were all surrounded by the envelope-like gastric mesangial structure; thus, the “envelope theory” was proposed. Subsequently, because the “gastric mesangium” has a fan shape and its characteristics are different to that of the traditional mesangial-free, the concept of a “generalized mesangium” was further proposed[22], which refers to the fascia and/or serosa envelope surrounding organs and their vessels hanging on the posterior wall of the body regardless of its shape or whether it is free. Based on the definition of a generalized mesangium, Lin et al[23] deduced and verified the existence of a proximal segment of the dorsal mesogastrium and the morphology and structure of its table model, which provided guidance for determining the scope of gastric mesangial resection. In thorough primary resection and systemic lymph node dissection, to ensure the integrity of the envelope-like gastric mesentery and prevent cancer leakage, Hao et al[24] introduced the concept of “CME” and combined with traditional laparoscopic D2 surgery method, proposed the “GC D2 + complete gastric mesial resection”, known as the “GC D2 + CME surgery”.

Since 2014, we have actively explored membrane anatomy under the guidance of laparoscopic gastric D2 + CME radical surgery, and found that the gastric mesentery and mesangial bed is loose, tired vascular fascia space, along the fascia space can safely and effectively realize the complete resection of GC lesions, tumor within the scope of lymph node cleaning and the whole piece of gastric mesentery resection. To verify whether the membrane-dissection-guided laparoscopic radical gastrectomy with D2 lymph node dissection + CME can ensure complete resection of the primary tumor, lymph nodes, and gastric mesentery, thereby significantly reducing the occurrence of intraoperative cancer cell leakage, we conducted a prospective cohort analysis of 135 patients with GC that were treated at the Putian First Hospital[25]. The patients were divided into two groups based on the surgical approach: The membrane dissection (MD) group, which underwent laparoscopic radical gastrectomy with D2 lymph node dissection + CME guided by membrane anatomy; and the D2 group, which underwent traditional laparoscopic radical gastrectomy with D2 lymph node dissection. Abdominal lavage fluid was collected pre- and postoperatively from patients with GC, and the expression of carcinoembryonic antigen (CEA) and cytokeratin-19 (CK-19) message RNA (mRNA) in the abdominal lavage fluid was detected using reverse transcription polymerase chain reaction (RT-PCR). This serves as a reliable tool for clinically detecting free cancer cells and predicting peritoneal metastasis. We further evaluated the safety, feasibility, and short-term efficacy of laparoscopic D2 + CME radical resection guided by membrane anatomy.

MATERIALS AND METHODS

Study participants

This was a clinical prospective cohort study. According to the specific enrollment criteria, from February 2022 to January 2024, 135 patients who underwent radical laparoscopic gastrointestinal surgery in Putian First Hospital, were divided into groups according to different surgical methods: MD group (laparoscopic D2 + CME radical group) and D2 group (traditional laparoscopic D2 radical group). This study was approved by the Ethics Committee of the Putian First Hospital (No. 2021-019).

Patient inclusion criteria

The following inclusion criteria were used to select patients: (1) Age: 18-85 years; (2) Body mass index (BMI) < 30 kg/m²; (3) Gastric adenocarcinoma confirmed by gastroscopy and tissue biopsy before surgery; (4) No serious underlying disease [American Society of Anesthesiologists (ASA) preoperative grade 3]; (5) Patients who underwent routine preoperative examination (e.g., chest and abdominal computed tomography, magnetic resonance imaging, and abdominal ultrasound), with no observed peritoneal and distant organ metastasis; (6) Patients who had complete clinical data; (7) Signed the informed consent form; and (8) The surgery could be removed by “R0” in accordance with the guidelines to meet the radical treatment standard.

The following exclusion criteria were used: (1) Severe mental disorders; (2) Pregnant or lactating women; (3) Had GC combined with other malignant tumors and/or a history of other malignant tumors within 5 years; (4) Received preoperative neoadjuvant therapy (e.g., chemotherapy and radiotherapy); and (5) Stomach cancer combined with bleeding, perforation, and obstruction that required emergency surgery.

Methods

Collection of specimens: For routine preoperative preparation, patients were in the leg position, pneumoperitoneum was established and pressure in the pneumoperitoneum was maintained at 12-15 mmHg. Laparoscopic exploration was performed to confirm no planting metastasis, no distant organ metastasis, and whether laparoscopic radical resection of GC was feasible. Before the surgery, 150 mL warm saline was injected into the abdominal cavity, being careful to avoid directly washing the primary focus, and the left and right sections of the liver, spleen fossa, and transverse colon were rinsed. The saline solution (at least 100 mL) was aspirated, gently stirred, marked as preoperative abdominal irrigation fluid specimen, and immediately stored in a cryogenic refrigerator at 4 °C until further inspection. After tumor resection and lymph node dissection, and before the initiation of digestive tract reconstruction, the full set of attrator devices were replaced and the wound was rinsed with 150 mL warm saline. The saline was again aspirated (at least 100 mL), gently stirred, marked as the postoperative abdominal irrigation fluid specimen, and stored in the cryogenic refrigerator at 4 °C until further inspection.

