Laparoscopy-assisted gastrectomy remains safe and feasible for gastric cancer patients with a history of upper abdominal surgery

Abstract

BACKGROUND

Laparoscopy-assisted radical gastrectomy is a well-established treatment for gastric cancer. However, its safety and feasibility in patients with a history of upper abdominal surgery (UAS) remain controversial.

AIM

To evaluate perioperative outcomes and oncologic radicality in gastric cancer patients with prior UAS undergoing laparoscopic gastrectomy, using a novel adhesion-risk stratification system.

METHODS

This retrospective study included gastric cancer patients with a history of UAS who underwent radical gastrectomy. The primary endpoints were a composite indicator of “difficult surgery” (defined as operative time > 210 minutes or conversion to open surgery) and the achievement of a “textbook outcome” (TO), a composite all-or-none measure of optimal surgical quality. Propensity score matching and inverse probability of treatment weighting were used to adjust for potential confounders.

RESULTS

A total of 179 patients were included (laparoscopic: 83; open: 96). Baseline characteristics were well balanced. Operative time was longer in the laparoscopic group (P < 0.001), with a conversion rate of 9.6%. Both groups achieved 100% R0 resection with comparable lymph node yields (median, 35.0). The laparoscopic group demonstrated significant recovery advantages, including earlier time to first flatus, earlier resumption of oral intake, shorter postoperative hospital stay, and lower pain scores (all P < 0.001). Overall complication rates were similar between groups. Notably, the laparoscopic group achieved a significantly higher TO rate (78.3% vs 52.1%; P < 0.001). A history of high adhesion-risk surgery was not an independent predictor of “difficult surgery” in the laparoscopic cohort (P = 0.261).

CONCLUSION

In gastric cancer patients with prior UAS, laparoscopy-assisted gastrectomy is safe and feasible, offering faster postoperative recovery and higher TO rates without compromising oncologic radicality.

Key Words: Gastric cancer; Laparoscopic surgery; Upper abdominal surgery history; Safety; Adhesion-risk stratification

Core Tip: This study used propensity score matching and inverse probability of treatment weighting to evaluate the safety and feasibility of laparoscopy-assisted gastrectomy in gastric cancer patients with a history of upper abdominal surgery. The results showed that although laparoscopy-assisted gastrectomy required a longer operative time, it was associated with faster postoperative recovery and a shorter hospital stay. Moreover, even among patients at high risk of intra-abdominal adhesions, the laparoscopic approach was not an independent predictor of surgical difficulty. Therefore, laparoscopic surgery should not be considered a contraindication in gastric cancer patients with prior upper abdominal surgery.

INTRODUCTION

Gastric cancer remains a leading cause of cancer-related mortality worldwide, and radical gastrectomy is the cornerstone of curative treatment[1]. Over the past decades, laparoscopy-assisted radical gastrectomy (LARG) has evolved into a mature and widely adopted approach, supported by robust evidence of its minimally invasive benefits[2-4]. However, its use in patients with a history of upper abdominal surgery (UAS) remains technically challenging. Intra-abdominal adhesions from prior procedures may increase operative difficulty and the risk of conversion to open surgery.

Although the benefits of laparoscopy have been demonstrated in other surgical fields for patients with previous abdominal operations[5-7], its role in gastric cancer patients with a history of UAS - an anatomically complex region - remains debated.

A key limitation of the existing literature is the frequent treatment of prior surgery as a simple binary variable (present vs absent). This approach lacks nuanced, procedure-specific risk stratification that accounts for the varying potential of different surgeries to cause severe adhesions that complicate subsequent dissection.

To address this gap, the present study aimed to comprehensively evaluate the safety, feasibility, and oncological adequacy of LARG in gastric cancer patients with a history of UAS. By introducing a novel, procedure-based adhesion-risk stratification system, we performed detailed subgroup analyses to determine whether the potential benefits of the laparoscopic approach are maintained consistently across different levels of anticipated adhesion risk.

MATERIALS AND METHODS

Study design and patient recruitment

This retrospective study was approved by the Ethics Committee of the National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences, and Peking Union Medical College, and was conducted in accordance with the ethical principles of the World Medical Association Declaration of Helsinki. We analyzed consecutive patients with gastric cancer and a history of UAS who underwent radical gastrectomy at our institution between January 2022 and April 2025.

The abdominal cavity was conceptually divided into four quadrants, with the umbilicus used as the central landmark. Previous surgeries performed cephalad to the umbilical plane were defined as upper abdominal operations. Inclusion criteria were: (1) Histologically confirmed gastric adenocarcinoma; (2) No evidence of distant metastasis (M0); and (3) A history of UAS.

