In recent years, the incidence of adenocarcinoma of esophagogastric junction (AEG) has been increasing progressively (1). Radical resection serves as the cornerstone for long-term survival in patients with AEG. It is well-established that achieving a negative proximal resection margin (PRM) and performing thorough lymphadenectomy are critical components of radical resection. It has long been recognized that a positive PRM significantly compromises survival for patients with AEG (2, 3). Therefore, to ensure a negative PRM, studies have been conducted to determine the maximum distance of tumor infiltration toward the oral side, which serves as the standard for determining the length of distal esophageal resection during surgery. However, due to inevitable heterogeneity in patients enrolled across different centers, conclusions from various studies have been inconsistent. For instance, Ito H et al. (4) recommended a PRM length of 6 cm for T3/T4 patients to achieve pathological negativity, while Mariette et al. (2) suggested a length of 8 cm, and Papchristou (5) advocated for a length of up to 12 cm. Furthermore, some authors observed local recurrence in advanced patients even with negative PRM, hypothesizing that this might be due to insufficient length of PRM (6). Consequently, they recommended increasing the length even when the PRM was negative, in order to reduce the risk of recurrence.

According to Siewert classification, AEG is divided into three subtypes (7). Siewert type I refers to tumors whose center located 1-5cm above the esophagogastric junction (EGJ), primarily operated on via trans-thoracic approach; Siewert type III involves tumors with center 2-5cm below the EGJ, mainly managed through trans-abdominal approach. In contrast, Siewert type II AEG, defined as tumors with center located between 1 cm above and 2 cm below the EGJ, had been subject to considerable debate regarding the optimal surgical approach (8–10). Historically, surgeons opted for transthoracic approach for Siewert type II AEG to achieve adequate length of PRM mentioned above. However, studies by Hulscher et al. (11) and the Japan Clinical Oncology Group (JCOG) 9502 trial (12) found no significant difference in long-term survival between transthoracic and abdominal-transhiatal approaches, besides, the transthoracic approach associated with poorer short-term recovery. A meta-analysis of 1,155 patients included 8 studies further confirmed both approaches yield comparable long-term survival, but the abdominal-transhiatal approach significantly reduced 30-day mortality, pulmonary complications, and shortened hospital stays (13). Meanwhile, Kurokawa Y et al. (14) discovered that Siewert type II AEG patients with esophageal invasion exceeding 3cm had a higher rate of mediastinal lymph node metastasis, and transthoracic lymphadenectomy could offer survival benefits in such cases. Therefore, for patients with Siewert type II AEG invading the esophagus less than 3cm, the abdominal-transhiatal approach is recommended as the standard surgical procedure. Nevertheless, it is undeniable that the PRM length is closely related to the surgical approach, and the length of PRM is restricted by the limited subdiaphragmatic space in the abdominal-transhiatal approach. Kinoshita T et al. (15) reported a median length of PRM in this procedure was 14.5mm (range 10-23mm). While Sugita S et al. (16) observed a median of only 9mm (range 3-26mm) in laparoscopic-assisted abdominal-transhiatal procedure. Consequently, whether the length of PRM in abdominal-transhiatal approach for Siewert type II AEG affects survival and prognosis remains a key clinical concern.

In China, locally advanced AEG accounts for a considerable proportion. According to guidelines recommendations, perioperative chemotherapy combined with surgical resection is the standard treatment strategy for locally advanced AEG (17, 18). However, many previous studies focused on the safe length of PRM for AEG have excluded patients receiving neoadjuvant chemotherapy (NAC) (19–22), and thus the optimal PRM length for Siewert type II AEG patients after NAC remains undetermined. Accordingly, we conducted this study utilizing a laparoscopic-assisted abdominal-transhiatal approach to treat patients with locally advanced Siewert type II AEG. The primary objective was to investigate the association between PRM length and survival outcomes, with particular emphasis on determining the optimal PRM length following NAC.

To determine the optimal PRM length for AEG, two viewpoints had historically emerged (26). One advocated that the length should ensure pathological negativity to the greatest extent possible; the other, from a prognostic perspective, employed statistical methods to identify the optimal PRM length associated with the highest survival rate. A significant discrepancy was observed between the two viewpoints: the former generally considered a length of 6–8 cm (or more) necessary to achieve pathological negativity in all patients (2, 4, 5), whereas the latter revealed that a length exceeding 2 cm yield significantly better survival outcomes than shorter length, thus advocating 2 cm was sufficient (22).