Total RNA extraction: Frozen samples were removed from a -80 °C refrigerator. The samples were thawed and the cell suspension was mixed well and incubated at room temperature for 10 minutes until completely dissolved. Subsequently, 200 μL of chloroform was added, the sample was vortexed for 1 minute, and incubated at room temperature for 10 minutes. After centrifugation at 4 °C and 12000 rpm, the sample was separated into three layers: Colorless supernatant (including RNA), white middle layer (including protein and DNA), and pink lower layer (organic phase). The supernatant was carefully removed and placed in a 1.5 mL centrifuge tube without RNA enzyme. Next, an equal volume of isopropyl alcohol (500 μL) was added, and the supernatant was mix well and incubated at room temperature for 10 minutes. Subsequently, after centrifugation at 4 °C and 12000 rpm for 10 minutes, a white RNA precipitate was observed at the bottom of the centrifuge tube. The supernatant was discarded and 1 mL of 75% alcohol was used to wash the precipitate; this solution was then centrifuged for 5 minutes at 4 °C and 12000 rpm. The supernatant was carefully removed and discarded, the centrifuge tube cover was left open, and the remaining solution was allowed to dry at room temperature for 10 minutes. Subsequently, 10-30 μL of diethyl pyrocarbonate (DEPC) water was added according to the amount of RNA precipitation to dissolve the RNA; the RNA reverse transcription reaction was conducted either immediately or the samples were stored at -80 °C until further analysis.

Evaluation of RNA concentration: RNA solution (2 μL) was added to 98 μL of DEPC water to dilute it by 50 times. Subsequently, absorbance was measured at 260 nm and 280 nm using an Ultraviolet spectrophotometer, and the values of λ = 260 and λ = 280 were recorded. The λ260/λ280 ratio was then calculated; if the ratio was maintained in the range of 1.7-2.1, the RNA purity was good. The RNA concentration was calculated by λ260 value, and the RNA concentration was calculated as λ26040 dilution of 1000 (μg/μL).

RNA reverse transcription reaction for complementary DNA synthesis

Configuring the reverse transcription reaction solution: The reverse transcription reaction solution (20 μL) was placed on ice. The reaction solution configuration components are as follows: (1) 4 μL of 5 × PrimeScript RT Master Mix; (2) 1 μg of RNA (the number of μL of the required RNA was calculated according to the extracted RNA concentration); and (3) 20 μL of enzyme-free sterile water.

Reverse transcription reaction: Twenty samples were used for reverse transcription. The samples were flashed away and the reverse transcription reaction was performed. The reaction conditions were as follows: Reverse transcription reaction at 37 °C for 15 minutes; reverse transcriptase inactivation reaction at 85 °C for 5 seconds and then stored at 4 °C (save). After the reverse transcription reaction, the complementary DNA products were stored at -20 °C or analyzed further by real-time PCR.

Control group with benign lesions

To distinguish between high or low relative expression of CEA and CK-19 mRNAs in peritoneal washes, we selected 14 patients with benign gastric lesions as negative controls. The intraperitoneal irrigation fluid was collected and processed similarly as mentioned above, and the relative expression levels of CEA mRNA and CK-19 mRNA were detected (the mean + two times the SD was set as the threshold). Fluwash above this threshold was defined as a positive cancer leak, and tagged as either CEA + or CK-19 +. The combination of two tumor markers simultaneously positive for cancer leakage was marked as CEA and CK-19 +. Patients below the defined threshold were defined as negative for cancer leaks, and tagged as either CEA - or CK-19 -; other tumor markers were labeled as CEA and CK-19.

The positive rate of postoperative cancer leakage was defined as the number of patients with negative preoperative cancer leakage in each group. The calculation formula was as follows: Positive rate of postoperative cancer leakage = (negative preoperative and positive postoperative)/(all cases in the same group who are negative for preoperative cancer).

Statistical analysis

All statistical analyses were performed using IBM SPSS (version 25.0). The data are expressed as the mean ± SD. The t-test or rank-independent test were used for two independent samples. Class data are expressed as frequencies or percentages and were compared using the χ² test. Statistical significance was set at P < 0.05.

RESULTS

Comparison of the clinical baseline data of the enrolled patients

In total, 145 patients with GC were enrolled, among them, 10 patients were excluded due to postoperative pathological criteria: Three patients had lymphoma and three had neuroendocrine carcinoma, three specimens had positive resection margins, and one squamous carcinoma. Therefore, 135 patients were enrolled in this study (102 men and 33 women, with an average age 67.54 ± 9.46 years), Among them, 69 and 66 were in the MD and D2 groups, respectively. There was no statistical difference in clinical baseline data, such as age, sex, BMI, ASA score, and tumor site, between the two groups (P > 0.05), Basic clinical baseline data and pathological information are presented in Table 1. In this study, 14 patients with benign gastric lesions were also selected as negative controls, including eight men and six women, aged 67.00 ± 12.73 years, and their composition was representative.

Table 1 Comparison of membrane dissection and D2 groups, mean ± SD/n (%).