Exclusion criteria: (1) Gastric stump cancer; (2) Emergency surgery for conditions such as pyloric obstruction, perforation, or hemorrhage; and (3) Poor cardiopulmonary function precluding tolerance of laparoscopic surgery. Procedures such as diagnostic or therapeutic abdominal paracentesis (e.g., for liver or pancreatic lesions) and endoscopic interventions (e.g., mucosal resection, submucosal dissection, endoscopic retrograde cholangiopancreatography) were not considered a formal history of UAS in this study. Surgeries for renal or adrenal diseases performed via a retroperitoneal approach were also excluded. Patients with gastric stump cancer - who were predominantly treated via an open approach during the study period - were excluded to avoid selection bias (Figure 1).

Figure 1  The inclusion and exclusion criteria.

Surgical procedures: All patients and their families were thoroughly counseled regarding the potential benefits and risks of both laparoscopic and open approaches, and the final decision was made jointly. All operations were performed by an experienced senior surgeon proficient in both techniques, with experience of more than 100 laparoscopic and open radical gastrectomies before the study period.

For laparoscopy-assisted gastrectomy, general anesthesia was administered with the patient in the lithotomy position. A standard five-port technique was used. Pneumoperitoneum was established and maintained at 15 mmHg. After laparoscopic mobilization and lymphadenectomy, an upper midline incision (5-8 cm) was made for specimen extraction and extracorporeal anastomosis.

For open gastrectomy, the patient was placed in the supine position, and the entire procedure was performed through an upper midline incision (approximately 20 cm).

All patients underwent D2 lymphadenectomy. The extent of lymph node dissection and the anastomotic method were standardized according to the type of gastrectomy, as follows.

Proximal gastrectomy: Lymph node stations 1, 2, 3, 4sa, 4sb, 7, 8a, 9, 10, and 11 were dissected, followed by esophagogastric anastomosis.

Distal gastrectomy: Lymph node stations 1, 3, 4sb, 4d, 5, 6, 7, 8a, 9, 11p, and 12a were dissected, followed by Billroth I or II reconstruction.

Total gastrectomy: Lymph node stations 1, 2, 3, 4sa, 4sb, 4d, 5, 6, 7, 8a, 9, 10, 11, and 12a were dissected, followed by Roux-en-Y esophagojejunostomy.

Two experienced gastrointestinal pathologists reviewed surgical specimens. Pathologic staging was performed according to the 8^th edition of the American Joint Committee on Cancer staging manual for gastric cancer. Postoperative pain was assessed daily by nursing staff using the visual analog scale (range 0-10). All clinical data were prospectively maintained in a dedicated database.

Study definitions and stratification

Adhesion-risk stratification: Based on the type of prior surgery, patients were stratified into high- or low-adhesion-risk groups. The high-risk group included procedures anticipated to cause significant adhesions that directly impede gastric mobilization or D2 lymphadenectomy: Repair of gastroduodenal perforation, gastric resection, cholecystectomy with common bile duct exploration, distal pancreatectomy, and colectomy involving the transverse colon. The low-risk group comprised procedures expected to have a limited impact on the perigastric field: Cholecystectomy alone, elective splenectomy, small bowel resection, liver resection (excluding segments adjacent to the lesser curvature), and abdominal wall surgery. This classification represents a clinically informed but assumption-based stratification designed to distinguish procedures likely to involve a substantial perigastric adhesive burden from those expected to have a limited impact. It does not account for individual variation in adhesion severity or factors such as the time elapsed since the prior surgery or a history of abdominal infection.

Difficult surgery: A composite binary endpoint was defined as an operative time exceeding 210 minutes or conversion from laparoscopic to open surgery.

Textbook outcome: An all-or-none composite measure of optimal surgical quality. Textbook outcome (TO) was considered achieved only if a patient met all of the following criteria: (1) R0 resection; (2) Retrieval of ≥ 15 lymph nodes; (3) No severe postoperative complications (Clavien-Dindo grade ≥ III); (4) No unplanned re-operation; (5) No postoperative mortality within 90 days; and (6) A postoperative hospital stay < 11 days.

The operative time threshold of > 210 minutes was selected based on institutional benchmarks. In our experience, standard laparoscopic radical gastrectomy is typically completed within 180-200 minutes. A duration exceeding 210 minutes - approximately the 75^th percentile among historical uncomplicated cases - was considered indicative of a substantially prolonged procedure, most commonly attributable to adhesiolysis. This cutoff is consistent with values used in prior studies evaluating surgical difficulty in patients with a history of abdominal surgery.

Statistical analysis

Statistical analyses were performed using SPSS software (version 24.0; IBM Corp.). Continuous variables are presented as mean ± SD or as median with interquartile range, as appropriate, and were compared using the Student t-test or the Mann-Whitney U test. Categorical variables are presented as counts and percentages and were compared using the χ² test or Fisher’s exact test.

To address potential selection bias and improve comparability between the laparoscopic and open groups, propensity score matching (PSM) and inverse probability of treatment weighting (IPTW) were applied. The propensity score was estimated using logistic regression with the following covariates: Age, sex, body mass index, American Society of Anesthesiologists physical status score, receipt of neoadjuvant chemotherapy, and clinical tumor-node-metastasis (TNM) stage.

