Impact of laparoscopic gastrectomy on long‐term prognosis of patients with primary T3 or more advanced gastric cancer: A propensity score matching analysis

Abstract Background There is no consensus regarding a better long‐term prognosis with laparoscopic gastrectomy than with open surgery in patients with advanced gastric cancer, especially in patients with T3 or more advanced gastric cancer. We investigated the impact of laparoscopic gastrectomy on the long‐term prognosis of patients who underwent radical gastrectomy for primary T3 or more advanced gastric cancer. Methods This single‐center, retrospective cohort study included 294 consecutive patients who underwent radical gastrectomy for primary T3 or more advanced gastric cancer from April 2008 through April 2017. We compared overall survival between laparoscopic and open surgery, using propensity score matching to adjust for baseline characteristics. We also investigated prognostic factors for overall survival by a forward stepwise procedure of Cox proportional hazards regression for multivariate analysis. Results There were 136 (46.3%) and 158 (53.7%) patients in the laparoscopy and open groups, respectively. The median follow‐up period was 39 mo. After matching, there were 97 patients in each group, with no significant differences in background characteristics. After matching, the overall survival was significantly worse in the open group than in the laparoscopy group (P < 0.001). Multivariate analyses also showed that open surgery was an independent poor prognostic factor for overall survival (hazard ratio: 2.160, 95% concordance interval: 1.365–3.419, P = 0.001). Conclusion Overall survival may be better with laparoscopic gastrectomy than with open surgery for patients with primary T3 or more advanced gastric cancer.


| Indications for the procedure
Each patient underwent radical gastrectomy with D2 nodal dissection. Laparoscopic surgery or open surgery were indicated for all advanced gastric cancers. The surgeon selected the surgical approach, which was not changed throughout the procedure. For quality control, only one surgeon, who was a professional in gastrectomy and responsible for the quality of the laparoscopic surgery, was involved in this study, as in the JCOG-0912 study protocol. The procedure for nodal dissection and reconstruction was the same in both groups.

| Postoperative management with S-1 adjuvant chemotherapy
Postoperative adjuvant chemotherapy with S-1 was administered to patients with stage II and stage III disease for a maximum of 1 y, according to the Japanese gastric cancer treatment guidelines. 8 The regimen was initiated with 80-120 mg/d for 4 weeks, followed by 2 weeks of rest. If side effects were observed, the dose was gradually reduced from 120 mg/d to 100 mg/d or from 100 mg/d to 80 mg/d, according to the guidelines. The treatment was discontinued if the side effects could not be controlled with dose optimization or two or more steps of dose reduction, or if recurrence of disease was confirmed during adjuvant chemotherapy. For recurrent cases, additional chemotherapy was administered according to the Japanese gastric cancer treatment guidelines. 8 Patients did not receive any treatment other than adjuvant chemotherapy with S-1 until recurrence.
The patients were followed up at an outpatient clinic.
Hematological tests were performed at least every 2-3 weeks during S-1 chemotherapy and at least every 3 mo for 5 y after chemotherapy completion. Computed tomography (CT) was performed every 6 mo, while endoscopy was performed every year until 5 y after surgery.

| Body composition analysis
Before surgery, we measured the visceral fat area and skeletal muscle mass on plain CT images using the graphical analysis software Ziostation (Ziosoft, Newark, CA, USA). Visceral fat mass was measured at the umbilical level, and skeletal muscle mass was measured at the third lumbar vertebral level. Visceral fat mass and skeletal muscle mass measured on a single CT image slice were divided by the patient's height in meters squared to obtain the visceral adipose tissue index (VAI) and skeletal muscle mass index (SMI), respectively. 9 The cutoff values for VAI and SMI were set separately for men and women using the cutoff values previously reported for Asians; the cutoff values for VAI were 35.43 cm 2 /m 2 for men and 24.85 cm 2 /m 2 for women, while those for SMI were 40.80 cm 2 /m 2 for men and 34.90 cm 2 /m 2 for women. [9][10][11] Patients with VAI and SMI values above and below the cutoff values were considered to have high and low VAI/SMI, respectively.

| Outcomes
The primary outcome was OS, defined as the period between surgery and death. The secondary outcomes were relapse-free survival (RFS), cancer-specific survival (CSS), and other-cause survival (OCS).
RFS was defined as the period between surgery and recurrence or death, whichever occurred first. Preoperative chronic inflammation was defined by a C-reactive protein (CRP) level of ≥0.5 mg/dL just before the surgery, and OS was compared between laparoscopic and open surgery in patients with or without preoperative chronic inflammation.
Postoperative complications were defined as complications of grade ≥2 according to the Clavien-Dindo classification (CD) that occurred within 30 d after surgery. Complications of CD grade ≥3 were considered severe complications. The body weight loss (BWL) rate was calculated for 1 mo, 6 mo, and 1 y.

| Statistical analyses
The PSM method was used to adjust for differences in the background characteristics of patients and selection bias in this nonrandomized study. A logistic regression model including age, sex, clinical stage, surgical procedure, comorbidities, and preoperative CRP as factors was used to estimate the propensity score, and 1:1 matching between the two groups was achieved using the nearest-neighbor matching method. The caliper size was 0.20.
Postoperative outcomes were compared between the two groups after matching.
For continuous variables, we used the Mann-Whitney U-test, and for categorical variables, we used the Chi-square test or Fisher's exact test. The log-rank test was used for long-term survival. We used a forward stepwise procedure of Cox proportional hazards regression for multivariate analysis to identify prognostic factors for OS, and calculated hazard ratios (HRs). All data were analyzed using EZR software (Saitama Medical Center, Jichi Medical University, Japan). P < 0.05 was considered statistically significant. Table 1 shows the patients' characteristics. In total, 294 patients met the eligibility criteria, and they were divided into laparoscopy  Table 2 shows the postoperative outcomes after matching. The operating time was longer (P < 0.001), while intraoperative blood loss was lesser in the laparoscopy group (P < 0.001) than in the open group. There were no between-group differences in the rates of total postoperative complications (P = 1.000) and severe postopera-   Figure 1D). Figure 2 shows the OS based on the adjuvant chemotherapy administered after matching. OS was better in the order of S-1 completion group, S-1 discontinuation group, and group without S-1 (P = 0.018).

| Prognostic factors for OS
The results of multivariate analysis with a forward stepwise procedure of the prognostic factors for OS in all patients are shown in  showed that the time from relapse to death did not differ between surgery with adjuvant S-1 chemotherapy and surgery alone, 13 and that differences in compliance may not affect the time from relapse to death. Therefore, good compliance with laparoscopic surgery may contribute to differences in RFS, which may directly correlate with OS and CSS. We also reported that laparoscopic surgery may prolong RFS compared to that after open surgery in patients with gastric cancer showing reduced muscle loss. 14 A meta-analysis has shown that the side effects of chemotherapy are more severe in patients with reduced muscle loss. 15 In this study, more than half the  Abbreviations: CHF, chronic heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CRP, C-reactive protein; IQR, interquartile range; SD, standard deviation; SMI, skeletal muscle mass index; VAI, visceral adipose tissue index.