Overview and preferences of SNUH and NMUH

After screening patients for inclusion and exclusion criteria (Supplementary Fig. 1), 1166 cases including 780 males and 386 females from SNUH and 847 cases including 552 males and 295 females from NMUH were enrolled in this study (Supplementary Table 1). Their ages ranged from 28 to 88 years in SNUH and from 26 to 87 years in NMUH. In the SNUH group, 315 and 851 patients underwent LADG and TLDG, respectively. In the NMUH group, only 90 underwent LADG, while 757 underwent TLDG. As shown in Table S1, there were significant differences in age (p = 0.001), pathological T stage (pT) (p < 0.001), pathological N stage (pN) (p < 0.001), LADG methods (p < 0.001), TLDG methods (p < 0.001), and tumor size (p < 0.001) between SNUH and NMUH, except for sex (p = 0.420) and tumor location (p = 0.460). Notably, 71.78% of the enrolled GC patients were diagnosed at the pT1 stage in SNUH and 51.83% at the pT1 stage in NMUH. This indicated some differences in the preferences and baseline parameters between SNUH and NMUH.

At SNUH, we found that LADG RY (n = 2, 0.17%), LADG Uncut (n = 0, 0%), TLDG RY (n = 23, 1.97%), and TLDG Uncut (n = 20, 1.71%) were rarely performed. B-I (460, 39.45%) and B-II (n = 661, 56.69%) were performed more frequently in the SNUH group than RY and Uncut . In the NMUH group, TLDG B-II (n = 464, 54.78%) and TLDG Uncut (n = 209, 24.68%) were performed, thus indicating a preference for the TLDG method. The LADG method was also performed, but similar to the LADG RY/Uncut to SNUH, the total number of LADG cases was 90 (10.62%) in NMUH, especially in LADG B-I (n = 1, 0.12%). Irrespective of LADG or TLDG, B-I was seldom performed in NMUH.

Pathological T1 stage gastric cancer was more prevalent in SNUH (71.78%) in all laparoscopic distal gastrectomy cases, whereas in NMUH, only half of the patients were at the pT1 stage (51.83%). Along with more early pT stage cases in SNUH, For pT1 cases, we found 76.8% was pN0 at SNUH and 57.14% at NMUH.

The overall complication rate did not differ between SNUH and NMUH

The number of patients with overall complications from both centers is shown in Table 1. No mortality occurred in any of the patients in our study. As shown in Table 1, SNUH and NMUH had overall LADG complication rates of 17.77% and 18.89%, respectively, with no significant difference (p = 0.809) in the LADG subgroup. The only available comparison showed no significant difference in the overall complication rate between the SNUH and LADG groups (p = 0.851).

Table 1 The overall information of complication rates of SNUH and NMUH

In the TLDG subgroup, the overall complication rate of TLDG cases from SNUH was 13.28%, which did not differ from the 11.49% for NMUH (p = 0.279). The B-I analysis could not be completed because of the lack of data. The overall complication rates of TLDG B-II (p = 0.853), RY (p = 0.203), and uncut (p > 0.999) analyses did not show any significant differences between SNUH and NMUH.

Finally, we combined all laparoscopic distal gastrectomy cases for comparison purposes between the two centers, which was 14.49% for SNUH versus 12.28% for NMUH (p = 0.152). These outcomes aligned with our expectations as they did not yield significantly different results. Additionally, we conducted a multivariate analysis of variance to assess whether factors such as age, sex, and cancer stage would impact the outcomes. However, the analysis yielded negative results (Supplementary Table 2), thus indicating that these factors did not influence our results.

SNUH and NMUH did not differ regarding the anastomosis-related complication rates

Gastrointestinal anastomosis is one of the most important procedures performed during laparoscopic distal gastrectomy. As a result, we consider the anastomosis-related complications to have a closer correlation to the quality of laparoscopic surgery. The most serious complications associated with anastomosis are GI leakage and GI bleeding. Therefore, we categorized these as anastomosis-related complications. We conducted the same analysis as previously described for the overall complications (Table 2). The anastomosis-related complication of all LADG cases from SNUH (2.54%) and NMUH (5.55%) did not demonstrate a significant difference (p = 0.386), as well as the LADG B-II analysis (p = 0.206). In the TLDG analysis, B-II (p = 0.324), RY(p = 0.548) and Uncut (p = 0.259) analysis confirmed no difference between SNUH and NMUH again regarding anastomosis-related complications. Among all TLDG cases, 2.82% from SNUH and 1.59% from NMUH demonstrated negative results (p = 0.095). In general, when combining LADG and TLDG cases, there was no statistical difference between SNUH (2.74%) and NMUH (2.01%) for anastomosis-related complications (p = 0.251). We performed a multivariate analysis of variance again to detect if other factors would influence the outcomes, and the findings remained negative (Supplementary Table 3).

Table 2 The overall information of anastomosis-related complication rates of SNUH and NMUHPostoperative hospital stay of SNUH was shorter than that of NMUH

The detailed postoperative data are listed in Table 3. We have found that postoperative time of all LADG cases from SNUH (8(7,11))was statistically shorter than that of NMUH (9(8,10)) (p = 0.0163) in the LADG subgroup.The comparison of the postoperative time after LADG B-II did not show any positive results between SNUH and NMUH (p = 0.4095) while other comparisons were not available.