For PSM, a 1:1 nearest-neighbor matching algorithm without replacement was applied with a caliper width of 0.2 standard deviations of the logit of the propensity score. For IPTW, stabilized weights were calculated. Balance in baseline characteristics after adjustment was assessed using standardized mean differences (SMDs), with an SMD < 0.1 indicating good balance. Primary outcomes were analyzed in three cohorts: The original (unadjusted) cohort, the PSM-matched cohort, and the IPTW-weighted cohort to assess robustness.

All analyses followed the intention-to-treat principle; patients who required conversion from laparoscopic to open surgery were analyzed in the laparoscopic group. A two-sided P value < 0.05 was considered statistically significant.

RESULTS

Patient characteristics and intergroup balance

Baseline patient demographics and clinical characteristics are summarized in Table 1. In the original, unadjusted cohort (laparoscopic: n = 83; open: n = 96), the groups differed significantly with respect to age and the proportion of patients receiving neoadjuvant chemotherapy (all P < 0.05). After PSM and IPTW, baseline covariates were well balanced between the laparoscopic and open groups. This balance was supported by the absence of statistically significant between-group differences (all P > 0.05) and by SMDs consistently below 0.1 for all variables in both adjusted cohorts.

Table 1 Baseline characteristics of patients, n (%).

Characteristics	Overall cohort (unadjust)				PSM cohort				IPTW cohort
	Laparoscopic (n = 83)	Open (n = 96)	P value	SMD	Laparoscopic (n = 74)	Open (n = 74)	P value	SMD	Laparoscopic (n = 70.35)	Open (n = 86.23)	P value	SMD
Age (year)	59.0 (52.0, 65.0)	58.5 (53.8, 64.0)	0.867	0.04	58.0 (52.0, 64.8)	58.0 (53.0, 64.0)	0.885	0.03	59.00 (52.0, 65.0)	58.00 (53.0, 64.0)	0.933	< 0.01
Sex			0.505	0.06	-		1.000	0.01	-
Female	33 (39.8)	44 (45.8)			30 (40.5)	31 (41.9)			28.26 (40.2)	36.70 (42.6)	0.756	0.05
Male	50 (60.2)	52 (54.2)			44 (59.5)	43 (58.1)			42.08 (59.8)	49.53 (57.4)
BMI (kg/m2)	23.6 (22.7, 25.4)	24.6 (23.2, 25.7)	0.424	0.15	23.6 (22.5, 25.4)	24.0 (22.4, 25.5)	0.775	0.06	23.60 (22.6, 25.4)	23.97 (22.6, 25.6)	0.718	0.07
ASA score			0.372	0.12	-		0.799	0.04	-
1	15 (18.1)	18 (18.8)			14 (18.9)	15 (20.3)			11.65 (16.6)	5.15 (17.6)	0.893	0.07
2	41 (49.4)	38 (39.6)			38 (51.4)	34 (45.9)			34.09 (48.5)	38.63 (44.8)	-	-
3	27 (32.5)	40 (41.7)			22 (29.7)	25 (33.8)			24.61 (35.0)	32.45 (37.6)	-	-
Neoadjuvant chemotherapy			1.000	0.01	-		0.838	0.03	-
No	69 (83.1)	79 (82.3)			60 (81.1)	58 (78.4)			57.90 (82.3)	70.87 (82.2)	0.984	< 0.01
Yes	14 (16.9)	17 (17.7)			14 (18.9)	16 (21.6)			12.45 (17.7)	15.36 (17.8)	-	-
cTNM stage			0.272	0.06	-		0.929	0.01	-		0.783	0.09
I	26 (31.3)	36 (37.5)			26 (35.1)	25 (33.8)			22.95 (32.6)	30.46 (35.3)
II	31 (37.3)	40 (41.7)			31 (41.9)	30 (40.5)			27.33 (38.8)	35.31 (41.0)
III	26 (31.3)	20 (20.8)			17 (23.0)	19 (25.7)			20.07 (28.5)	20.46 (23.7)

PSM: Propensity score matching; IPTW: Inverse probability of treatment weighting; SMD: Standardized mean difference; BMI: Body mass index; ASA: American Society of Anesthesiologists; cTNM: Clinical tumor-node-metastasis.

Details of UAS are presented in Table 2. In the overall cohort, the distribution of previous surgical approaches (laparoscopic vs open) and the range of specific procedures performed (e.g., cholecystectomy alone, repair of gastroduodenal perforation, gastric resection) were comparable between the laparoscopic and open groups. This balance in surgical history composition was maintained in both the PSM and IPTW cohorts, with no significant differences observed (all P > 0.05), thereby minimizing potential confounding from this factor.

Table 2 Details of previous surgical procedures, n (%).