Basic information	Total (n = 135)	Membrane anatomy group (n = 69)	D2 group (n = 66)	Statistic	P value
Age	67.54 ± 9.46	67.88 ± 9.19	67.18 ± 9.79	t = 0.43	0.668
Sex				χ2 = 0.21	0.650
Male	102 (75.56)	51 (73.91)	51 (77.27)
Female	33 (24.44)	18 (26.09)	15 (22.73)
ASA				-	0.666
I	6 (4.44)	2 (2.90)	4 (6.06)
II	79 (58.52)	42 (60.87)	37 (56.06)
III	50 (37.04)	25 (36.23)	25 (37.88)
BMI	22.48 ± 2.91	22.71 ± 3.06	22.23 ± 2.75	t = 0.96	0.339
AFP	31.29 ± 213.39	26.00 ± 179.35	36.81 ± 245.28	t = -0.29	0.770
CEA	8.62 ± 24.23	7.73 ± 21.06	9.55 ± 27.28	t = -0.44	0.664
CA125	13.89 ± 16.70	15.20 ± 20.59	12.53 ± 11.32	t = 0.93	0.354
CA19-9	284.42 ± 1703.98	196.39 ± 1389.93	376.45 ± 1986.93	t = -0.61	0.541
Tumor maximum diameter (cm)	4.88 ± 2.74	4.81 ± 2.67	4.96 ± 2.82	t = -0.31	0.760
Degree of differentiation				χ2 = 0.02	0.888
Middle/high	54 (40.00)	28 (40.58)	26 (39.39)
Low	81 (60.00)	41 (59.42)	40 (60.61)
Tumor site				χ2 = 7.17	0.067
Top	58 (42.96)	37 (53.62)	21 (31.82)
Middle	24 (17.78)	9 (13.04)	15 (22.73)
Bottom	32 (23.70)	15 (21.74)	17 (25.76)
Mixed	21 (15.56)	8 (11.59)	13 (19.70)
Lauren classification				χ2 = 0.13	0.938
Diffusivity	35 (25.93)	17 (24.64)	18 (27.27)
Mixed	65 (48.15)	34 (49.28)	31 (46.97)
Visible peristalsis	35 (25.93)	18 (26.09)	17 (25.76)
Periodization1				χ2 = 0.65	0.885
T1	23 (17.04)	13 (18.84)	10 (15.15)
T2	15 (11.11)	7 (10.14)	8 (12.12)
T3	57 (42.22)	30 (43.48)	27 (40.91)
T4	40 (29.63)	19 (27.54)	21 (31.82)
N period1				χ2 = 3.33	0.344
N0	42 (31.11)	23 (33.33)	19 (28.79)
N1	23 (17.04)	12 (17.39)	11 (16.67)
N2	22 (16.30)	14 (20.29)	8 (12.12)
N3	48 (35.56)	20 (28.99)	28 (42.42)
TNM grade1				χ2 = 1.39	0.500
I	33 (24.44)	17 (24.64)	16 (24.24)
II	30 (22.22)	18 (26.09)	12 (18.18)
III	72 (53.33)	34 (49.28)	38 (57.58)

¹Tumor node metastasis staging was performed using the American Joint Committee on Cancer tumor node metastasis staging system.

AFP: Alpha fetoprotein; ASA: American Society of Anesthesiologists; BMI: Body mass index; CA125: Carbohydrate antigen 125; CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen; TNM: Tumor node metastasis.

Results of CEA mRNA test of peritoneal wash solution

CEA mRNA expression levels before and after surgery were determined using RT-PCR and calculated for all enrolled patients. The CEA mRNA expression levels were 0.24 ± 0.43 and 0.33 ± 0.60 (P = 0.318) in the MD and D2 groups respectively, and for the postoperative peritoneal irrigation fluid it was 0.45 ± 0.67 ± 0.74 in the MD and D2 groups, respectively; however, no significant difference (P = 0.081) was observed (Table 2).

Table 2 Carcinoembryonic antigen cancer leakage ratios for the membrane dissection and D2 surgeries at different T stages, n (%).

T grade	Total (n = 123)	Membrane anatomy group (n = 63)	D2 group (n = 60)	Statistic	P value
Total T stage				χ2 = 6.15	0.013
CEA -	101 (82.11)	57 (90.48)	44 (73.33)
CEA +	22 (17.89)	6 (9.52)	16 (26.67)
T1					0.435
CEA -	22 (95.65)	13 (100.00)	9 (90.00)
CEA +	1 (4.35)	0 (0.00)	1 (10.00)
T2					1.000
CEA -	10 (76.92)	5 (83.33)	5 (71.43)
CEA +	3 (23.08)	1 (16.67)	2 (28.57)
T3				χ2 = 0.76	0.384
CEA -	44 (86.27)	24 (92.31)	20 (80.00)
CEA +	7 (13.73)	2 (7.69)	5 (20.00)
T4					0.146
CEA -	25 (69.44)	15 (83.33)	10 (55.56)
CEA +	11 (30.56)	3 (16.67)	8 (44.44)

CEA: Carcinoembryonic antigen.

In this study, 123 patients were positive for CEA, including 63 in MD group and 60 in the D2 group. CEA mRNA expression was 0.12 ± 0.18 and 0.16 ± 0.21, respectively, with no statistical difference (P = 0.239), and 0.27 ± 0.36 and 0.56 ± 0.65 in the D2 group, which was statistically significant (P = 0.003) (Table 3).

Table 3 Carcinoembryonic antigen + ratios of the membrane dissection and D2 procedures in the T3 + T4 stage, n (%).

T3 + T4	Total (n = 87)	Membrane anatomy group (n = 63)	D2 group (n = 60)	Statistic	P value
CEA				χ2 = 4.72	0.030
CEA -	69 (79.31)	39 (88.64)	30 (69.77)
CEA +	18 (20.69)	5 (11.36)	13 (30.23)

CEA: Carcinoembryonic antigen.

In the T3 + T4 stage, there was no statistical difference in preoperative CEA expression between the two groups (P = 0.145), while the postoperative D2 group had significantly higher relative CEA expression than the MD group (P = 0.007).