Upon closer analysis, both viewpoints presented certain limitations. In the first case, the required length of PRM was determined by a few extreme cases in the cohort – patients with the farthest spread of the tumor through mucosa and/or submucosa or skipping metastasis. Applying the length obtained from extreme cases as universal standard would necessitate extensive distal esophageal resection in all patients, inevitably increasing the rate of thoracotomies (2, 10, 27) and corresponding perioperative risks (11). Moreover, opponents noted that patients with skip metastases often presented with extensive lymphatic invasion; and their prognosis exhibited poor even with additional esophageal resection (28). Meanwhile, the second viewpoint underestimated the importance of the PRM status. While it might achieve the best overall survival at the population level, individual patient could still have opportunities for positive PRM and then suffer high postoperative recurrence rates and poor survival outcomes.

Therefore, this study explores a novel therapeutic strategy that prioritizes achieving negative status of PRM before investigating the relationship between the length and survival outcomes. We propose that this approach avoids the rigidity of a one-size-fits-all method while also circumventing the limitation of solely focusing on survival optimization in large cohorts at the expense of individual outcomes. In summary, this strategy emphasizes both the importance of negative status of PRM and the consideration of survival impact. Although this method has been reported in gastric cancer research (24, 25), it is less documented in AEG (19) and has not been reported in patients with AEG following NAC.

In this study, we found that patients with positive PRM had significantly more advanced pathological stages compared to those with negative status, and the difference was statistically significant. However, no statistically significant differences were observed in various clinicopathological characteristics or the length of PRM between the two groups. This suggested that a positive PRM was not related to surgical technique or quality but might be associated with stage and adverse biological behavior (29–32). Survival analysis indicated that patients with positive PRM had worse OS and RFS than those with negative margins, with the difference in RFS being statistically significant. These findings were consistent with reports in the literature from 2021 (31). Multivariate analysis further confirmed that PRM status was an independent risk factor for RFS in patients with Siewert type II AEG.

Given that the status of the PRM was associated with survival outcomes and prognosis, we excluded patients with positive PRM in the subsequent analysis, focusing specifically on those of locally advanced Siewert type II AEG with negative PRM. In this cohort, patients were divided into two groups based on the treatment strategy received: SA group and NAC group. In the SA group, survival analysis indicated that PRM length was not an independent factor influencing patients’ survival or prognosis, which was consistent with the findings of Feng et al. (19). A more detailed comparison revealed that, in our study, the patients in SA group had a TNM stage III rate of 76.5%, a PRM length of 1.24 ± 0.88 cm, and a 5-year overall survival rate of 54.9%. In contrast, Feng et al. reported a stage III rate of 54.8%, a proximal margin length of 2.24 ± 0.99 cm, and a comparable survival rate (54.7%). Thus, despite differences in PRM length, the two cohorts achieved similar survival outcomes, supporting the conclusion that PRM length was not associated with survival. To further strengthen the reliability of this conclusion, we referred to previous studies and evaluated different cutoff values for PRM length, comparing survival outcomes above and below these thresholds (25). The results showed that regardless of whether the cutoff was set at 5 mm, 10 mm, or 20 mm, there were no statistically significant differences in OS or RFS between the subgroups. In particular, even when comparing patients with PRM lengths ≤5 mm and those with lengths >20 mm, no statistically significant survival difference was observed. Therefore, it could conclude that in SA group with negative status PRM, PRM length was not associated with patient survival.

In addition to confirming that the PRM length was not associated with survival outcomes in locally advanced AEG patients with negative PRM in the SA group, this study further investigated the relationship between PRM length and survival prognosis in patients who underwent NAC followed surgery. The latter held significant clinical relevance for two main reasons: firstly, previous studies rarely addressed the impact of PRM length on prognosis in patients receiving NAC (19, 20); secondly, in these patients, tumors might exhibit varying degrees of regression. Consequently, determining tumor boundaries and PRM length after NAC posed greater challenges compared to surgical resection alone in patients with locally advanced AEG. To address this, we employed techniques such as visual inspection, palpation, and intraoperative gastroscopy (33) to generally determine the upper boundary of the tumor. Furthermore, considering tumor regression after chemotherapy, we precisely measured the PRM length by examining formalin-fixed specimens microscopically.