Characteristics	Overall cohort (unadjust)			PSM cohort			IPTW cohort
	Laparoscopic (n = 83)	Open (n = 96)	P value	Laparoscopic (n = 74)	Open (n = 74)	P value	Laparoscopic (n = 70.35)	Open (n = 86.23)	P value
Previous surgical approach			1.000	-	-	1.000	-	-	0.733
Laparoscopic	43 (51.8)	49 (51.0)		37 (50.0)	37 (50.0)		33.89 (48.2)	43.84 (50.9)
Open	40 (48.2)	47 (49.0)		37 (50.0)	37 (50.0)		36.46 (51.8)	42.38 (49.2)
Previous surgical procedures			0.816	-		0.744	-		0.698
Abdominal lesion resection1	3 (3.61)	5 (5.21)	-	3 (4.05)	4 (5.41)	-	2.26 (3.21)	3.94 (4.57)
Cholecystectomy alone	23 (27.7)	20 (20.8)	-	19 (25.7)	15 (20.3)	-	20.53 (29.2)	17.73 (20.6)
Cholecystectomy and billary tract exploration	10 (12.0)	10 (10.4)	-	9 (12.2)	9 (12.2)	-	9.10 (12.9)	10.67 (12.4)
Colectomy	0 (0.00)	2 (2.08)	-	0 (0.00)	2 (2.70)	-	0.00 (0.00)	2.35 (2.73)
Distal pancreatectomy	3 (3.61)	4 (4.17)	-	1 (1.35)	4 (5.41)	-	2.03 (2.88)	3.80 (4.40)
Gastric lesion resection	9 (10.8)	12 (12.5)	-	8 (10.8)	10 (13.5)	-	7.62 (10.8)	10.85 (12.6)
Liver lesion resection	7 (8.43)	10 (10.4)	-	7 (9.46)	7 (9.46)	-	6.01 (8.55)	8.10 (9.39)
Repair of gastroduodenal perforation	14 (16.9)	10 (10.4)	-	13 (17.6)	7 (9.46)	-	11.69 (16.6)	9.07 (10.5)
Small intestine lesion resection	9 (10.8)	14 (14.6)	-	9 (12.2)	10 (13.5)	-	7.73 (11.0)	12.20 (14.2)
Splenectomy alone	5 (6.02)	9 (9.38)	-	5 (6.76)	6 (8.11)	-	3.37 (4.80)	7.51 (8.71)

¹Abdominal lesions were undiagnosed abdominal masses of unknown origin such as fibromatosis, liposarcoma, neurinoma, and tuberculosis of the abdominal cavity.

PSM: Propensity score matching; IPTW: Inverse probability of treatment weighting.

Operative and pathological outcomes

These outcomes are summarized in Table 3. Operative time was significantly longer in the laparoscopic group than in the open group (overall cohort: 229 minutes vs 214 minutes; P < 0.001), and this difference persisted in the PSM- and IPTW-adjusted analyses (all P < 0.001). The conversion rate from laparoscopic to open surgery was 9.6% in the overall laparoscopic cohort, with rates ranging from 10.8% to 11.2% in the adjusted cohorts; no conversions occurred in the open group (P < 0.05). No significant between-group differences were observed in estimated blood loss, the incidence of iatrogenic injury, or the distribution of gastrectomy types performed (all P > 0.05). Both groups achieved an R0 resection rate of 100%. The total number of harvested lymph nodes was comparable (median, 35.0 in both groups in the overall cohort; P = 0.510), and the distribution of pathological TNM stages did not differ significantly (P > 0.05).

Table 3 Surgical and pathological outcomes, n (%).