According to the threshold calculated for benign tumors, 9.52% (6/63) patients in the MD group were postoperative CEA +, and 26.66% (16/60) in the D2 group were postoperative CEA +, which was significantly higher than that in the MD group (P = 0.013). In the T3 + T4 stage, the rates of postoperative CEA + were 11.36% and 30 / 43% (13/43) in the D2 group, respectively; thus, the rate of postoperative CEA + in the D2 group was significantly higher than that in the MD group (P = 0.030) (Table 4).

Table 4 Relationship between carcinoembryonic antigen cancer leakage and patient clinical pathology characteristics, n (%).

Variable	Total (n = 123)	CEA - (n = 101)	CEA + (n = 22)	Statistic	P value
Degree of differentiation				χ2 = 7.67	0.006
Medium-high	49 (39.84)	46 (45.54)	3 (13.64)
Low	74 (60.16)	55 (54.46)	19 (86.36)
Violated vasculature				χ2 = 3.88	0.049
Yes	51 (41.46)	46 (45.54)	5 (22.73)
No	72 (58.54)	55 (54.46)	17 (77.27)
T grade					0.049
T1	23 (18.70)	22 (21.78)	1 (4.55)
T2	13 (10.57)	10 (9.90)	3 (13.64)
T3	51 (41.46)	44 (43.56)	7 (31.82)
T4	36 (29.27)	25 (24.75)	11 (50.00)
N grade					0.025
N0	38 (30.89)	35 (34.65)	3 (13.64)
N1	20 (16.26)	19 (18.81)	1 (4.55)
N2	19 (15.45)	14 (13.86)	5 (22.73)
N3	46 (37.40)	33 (32.67)	13 (59.09)
TNM grade				χ2 = 8.87	0.012
I	31 (25.20)	28 (27.72)	3 (13.64)
II	26 (21.14)	25 (24.75)	1 (4.55)
III	66 (53.66)	48 (47.52)	18 (81.82)
Modus operandi				χ2 = 6.15	0.013
Membrane anatomy	63 (51.22)	57 (56.44)	6 (27.27)
D2	60 (48.78)	44 (43.56)	16 (72.73)

CEA: Carcinoembryonic antigen; TNM: Tumor node metastasis.

Clinicopathological factors affecting postoperative CEA +

Univariate analysis showed that the postoperative degree of differentiation (P = 0.011), N stage (P = 0.049), TNM stage (P = 0.027), and surgical mode (P = 0.017) affected the occurrence of postoperative CEA + (Table 5). Of these factors, the degree of differentiation (P = 0.013), and surgical mode (P = 0.013) were independent risk factors for postoperative CEA + (Table 6).

Table 5 Single analysis affecting carcinoembryonic antigen cancer leakage.

Variable	OR (95%CI)	P value
Degree of differentiation		0.011
Low	1.00 (Ref.)
Medium-high	0.19 (0.05-0.68)
N grade		0.049
N0	1.00 (Ref.)
N1	0.61 (0.06-6.32)
N2	4.17 (0.88-19.82)
N3	4.60 (1.20-17.60)
TNM grade		0.027
I	1.00 (Ref.)
II	0.37 (0.04-3.82)
III	3.50 (0.95-12.95)
Modus operandi		0.017
Membrane anatomy	1.00 (Ref.)
D2	3.45 (1.25-9.55)

OR: Odds ratio; CI: Confidence interval; TNM: Tumor node metastasis.

Table 6 ultivariate analysis of carcinoembryonic antigen cancer leakage.

Variable	OR (95%CI)	P value
Degree of differentiation		0.013
Low	1.00 (Ref.)
Medium-high	0.17 (0.04-0.77)
Modus operandi		0.013
Membrane anatomy	1.00 (Ref.)
D2	4.23 (1.29-13.43)

OR: Odds ratio; CI: Confidence interval.

Results of CK-19 mRNA in peritoneal irrigation solution

Postoperative CK-19 mRNA expression levels were 0.38 ± 0.42 and 0.70 ± 1.05 in the MD and D2 groups, respectively (P = 0.028). Of the 60 patients in the MD group, six were CK-19 + (10.00%, 6/60), and of the 58 patients in the D2 group, 14 were CK-19 + (24.14%, 14/58) (P = 0.041) (Table 7).

Table 7 Cytokeratin-19 cancer leak ratios for two surgeries in different T stages, n (%).

T grade	Total (n = 118)	Membrane anatomy group (n = 63)	D2 group (n = 60)	Statistic	P value
T grade				χ2 = 4.19	0.041
CK-19 -	98 (83.05)	54 (90.00)	44 (75.86)
CK-19 +	20 (16.95)	6 (10.00)	14 (24.14)
T1
CK-19 -	21 (100.00)	12 (100.00)	9 (100.00)
CK-19 +	0 (0.00)	0 (0.00)	0 (0.00)
T2					1.000
CK-19 -	12 (92.31)	6 (100.00)	6 (85.71)
CK-19 +	1 (7.69)	0 (0.00)	1 (14.29)
T3				χ2 = 0.27	0.605
CK-19 -	40 (83.33)	22 (88.00)	18 (78.26)
CK-19 +	8 (16.67)	3 (12.00)	5 (21.74)
T4					0.156
CK-19 -	25 (69.44)	14 (82.35)	11 (57.89)
CK-19 +	11 (30.56)	3 (17.65)	8 (42.11)

CK-19: Cytokeratin-19.