Table 3 The overall information regarding the length of postoperative stay for SNUH and NMUH

Consistent with the LADG sub-analysis, although B-II (p = 0.4094), RY (p = 0.9816), and Uncut (p = 0.4035) sub-analyses demonstrated negative results, all TLDG cases from SNUH (7(6,9)) had a shorter postoperative stay than those with NMUH(8,(7,9)) (p = 0.0017). Taken together, SNUH had a shorter postoperative hospital stay (7(7,10) days) than NMUH (8(7,9) days) (p = 0.0001). A multivariate analysis of variance confirmed that other factors did not influence the outcomes (Supplementary Table 4).

Resected lymph nodes demonstrated no significant difference between SNUH and NMUH

We then analyzed the number of resected lymph nodes using different methods (Supplementary Table 5). The number of resected lymph nodes was confirmed by pathologists from both centers. Since D2 lymphadenectomy was performed in all cases of NMUH, D1 + lymphadenectomy was performed in early cases, and D2 lymphadenectomy was performed in advanced cases in SNUH, we selected all cases from NMUH and only D2 cases from SNUH. Based on these criteria, we included 187 LADG and 311 TLDG cases from SNUH. As previously reported, resected lymph nodes after B-II did not show a significant difference (p = 0.9718) between both centers, and the same situation emerged in the LADG subgroup compared with 37.43 ± 15.25 in SNUH and 39.67 ± 10.31 in NMUH (p = 0.2038). In the TLDG comparison, B-II (p = 0.3738), RY (p = 0.3263), and Uncut (p = 0.4710) showed similar results. Although resected LNs in SNUH with 40.92 ± 15.18, significantly fewer than 42.81 ± 11.23 of NMUH (p = 0.0263); however, after combining LADG and TLDG, no difference in resected LNs emerged between the two centers. Both centers retrieved more than 40 lymph nodes, which is much more than the number that the NCCN guidelines required.

The lymph node metastasis rate was significantly higher in NMUH than in SNUH

Since the average number of retrieved LNs exceeded 40, this could result in some interesting facts. We then analyzed the distribution of LNs at different T stages (Table 4). Comparing the resected LNs between the two centers based on the pT stage, we found that the LNs of pT1a (p = 0.0002), pT1b (p = 0.0004), and pT2 (p = 0.0446)from SUNH were significantly fewer than those of NMUH, but no difference was observed in pT3 and pT4a stage tumors between SNUN and NMUH. This is expected because all cases of NMUH were performed with D2 dissection, but early cases of SNUH may be performed with D1 + dissection. Generally, the resected number of LNs is approximately 40, which is adequate for the pathological analysis.

Table 4 The resected LNs and LNM rate of the tumors in NMUH and SNUH

We then calculated the lymph nodes metastasis (LNM) rate for each T stage GC, as listed in Table 4. In SNUH, the LNM rate of pT1a, pT1b, pT2, pT3, pT4a and T4b were 1.46%, 20.29%, 48.30%, 65.05%, 70.91% and 100%, respectively, while LN metastasis rate in NMUH of pT1a, pT1b, pT2, pT3, pT4a and pT4b were 9.36%, 31.37%, 51.37%, 75.38%, 85.87% and 100%. Interestingly, the LNM rate in NMUH was significantly higher than that in SNUH (nearly 10% higher at almost every stage). Statistical significance was observed for the pT1a (p < 0.001), pT1b (p = 0.002), and pT4a (p = 0.0027) stages.

We then analyzed the difference in tumor sizes of both centers, and the inspection of tumor size might influence LNM (Supplementary Fig. 2). More interestingly, the tumor size of pT1a (p = 0.0001), pT1b (p = 0.0001), and combined (p = 0.0001) in SNUH was significantly larger than that in NMUH and regarding pT4a there was no significant difference (p = 0.0758), thus indicating that size did not play a role in the LNM rate variance.

No significant difference in the 5-year OS between SNUH and NMUH

We analyzed the overall survival time of patients enrolled from the SNUH and NMUH groups. Combining all enrolled patients, SNUH exhibited a 5-year overall survival rate of 92.22%, while that of NMUH was 84.94% (χ2 = 11.534, p = 0.001). Subsequently, we further compared the 5-year OS of SNUH and NMUH patients based on stages I, II, and III, which revealed no significant difference between the two centers (Fig. 1, Table 5). Similarly, no statistically significant differences were observed between the two centers when further comparisons were made based on the specific TNM stages (Fig. 1, Table 5). In addition, a risk analysis of the prognosis of the two centers was conducted. A multivariate Cox analysis demonstrated that more advanced T and N stages would affect the prognosis, as expected. However, neither different centers nor laparoscopic methods influenced prognosis according to the analysis (Supplementary Table 6).

Fig. 1

5-year overall survival for different stages in SNUH and NMUH

Table 5 No significant differences in the 5-year prognosis were observed between SNUH and NMUH