Characteristics	Overall cohort (unadjust)			PSM cohort			IPTW cohort
	Laparoscopic (n = 83)	Open (n = 96)	P value	Laparoscopic (n = 74)	Open (n = 74)	P value	Laparoscopic (n = 70.35)	Open (n = 86.23)	P value
Surgical outcome
Surgical procedure	-		0.951	-		0.815	-		0.976
Proximal gastrectomy	19 (22.9)	22 (22.9)		12 (16.2)	15 (20.3)		14.69 (20.9)	18.41 (21.4)
Distal gastrectomy	44 (53.0)	49 (51.0)		43 (58.1)	41 (55.4)		37.78 (53.7)	44.87 (52.0)
Total gastrectomy	20 (24.1)	25 (26.0)		19 (25.7)	18 (24.3)		17.87 (25.4)	22.95 (26.6)
Operation time	229 (20.4)	214 (20.1)	< 0.001a	229 (20.7)	212 (20.7)	< 0.001a	230.00 (220, 240)	210.00 (200, 225)	< 0.001a
Estimated blood loss	160 (140, 200)	180 (150, 200)	0.366	165 (150, 200)	180 (150, 200)	0.730	155.25 (140, 200)	180.00 (150, 200)	0.281
Conversion to open surgery1			0.002a	-		0.006a	-		0.001a
No	75 (90.4)	96 (100)		66 (89.2)	74 (100)		62.49 (88.8)	86.23 (100)
Yes	8 (9.64)	0 (0.00)		8 (10.8)	0 (0.00)		7.86 (11.2)	0.00 (0.00)
Iatrogenic injury2			0.803	-		0.743	-		0.502
No	77 (92.8)	91 (94.8)		68 (91.9)	70 (94.6)		64.90 (92.3)	81.76 (94.8)
Yes	6 (7.23)	5 (5.21)		6 (8.11)	4 (5.41)		5.45 (7.75)	4.47 (5.18)
Pathological outcome
R0 resection achieved	83 (100)	96 (100)	1.000	74 (100)	74 (100)	1.000	70.35 (100)	86.23 (100)	1.000
Total lymph node number	35.0 (27.0, 42.5)	35.0 (25.0, 38.5)	0.510	37.0 (28.0, 45.8)	34.5 (25.0, 37.8)	0.186	34.91 (26.69, 40.00)	35.00 (25.00, 38.10)	0.591
pTNM			0.668	-		0.762	-		0.708
I	31 (37.3)	33 (34.4)		27 (36.5)	26 (35.1)		25.39 (36.1)	28.04 (32.5)
II	25 (30.1)	35 (36.5)		22 (29.7)	26 (35.1)		21.88 (31.1)	32.14 (37.3)
III	27 (32.5)	28 (29.2)		25 (33.8)	22 (29.7)		23.09 (32.8)	26.05 (30.2)

^aP < 0.05.

¹Reasons for conversion to open surgery only included failure to place trocars and unseparated severe adhesion.

²Iatrogenic injury mainly included bowel injury, portal vein injury, and splenic bleeding.

PSM: Propensity score matching; IPTW: Inverse probability of treatment weighting; pTNM: Pathological tumor-node-metastasis.

Among the eight cases that required conversion from laparoscopic to open surgery, the most common technical challenge was dense, vascularized adhesions in the perigastric region, particularly around the hepatoduodenal ligament and lesser curvature. These adhesions obscured anatomic planes and increased the risk of visceral or vascular injury during dissection. In three cases, extensive adhesions between the abdominal wall and underlying viscera also impeded safe initial trocar placement. These findings were more frequent among patients in the high adhesion-risk subgroup, indicating that although conversion was not required in all high-risk cases, the most prohibitive adhesions were observed predominantly in this group.

Postoperative recovery and safety profile

Postoperative recovery and complication data are summarized in Table 4. The laparoscopic group demonstrated significant advantages in recovery metrics, including earlier time to first flatus, earlier resumption of oral intake, and a shorter postoperative hospital stay (all P < 0.001). Furthermore, patients in the laparoscopic group reported significantly lower pain scores on postoperative days 1 and 2 (all P < 0.05). No statistically significant differences were found between groups in the overall complication rate, specific complication profiles, re-operation rate, or perioperative mortality (all P > 0.05).

Table 4 Postoperative recovery and complications, n (%).