Relationship between CK-19 + and clinical baseline data and pathology after surgery

Statistical analysis of postoperative cancer leakage and clinical pathological data revealed no significant correlation between the incidence of CK-19 + and age, sex, BMI, preoperative ASA score, preoperative serum tumor marker level, tumor size, tumor site, Lauren classification, and nerve invasion (Table 8); however, a significant correlation was found with differentiation (P = 0.039), T stage (P = 0.015), N stage (P = 0.021), TNM stage (P = 0.010), and surgical method (P = 0.041) (Table 9). Furthermore, the rate of CK-19 + increased with TNM stage, with one (3.45%) case of CK-19 + detected out of 29 patients with stage I GC, two (8.33%) cases in patients with stage II GC, and 17 (26.15%) cases out of 65 patients with stage III GC.

Table 8 Cytokeratin-19 + ratio in T3 + T4 stage, n (%).

T3 + T4	Total (n = 84)	Membrane anatomy group (n = 63)	D2 group (n = 60)	Statistic	P value
CK-19				χ2 = 3.33	0.068
CK-19 -	65 (77.38)	36 (85.71)	29 (69.05)
CK-19 +	19 (22.62)	6 (14.29)	13 (30.95)

CK-19: Cytokeratin-19.

Table 9 Relationship between cytokeratin-19 cancer leakage and patients, n (%).

	Total (n = 118)	CK-19 - (n = 98)	CK-19 + (n = 20)	Statistic	P value
Degree of differentiation				χ2 = 4.27	0.039
Medium-high	48 (40.68)	44 (44.90)	4 (20.00)
Low	70 (59.32)	54 (55.10)	16 (80.00)
T grade					0.015
T1	21 (17.80)	21 (21.43)	0 (0.00)
T2	13 (11.02)	12 (12.24)	1 (5.00)
T3	48 (40.68)	40 (40.82)	8 (40.00)
T4	36 (30.51)	25 (25.51)	11 (55.00)
N grade					0.021
N0	37 (31.36)	35 (35.71)	2 (10.00)
N1	17 (14.41)	16 (16.33)	1 (5.00)
N2	20 (16.95)	15 (15.31)	5 (25.00)
N3	44 (37.29)	32 (32.65)	12 (60.00)
TNM grade					0.010
I	29 (24.58)	28 (28.57)	1 (5.00)
II	24 (20.34)	22 (22.45)	2 (10.00)
III	65 (55.08)	48 (48.98)	17 (85.00)
Modus operandi				χ2 = 4.19	0.041
Membrane anatomy	60 (50.85)	54 (55.10)	6 (30.00)
D2	58 (49.15)	44 (44.90)	14 (70.00)

CK-19: Cytokeratin-19; TNM: Tumor node metastasis.

Clinicopathological factors affecting postoperative CK-19 +

Univariate analysis revealed that the degree of differentiation (P = 0.047), TNM stage (P = 0.030), and surgical method (P = 0.047) significantly affected the occurrence of postoperative CK-19 + (Table 10). Based on these factors, we concluded that surgical method (P = 0.048) was an independent risk factor for postoperative CK-19 + (Table 11).

Table 10 Single analysis of the cytokeratin-19 cancer leakage.

Variable	OR (95%CI)	P value
Degree of differentiation		0.047
Low	1.00 (Ref.)
Medium-high	0.31 (0.10-0.98)
TNM grade		0.030
I	1.00 (Ref.)
II	2.55 (0.22-29.93)
III	9.92 (1.25-78.58)
Modus operandi		0.047
Membrane anatomy	1.00 (Ref.)
D2	2.86 (1.02-8.07)

OR: Odds ratio; CI: Confidence interval; TNM: Tumor node metastasis.

Table 11 analysis of the cytokeratin-19 cancer leakage.

Variable	OR (95%CI)	P value
Modus operandi		0.048
Membrane anatomy	1.00 (Ref.)
D2	3.15 (0.97-10.23)

OR: Odds ratio; CI: Confidence interval.

Correlation analysis of CEA with CK-19

In this study, the two tumor markers were analyzed together to reduce the possibility of false-positive results. Similarly, excluding patients with positive preoperative cancer leakage, two patients in the MD group were positive for postoperative tumor markers, namely, CEA and CK-19 (3.45%, 2/58), while 10 patients in the D2 group were positive for CEA and CK-19 (17.54%, 10/57), with a significant difference (P = 0.013) (Table 12). In the T3 + T4 stage, the rates of postoperative CEA and CK-19 positivity were 5.00% (2/40) and 21.95% (9/41) in the MD and D2 groups, respectively, and the ratio of postoperative CEA and CK-19 positivity in D2 was significantly higher than that in the MD group (P = 0.026) (Table 13).

Table 12 Concomitant cancer leak ratio for membrane dissection and D2 in different T stages, n (%).