Characteristics	Overall cohort (unadjust)			PSM cohort			IPTW cohort
	Laparoscopic (n = 83)	Open (n = 96)	P value	Laparoscopic (n = 74)	Open (n = 74)	P value	Laparoscopic (n = 70.35)	Open (n = 86.23)	P value
Time to first flatus	2.80 (2.60, 3.20)	3.35 (2.80, 3.60)	< 0.001a	2.80 (2.60, 3.20)	3.30 (2.80, 3.50)	< 0.001a	2.80 (2.60, 3.20)	3.39 (2.84, 3.60)	< 0.001a
Time to first oral intake	2.80 (2.60, 3.30)	3.40 (2.88, 3.60)	< 0.001a	2.80 (2.60, 3.30)	3.40 (2.80, 3.60)	0.001a	2.80 (2.60, 3.30)	3.40 (2.90, 3.60)	< 0.001a
Postoperative length of stay	9.00 (8.00, 10.0)	10.0 (9.00, 12.0)	< 0.001a	9.00 (8.00, 10.0)	10.0 (9.00, 12.0)	< 0.001a	9.00 (8.0, 10.0)	10.30 (9.0, 12.0)	< 0.001a
Pain score on postoperative day 1	4.00 (4.00, 5.00)	5.00 (5.00, 6.00)	< 0.001a	4.00 (4.00, 5.00)	6.00 (5.00, 6.00)	< 0.001a	4.00 (4.00, 5.00)	5.26 (5.00, 6.00)	< 0.001a
Pain score on postoperative day 2	2.00 (2.00, 3.00)	3.00 (2.00, 4.00)	0.002a	2.00 (2.00, 3.00)	3.00 (2.00, 4.00)	0.003a	2.00 (2.00, 3.00)	3.00 (2.00, 4.00)	0.007a
Pain score on postoperative day 3	1.00 (0.00, 2.00)	1.00 (1.00, 2.00)	0.067	1.00 (0.00, 1.75)	1.00 (1.00, 2.00)	0.048a	1.00 (0.00, 1.89)	1.00 (1.00, 2.00)	0.079
Complications1			0.361	-		0.183	-		0.219
No	59 (71.1)	57 (59.4)	-	52 (70.3)	45 (60.8)	-	50.21 (71.4)	50.22 (58.2)	-
Anastomotic or duodenal stump leakage	3 (3.61)	3 (3.12)	-	3 (4.05)	1 (1.35)	-	2.47 (3.51)	2.05 (2.38)	-
Anastomotic bleeding	1 (1.20)	0 (0.00)	-	1 (1.35)	0 (0)	-	0.83 (1.18)	0.00 (0.00)	-
Intestinal obstruction	7 (8.43)	12 (12.5)	-	7 (9.46)	12 (16.2)	-	6.82 (9.70)	10.48 (12.6)	-
Lymphatic leakage	4 (4.82)	3 (3.12)	-	4 (5.41)	1 (1.35)	-	2.89 (4.11)	2.43 (2.81)	-
Gastrointestinal dysfunction	4 (4.82)	8 (8.33)	-	3 (4.05)	5 (6.76)	-	3.31 (4.70)	7.66 (8.88)	-
Incision infection	5 (6.02)	13 (13.5)	-	4 (5.41)	10 (13.5)	-	3.82 (5.42)	13.38 (15.5)	-
Re operation			1.000	-		1.000			0.740
No	79 (95.2)	91 (94.8)		70 (94.6)	71 (95.9)		65.90 (93.7)	81.84 (94.9)
Yes	4 (4.82)	5 (5.21)		4 (5.41)	3 (4.05)		4.45 (6.33)	4.38 (5.08)
Perioperative death			1.000	-		1.000	-		0.720
No	81 (97.6)	94 (97.9)		72 (97.3)	73 (98.6)		68.40 (97.2)	84.57 (98.1)
Yes	2 (2.41)	2 (2.08)		2 (2.70)	1 (1.35)		1.95 (2.77)	1.66 (1.92)
Textbook outcome	-		< 0.001a	-		0.003a	-		< 0.001a
No	18 (21.7)	46 (47.9)		17 (23.0)	35 (47.3)		15.56 (22.1)	42.72 (49.6)
Yes	65 (78.3)	50 (52.1)		57 (77.0)	39 (52.7)		54.78 (77.9)	43.50 (50.5)

^aP < 0.05.

¹Postoperative complications mainly refer to surgery-related complications, including the seven conditions listed in the table. The diagnosis of all complications was based on one or more of the followings: The patient’s symptoms, physical examinations, laboratory tests (e.g., blood routine, drainage amylase and bilirubin tests, and bacterial culture and identification), upper gastrointestinal angiography, gastroscopy, and computed tomography examination.

PSM: Propensity score matching; IPTW: Inverse probability of treatment weighting.

Subgroup analysis by adhesion risk and predictors of surgical difficulty

Subgroup analyses based on the predefined adhesion-risk stratification were performed to further investigate the impact of prior UAS (Supplementary Table 1). Perioperative and pathological outcomes were compared between the laparoscopic and open groups within both the high- and low-adhesion-risk strata (Table 5).

Table 5 Subgroup analysis of perioperative and pathological outcomes between laparoscopic and open surgery stratified by risk status.

Characteristics	High risk group			Low risk group
	Laparoscopic (n = 36)	Open (n = 38)	P value	Laparoscopic (n = 47)	Open (n = 58)	P value
Surgical procedure			0.972	-		0.818
Proximal gastrectomy	9 (25.0)	9 (23.7)		10 (21.3)	13 (22.4)
Distal gastrectomy	17 (47.2)	19 (50.0)		27 (57.4)	30 (51.7)
Total gastrectomy	10 (27.8)	10 (26.3)		10 (21.3)	15 (25.9)
Operation time	229 (18.7)	212 (18.9)	< 0.001a	230 (21.8)	216 (20.9)	0.001a
Estimated blood loss	180 (150, 200)	170 (150, 198)	0.127	150 (125, 200)	200 (150, 200)	0.037a
Iatrogenic injury			0.610	-		1.000
No	34 (94.4)	37 (97.4)		43 (91.5)	54 (93.1)
Yes	2 (5.56)	1 (2.63)		4 (8.51)	4 (6.90)
Time to first flatus	2.80 (2.60, 3.20)	3.30 (2.82, 3.50)	0.008a	2.80 (2.60, 3.20)	3.40 (2.82, 3.60)	< 0.001a
Time to first oral intake	2.80 (2.60, 3.23)	3.30 (2.82, 3.58)	0.014a	2.80 (2.60, 3.40)	3.50 (2.90, 3.70)	0.004a
Postoperative length of stay	9.00 (8.00, 10.0)	10.0 (9.00, 11.8)	0.102	9.00 (8.00, 10.0)	11.0 (9.00, 13.0)	< 0.001a
Pain score on postoperative day 1	4.00 (4.00, 5.00)	5.00 (5.00, 6.00)	< 0.001a	4.00 (4.00, 5.00)	6.00 (5.00, 6.00)	< 0.001a
Pain score on postoperative day 2	2.00 (2.00, 3.00)	2.50 (2.00, 3.75)	0.062	2.00 (2.00, 3.00)	3.00 (2.00, 4.00)	0.016a
Pain score on postoperative day 3	1.00 (0.00, 1.00)	1.00 (1.00, 2.00)	0.065	1.00 (0.00, 2.00)	1.00 (1.00, 2.00)	0.419
Complications			0.527	-		0.614
No	24 (66.7)	20 (52.6)	-	35 (74.5)	37 (63.8)	-
Anastomotic or duodenal stump leakage	1 (2.78)	2 (5.26)	-	2 (4.26)	1 (1.72)	-
Anastomotic bleeding	1 (2.78)	0 (0.00)	-	0 (0.00)	0 (0.00)	-
Intestinal obstruction	4 (11.1)	4 (10.5)	-	3 (6.38)	8 (13.8)	-
Gastrointestinal dysfunction	1 (2.78)	4 (10.5)	-	3 (6.38)	4 (6.90)	-
Lymphatic leakage	2 (5.56)	1 (2.63)	-	2 (4.26)	2 (3.45)	-
Incision infection	3 (8.33)	7 (18.4)	-	2 (4.26)	6 (10.3)	-
Re operation			0.351	-		0.377
No	33 (91.7)	37 (97.4)		46 (97.9)	54 (93.1)
Yes	3 (8.33)	1 (2.63)		1 (2.13)	4 (6.90)
Perioperative death			1.000	-		0.586
No	36 (100)	37 (97.4)		45 (95.7)	57 (98.3)
Yes	0 (0.00)	1 (2.63)		2 (4.26)	1 (1.72)
R0 resection achieved	36 (100)	38 (100)		47 (100)	58 (100)
Total lymph node number	34.5 (19.5, 40.0)	34.0 (25.0, 38.0)	0.965	35.0 (28.0, 45.5)	35.0 (25.0, 40.0)	0.324
pTNM			0.609	-		0.804
I	12 (33.3)	13 (34.2)		19 (40.4)	20 (34.5)
II	11 (30.6)	15 (39.5)		14 (29.8)	20 (34.5)
III	13 (36.1)	10 (26.3)		14 (29.8)	18 (31.0)

^aP < 0.05.

PSM: Propensity score matching; IPTW: Inverse probability of treatment weighting; pTNM: Pathological tumor-node-metastasis.

The recovery benefits associated with the laparoscopic approach were consistently observed across both risk strata. In both the high- and low-risk subgroups, the laparoscopic group demonstrated significantly shorter times to first flatus and first oral intake, as well as lower pain scores on postoperative days 1 and 2, compared with the open group (all P < 0.05). Postoperative length of stay was also significantly shorter in the laparoscopic group within the low-risk subgroup (P < 0.001), with a similar trend in the high-risk subgroup (P = 0.102). Oncologic adequacy, reflected by a 100% R0 resection rate and comparable lymph node yield, was maintained in the laparoscopic group regardless of adhesion-risk status. Complication rates, re-operation, and mortality did not differ significantly between the laparoscopic and open groups in either subgroup.

Univariate logistic regression analysis within the laparoscopic cohort was performed to identify predictors of “difficult surgery” (Supplementary Table 2). Notably, classification in the high adhesion-risk group was not associated with an increased likelihood of surgical difficulty (odds ratio = 0.6, 95% confidence interval: 0.24-1.46, P = 0.261). Other factors, including body mass index, type of gastrectomy, and clinical TNM stage, also showed no significant associations (all P > 0.05). These findings indicate that a history of high-risk UAS, as defined in this study, does not constitute an absolute contraindication to laparoscopic radical gastrectomy and is not a reliable predictor of prohibitive surgical difficulty.

DISCUSSION

With ongoing advances in surgical techniques and instrumentation, LARG has become a standard and often preferred approach for gastric cancer, supported by robust evidence from high-quality randomized controlled trials[8-12]. However, its application in patients with a history of UAS remains a major clinical challenge. Such patients were frequently excluded from pivotal randomized controlled trials because of concerns about severe intra-abdominal adhesions, which are thought to increase operative difficulty, the risk of iatrogenic injury, and conversion rates[8,11,12]. Consequently, prior abdominal surgery has often been considered a relative, or even absolute, contraindication to laparoscopy. By focusing exclusively on this challenging cohort and using rigorous propensity-score adjustment alongside a novel procedure-based adhesion-risk stratification, our study demonstrates that LARG is safe and feasible and consistently preserves the core recovery advantages of minimally invasive surgery, regardless of the predefined risk associated with the prior operation.