	Total (n = 115)	Membrane anatomy group (n = 63)	D2 group (n = 60)	Statistic	P value
Total T stage				χ2 = 6.11	0.013
CEA and CK-19 -	103 (89.57)	56 (96.55)	47 (82.46)
CEA and CK-19 +	12 (10.43)	2 (3.45)	10 (17.54)
T1 grade
CEA and CK-19 -	21(100.00)	12 (100.00)	9 (100.00)
CEA and CK-19 +	0 (0.00)	0 (0.00)	0 (0.00)
T2 grade					1.000
CEA and CK-19 -	12 (92.31)	6 (100.00)	6 (85.71)
CEA and CK-19 +	1 (7.69)	0 (0.00)	1 (14.29)
T3 grade				χ2 = 0.52	0.470
CEA and CK-19 -	44 (95.65)	23 (100.00)	21 (91.30)
CEA and CK-19 +	2 (4.35)	0 (0.00)	2 (8.70)
T4 grade					0.121
CEA and CK-19 -	26 (74.29)	15 (88.24)	11 (61.11)
CEA and CK-19 +	9 (25.71)	2 (11.76)	7 (38.89)

CEA: Carcinoembryonic antigen; CK-19: Cytokeratin-19.

Table 13 Carcinoembryonic antigen and cytokeratin-19 + ratios in T3 + T4 stage, n (%).

T3 + T4	Total (n = 81)	Membrane anatomy group (n = 63)	D2 group (n = 60)	Statistic	P value
Simultaneous cancer leaks				χ2 = 4.96	0.026
CEA and CK-19 -	70 (86.42)	38 (95.00)	32 (78.05)
CEA and CK-19 +	11 (13.58)	2 (5.00)	9 (21.95)

CEA: Carcinoembryonic antigen; CK-19: Cytokeratin-19.

Statistical analysis of postoperative cancer leakage and clinical pathological data of patients showed that the incidence of postoperative CEA and CK-19 positivity was significantly correlated with the number of positive lymph nodes (P = 0.004), differentiation (P = 0.035), T stage (P = 0.006), N stage (P = 0.027), TNM stage (P = 0.022), and surgical method (P = 0.013) (Table 14).

Table 14 Association between simultaneous cancer leakage of carcinoembryonic antigen and cytokeratin-19 markers and clinicopathological characteristics of patients, n (%).

Variable	Total (n = 115)	CEA and CK-19 - (n = 103)	CEA and CK-19 + (n = 12)	Statistic	P value
Number of positive lymph nodes, median (Q1, Q3)	3.00 (0.00, 9.50)	2.00 (0.00, 8.00)	9.50 (6.25, 21.25)	Z = 2.948	0.004
Degree of differentiation				χ2 = 4.462	0.035
Low	68 (59.13)	57 (55.34)	11 (91.67)
Medium-high	47 (40.87)	46 (44.66)	1 (8.33)
T grade					0.006
T1	21 (18.26)	21 (20.39)	0 (0.00)
T2	13 (11.3)	12 (11.65)	1 (8.33)
T3	46 (40)	44 (42.72)	2 (16.67)
T4	35 (30.43)	26 (25.24)	9 (75.00)
N grade					0.027
N0	37 (32.17)	36 (34.95)	1 (8.33)
N1	16 (13.91)	16 (15.53)	0 (0.00)
N2	19 (16.52)	17 (16.50)	2 (16.67)
N3	43 (37.39)	34 (33.01)	9 (75.00)
TNM grade					0.022
I	29 (25.22)	28 (27.18)	1 (8.33)
II	23 (20)	23 (22.33)	0 (0.00)
III	63 (54.78)	52 (50.49)	11 (91.67)
Modus operandi				χ2 = 6.112	0.013
Membrane anatomy	58 (50.43)	56 (54.37)	2 (16.67)
D2	57 (49.57)	47 (45.63)	10 (83.33)

TNM: Tumor node metastasis; CEA: Carcinoembryonic antigen; CK-19: Cytokeratin-19.

Univariate analysis revealed that the number of positive lymph nodes (P = 0.022), degree of differentiation (P = 0.040), and surgical method (P = 0.026) were all factors affecting the occurrence of postoperative CEA and CK-19 positivity. Multivariate analysis revealed that the surgical mode (P = 0.042) was an independent risk factor for postoperative CEA and CK-19 positivity (Table 15).

Table 15 Single analysis of the simultaneous cancer leakage of the carcinoembryonic antigen and cytokeratin-19 markers.

Variable	OR (95%CI)	P value
Number of positive lymph nodes	1.05 (1.01-1.10)	0.022
Degree of differentiation		0.040
Low	1.00 (Ref.)
Medium-high	0.11 (0.01-0.90)
Modus operandi		0.026
Membrane anatomy	1.00 (Ref.)
D2	5.96 (1.24-28.55)

OR: Odds ratio; CI: Confidence interval.

Comparison of the recent efficacy of the two groups

Analysis of the recent efficacy in the two groups revealed that surgery time, intraoperative blood loss, number of lymph node dissections, first postoperative flatus, and first fluid feeding time were statistically significant (P < 0.05) (Table 16). Incision length, duration of postoperative hospital stays, and postoperative complications were not statistically significant (P > 0.05) (Table 17).

Table 16 Multivariate analysis of the simultaneous cancer leakage of carcinoembryonic antigen and cytokeratin-19 markers.

Variable	OR (95%CI)	P value
Modus operandi		0.042
Membrane anatomy	1.00 (Ref.)
D2	10.33 (1.63-65.65)

OR: Odds ratio; CI: Confidence interval.

Table 17 Comparison of recent surgical outcomes in the membrane dissection and D2 groups, mean ± SD/n (%).

Variable	Total (n = 135)	Membrane anatomy group (n = 63)	D2 group (n = 60)	Statistic	P value
Surgery time (minute)	307.33 ± 46.05	317.07 ± 53.43	297.15 ± 34.34	t = 2.59	0.011
Intraoperative bleeding (mL)	82.48 ± 54.73	73.04 ± 53.66	92.35 ± 54.49	t = -2.07	0.040
Number of lymph node dissections	48.22 ± 15.65	50.84 ± 16.82	45.48 ± 13.93	t = 2.01	0.046
Cut length (cm), median (Q1, Q3)	7.00 (5.00, 8.00)	7.00 (5.00, 8.00)	7.00 (5.00, 8.00)	Z = -0.03	0.972
Time to the first exhaust after surgery (day), median (Q1, Q3)	3.00 (2.00, 4.00)	3.00 (2.00, 4.00)	3.00 (3.00, 4.00)	Z = -2.61	0.009
Time of the first fluid intake (day), median (Q1, Q3)	7.00 (7.00, 8.00)	7.00 (6.00, 7.00)	8.00 (7.00, 8.00)	Z = -6.56	< 0.001
Number of days in hospital after surgery (day), median (Q1, Q3)	11.00 (10.00, 13.00)	11.00 (10.00, 14.00)	11.00 (10.25, 13.00)	Z = -0.17	0.867
Pulmonary infection				χ2 = 0.41	0.523
No	116 (85.93)	58 (84.06)	58 (87.88)
Yes	19 (14.07)	11 (15.94)	8 (12.12)
Anastomotic leakage					1.000
No	133 (98.52)	68 (98.55)	65 (98.48)
Yes	2 (1.48)	1 (1.45)	1 (1.52)
Celiac bleeding					0.489
No	134 (99.26)	69 (100.00)	65 (98.48)
Yes	1 (0.74)	0 (0.00)	1 (1.52)
Ileac passion					0.237
No	133 (98.52)	69 (100.00)	64 (96.97)
Yes	2 (1.48)	0 (0.00)	2 (3.03)
Gastroplegia					0.489
No	134 (99.26)	69 (100.00)	65 (98.48)
Yes	1 (0.74)	0 (0.00)	1 (1.52)
Inconal infection/dehiscence					0.489
No	134 (99.26)	69 (100.00)	65 (98.48)
Yes	1 (0.74)	0 (0.00)	1 (1.52)
Abdominal infection				χ2 = 0.00	1.000
No	132 (97.78)	67 (97.10)	65 (98.48)
Yes	3 (2.22)	2 (2.90)	1 (1.52)
Deep venous thrombosis in the lower limbs				χ2 = 0.13	0.715
No	120 (88.89)	62 (89.86)	58 (87.88)
Yes	15 (11.11)	7 (10.14)	8 (12.12)
Pulmonary embolism				χ2 = 0.13	0.717
No	129 (95.56)	65 (94.20)	64 (96.97)
Yes	6 (4.44)	4 (5.80)	2 (3.03)
postoperative complications				χ2 = 0.55	0.458
No	100 (74.07)	53 (76.81)	47 (71.21)
Yes	35 (25.93)	16 (23.19)	19 (28.79)

DISCUSSION

After the statistical analysis, the results of this study showed that the mean relative expression levels of CEA and CK-19 mRNAs (0.27 ± 0.36 and 0.38 ± 0.42) were significantly lower than those of D2 surgery (0.56 ± 0.65 and 0.70 ± 1.05) (P < 0.05). In the detection of two tumor markers, CEA and CK-19, after excluding the patients who had developed preoperative cancer leaks, the rate of postoperative cancer leakage in traditional D2 surgery was 26.67% and 24.14%, respectively, however, the rate of postoperative cancer leakage in the MD group was only 9.52% and 10.00%, respectively. These results suggest that MD surgery can reduce the number of tumor cells shed during GC radical resection, consistent with the findings of Xie et al’s study[26]. Through multifactorial analysis, the independent factors for postoperative CEA + were the degree of tumor differentiation and surgical mode and for postoperative CK-19 +, the independent factor was surgery, which showed that patients with more aggressive GC, i.e., increased differentiation and higher malignancy degree, were more likely to have abdominal free cancer cells, and that membrane anatomical surgery is key to reducing free cancer leakage.

In terms of surgery, compared with the D2 group, the number of dissected lymph nodes was significantly reduced in the MD group (P < 0.05), consistent with the results of Seybolt et al[27]. Because of the low blood volume during external gastrointestinal reconstruction it is difficult to estimate the blood volume; therefore, the blood volume during the laparoscopic surgery was calculated using the following formula: (Blood volume in the suctionator - flushing fluid volume) + (blood gauze weight - dry gauze weight), which was corrected in combination with surgical records and videos. Traditional D2 surgery for lymph node dissection based on vascular anatomy has increased risk of vascular damage and damage to the gastric mesangium and may result in residual lymph node tissue[28]. In contrast, because blood vessels, lymph nodes, and adipose tissue are wrapped in the gastric mesentery, in the MD surgery the entire gastric mesangium can be removed, reducing blood loss and increasing lymph nodes harvest during laparoscopy[29]. This study demonstrated that the operative separation of surgery in the MD group fits the anatomy of the stomach without deliberate searching for blood vessels and lymph nodes, and that this surgery type is feasible and reproducible

The average surgery time of the MD group was significantly greater than that of the D2 group (MD group: 317.07 ± 53.43 minutes; D2 group: 297.15 ± 34.34 minutes) (P < 0.05), which was consistent with the results of previous studies[30,31]. Traditional D2 surgery is mainly performed based on the experience of the surgeon. Although the surgery time is short, intraoperative damage is significant and anatomical surgery for GC requires careful surgery to keep the surgical field sufficiently clear to enter the correct level[32,33]; thus, extending the surgery time[34-37].