The primary concern regarding laparoscopy in patients with prior surgery is that adhesions may compromise patient safety and oncologic radicality. Our findings directly address this issue. Although operative time was significantly longer in the laparoscopic group (229 minutes vs 214 minutes, P < 0.0) - a result consistent with studies of laparoscopic colorectal surgery[6] and likely attributable to meticulous adhesiolysis - this did not translate into inferior outcomes. Specifically, we observed no significant differences in estimated blood loss, rates of iatrogenic injury (7.2% vs 5.2%, P = 0.803), or the spectrum and severity of postoperative complications. Most importantly, the laparoscopic approach achieved equivalent oncologic efficacy, with a 100% R0 resection rate and a comparable lymph node yield (median 35.0 vs 35.0, P = 0.510). These results corroborate and extend findings from prior smaller-scale studies[13-15], confirming that in experienced hands, the technical challenges posed by adhesions can be safely managed laparoscopically without compromising oncological principles.

Beyond safety and feasibility, a pivotal question is whether the benefits of minimally invasive surgery persist in this setting. Our data provide a clear affirmative answer. Patients undergoing LARG experienced markedly improved postoperative recovery, evidenced by earlier return of gastrointestinal function, a significantly shorter hospital stay, and substantially lower early postoperative pain scores (all P < 0.001). These advantages align with the core tenets of Enhanced Recovery After Surgery protocols. Notably, our subgroup analysis showed that these recovery benefits were consistently preserved across both high- and low-adhesion-risk strata. Even patients with a history of complex procedures derived significant benefits in terms of faster recovery and reduced pain, highlighting that the trauma-sparing effects of laparoscopy remain clinically meaningful despite the need for adhesiolysis.

To move beyond the simplistic binary of “prior surgery yes/no”, we introduced a novel stratification system based on the specific type of previous operation. The clinical utility of this stratification is twofold. First, it enables a more nuanced preoperative risk assessment. Second, and more importantly, our analysis yielded a key insight: Patients classified as high risk derived recovery benefits from laparoscopy that were comparable to those of low-risk patients, while maintaining similar safety and oncologic radicality. This finding challenges the practice of categorically excluding patients from a laparoscopic approach solely based on the nature of a prior procedure.

A particularly important finding further supports a more nuanced approach. In our analysis, classification as high adhesion risk was not associated with an increased likelihood of “difficult surgery” (odds ratio = 0.6, P = 0.261). This suggests that, although useful for descriptive stratification, the “high-risk” label is not a reliable standalone predictor of prohibitive difficulty. This may reflect several factors, including the paramount role of surgical expertise[13,14] and substantial interindividual biological variability in adhesion formation[16,17], which is poorly predicted by surgical history alone. Therefore, prior surgery should inform, rather than dictate, the surgical approach. A carefully conducted laparoscopy trial, with readiness for conversion, is a justifiable and often beneficial strategy for a broad spectrum of these patients.

This study has limitations inherent to its retrospective, single-center design. First, although we used PSM and IPTW to mitigate bias, residual confounding cannot be fully excluded, and prospective validation is warranted. Second, the adhesion-risk stratification applied in this study was based on anatomic and procedural assumptions rather than direct intraoperative assessment. We did not use a standardized adhesion scoring system during surgery, nor did we account for potentially influential variables such as the interval since the prior operation, a history of peritonitis or intra-abdominal infection, or the cumulative burden of multiple previous surgeries. Therefore, although this stratification enabled a preliminary, clinically informative subgroup analysis, future studies should incorporate objective intraoperative adhesion grading (e.g., standardized scoring systems) to develop more accurate and individualized risk prediction models. Finally, as a retrospective study conducted at a single high-volume tertiary institution, our findings may reflect the surgical team’s expertise and the center’s accumulated experience in managing complex gastric cancer. The proficiency required for laparoscopic dissection in the presence of adhesions is likely surgeon-dependent. Although our results demonstrate feasibility and safety in this setting, their generalizability to lower-volume centers or less experienced teams should be interpreted with caution. Future multi-institutional prospective studies are warranted to validate these findings across diverse clinical environments.

CONCLUSION

For selected gastric cancer patients with a history of UAS, LARG is a viable approach that preserves the advantages of minimally invasive surgery. A history of complex upper abdominal procedures should not be viewed as an absolute contraindication, as our study demonstrates comparable pathologic radicality - including R0 resection and adequate lymph node retrieval - along with favorable postoperative recovery. Clinical decision-making should be individualized and based on a comprehensive assessment rather than prior surgery type alone. Long-term oncologic outcomes warrant further investigation in future studies.

ACKNOWLEDGEMENTS

We are grateful to all staff members and collaborators for their participation in this study.