In terms of postoperative efficacy, the first postoperative flatus and first oral fluid intake time in the MD group were significantly earlier than those in the D2 group (P < 0.05), while no significant difference was observed in recent postoperative complications (P > 0.05), which is consistent with the results of Zhao et al[38]. The first postoperative flatus time and oral fluid intake are the best parameters for evaluating gastrointestinal functional recovery after abdominal surgery. During the surgery in the MD group, there was less tissue damage, blood loss, and exudation; the corresponding inflammatory response was small and the surgical burden was reduced, which is conducive to the rapid recovery of gastrointestinal function after surgery, which may be one of the reasons for the earlier postoperative flatus in the MD group. Simultaneously, we found that the incidence of recent postoperative complications was similar in the two groups, and pulmonary infection and lower limb venous thrombosis were the main complications. Postoperative complications were mild and controllable, and there were no perioperative deaths. The results of three recent large-scale randomized controlled trials indicated that the postoperative complication incidence and mortality of laparoscopic D2 radical gastrectomy were 15% and 0% in Japan[39]; 15.2% and 2% in China[40]; and 16.6% and 0.4% in Korea, respectively[41]. Although the incidence of postoperative complications in the MD group in this study was 23.19% (16/69), which is higher than that reported above, if the incidence of pulmonary infection and lower limb venous thrombosis is controlled, the incidence of postoperative complications can be significantly reduced, reflecting the safety and reliability of MD surgery.

Despite less postoperative injury and rapid postoperative gastrointestinal function recovery in the MD group, there was no difference in the duration of the postoperative hospital stay between the two groups (P > 0.05). Considering that the duration of the postoperative hospital stay of patients with GC may be affected by many factors, such as incision healing and the prevention and treatment of postoperative complications, while the incision length and incidence of postoperative complications were similar between the two groups of patients, there was no statistical difference in the duration of the postoperative hospital stay.

Due to limited time, this study could not further evaluate long-term efficacy in patients with GC. According to a previous meta-analysis[42], a positive cancer leak is significantly associated with the risk of postoperative recurrence. For patients with a positive cancer leak, the mean recurrence rates were 55.35% and 68.73% at 24 and 60 months after surgery, respectively. However, in patients without cancer leakage, the average recurrence rates were 16.77% and 31.36% at 24 and 60 months after surgery, respectively. For patients with a positive cancer leak, the mean mortality rates were 43.5%, 75%, and 72.3% at 12, 24, and 60 months after surgery, respectively. For patients without cancer leakage, the mean mortality rate was 16.6%, 43.2%, and 41.2% at 12, 24, and 60 months after surgery, respectively.

Based on the conclusion of the above study, through in-depth exploration of embryology, membrane anatomy, and D2 + CME surgery for GC, it can be concluded that laparoscopic D2 + CME radical resection of GC guided by membrane anatomy is significantly different from the traditional laparoscopic D2 radical resection of GC. In traditional laparoscopic D2 radical surgery for GC mainly surrounding the organs or tissues and their blood elements, the focus is on the resection of primary lesions, vascular root ligation, and systematic lymph node dissection. Although this surgery also conforms to the principle of GC radical resection, it does not set the boundary of the entire resection area, and is dependent on the different habits, experience, and level of the surgeon, the chosen surgery path, and surgery plane. Although the surgery time is reduced, vascular damage occurs, the gastric membrane integrity is reduced, and gastric mesial and lymphatic tissue residues are produced, thereby increasing the chance of tumor recurrence. However, the laparoscopic membrane anatomy-guided radical D2 + CME of GC focuses on the complete resection of the gastric mesentery based on the radical D2 surgery for GC. The fascial space between the six messages of the stomach is only the surgical plane, which sets a precise boundary for the complete resection of the entire mesangium; therefore, the surgeon is required to accurately identify the gastric membrane and its boundaries to operate at a specific surgical level. Despite the increased surgery time, this surgery ensures the integrity of the gastric mesangial and smoothness of the mesangial bed, which significantly reduces the incidence of GC cell leakage and intraoperative bleeding, the lymph node and adipose tissue dissection is more thorough, and the postoperative gastrointestinal function recovery is faster; thus, this surgery is superior, and improves the long-term survival of patients.

Although this study suggests that the degree of tumor differentiation is an independent influencing factor for positive postoperative cancer leakage, the choice of surgery type does not depend on the degree of differentiation. Choosing laparoscopic D2 + CME radical resection for GC under the guidance of membrane anatomy can reduce intraoperative trauma, improve lymph node dissection, reduce the gastrointestinal function recovery time, reduce the incidence of intraoperative residual tumors and dissemination, and reduce the chance of surgical complications and hospitalization days, effectively improving patient prognosis and survival rate.

CONCLUSION

Compared with the traditional laparoscopic GC D2 radical surgery, laparoscopic GC D2 + CME radical surgery has a lower incidence of postoperative cancer leakage, less intraoperative bleeding, improved lymph node dissection, earlier first postoperative flatus, and a recent curative effect; thereby improving the long-term survival rate of patients. This surgery is worthy of further clinical promotion and application. Furthermore, simple and quick detection methods have been used to explore the risk factors for cancer leakage in laparoscopic GC surgery, which is of great significance for accurate staging, guiding the treatment plan, and preventing postoperative recurrence and metastasis.