In this cohort, 53 patients (49%) received NAC. Comparative analysis of clinicopathological characteristics revealed that the NAC group had significantly smaller tumor size, lower rates of lymph-vascular invasion and perineural invasion, as well as lower pTNM staging compared to the SA group. These findings indicated that NAC could reduce tumor size, mitigate adverse histologic features, and achieve downstaging—results consistent with previous studies (34). However, although tumor size decreased in the NAC group, there was no statistically significant difference in PRM length between the two groups. This suggested that the reduction in tumor diameter after NAC did not occur along the longitudinal axis and therefore did not lead to a significant change in PRM length. Additionally, although RFS and OS were longer in the NAC group than in the SA group, the differences were not statistically significant. This was also consistent with some prior researches indicating that while NAC might improve resectability, it did not significantly prolong patient survival (35, 36). On the other side, the survival comparisons between NAC and SA groups in this study were descriptive and not intended to demonstrate treatment efficacy.

In patients receiving NAC, results revealed that PRM length was not an independent factor influencing survival or prognosis. Similarly, different cutoff values were applied to categorize PRM lengths, and survival outcomes were compared across subgroups. No statistically significant differences were observed in survival outcomes. Furthermore, a comparison of clinicopathological characteristics and survival between the two extreme subgroups—PRM length ≤5 mm and >20 mm—also showed no statistically significant differences. These findings indicated that among patients who had undergone NAC, as long as the status of PRM was negative, survival outcomes were not associated with the length of PRM.

This study indicated that the length of PRM was not associated with survival prognosis, provided that the status of PRM was negative. For AEG, submucosal infiltration may occur, making visual assessment of normal mucosa alone insufficient to determine a negative PRM. Particularly after NAC, tumor regression may render visual mucosal evaluation even less reliable for assessing margin status. Intraoperative FSA provided critical information on margin involvement. For patients with positive result of FSA, extended resection to achieve negative PRM offered the opportunity for R0 resection, which might improve postoperative survival (37, 38). However, previous studies have shown that even after conversion to negative status, survival in locally advanced patients does not improve (39), leading to doubts regarding the value of FSA.

Until 2025, an Indian study (40) demonstrated that survival improved after achieving negative margins, including in locally advanced Siewert type II AEG patients, thereby reaffirming the value of intraoperative FSA. In present study, five patients with PRM involved on FSA underwent re-resection intraoperatively, all achieving negative margins (R1-to-R0 group). Survival outcomes were compared among R1-to-R0 group, those with positive margins on histological pathology (R1 group), and those with negative margins (R0 group). Although limited by small sample size, this exploratory analysis showed that patients in R1-to-R0 group had significantly better survival than R1 group (RFS: 65.6 ± 14.9 vs. 16.0 ± 4.3, p=0.025; OS: 67.6 ± 14.0 vs. 46.0 ± 11.5, p=0.264) and had no statistically significant difference compared to R0 group (RFS: 65.6 ± 14.9 vs. 59.6 ± 4.0, p=0.595; OS: 67.6 ± 14.0 vs. 70.3 ± 3.9, p=0.814). These findings are consistent with emerging evidence from recent studies, suggesting that intraoperative margin revision may offer survival benefits. Therefore, our findings support the routine use of intraoperative FSA, which is consistent with current literature recommendations.

Our results contradicted the conclusions of Lee’s study (6). Lee observed and analyzed 29 patients with locally advanced Siewert type II AEG who underwent transabdominal total gastrectomy and D2 lymphadenectomy, with a PRM length of 7.5 ± 6.6 mm. During the follow-up period, the local recurrence rate in this group was higher than that in patients with Siewert type III AEG and upper third gastric cancer. Therefore, the author inferred that extending the PRM length may reduce the local recurrence rate. A careful analysis revealed several issues about Lee’s study: First, the sample size of locally advanced Siewert type II AEG was relatively small, compromising the reliability of its conclusions. Second, patients with Siewert type II and III, as well as upper third gastric cancer, belonged to different etiological categories, and the factors related to local recurrence were complex and could not be simply attributed to insufficient PRM length. Third, Lee’s study lacked data to support its conclusions and did not specify an optimal PRM length. Fourth, half of the AEG patients with local recurrence in Lee’s study also had distant metastases, which significantly impacted overall survival. In summary, we believed that Lee’s conclusions required further validation. Meanwhile, compared to extending the PRM length in locally advanced patients with negative PRM status, we considered it more important and feasible to select appropriate treatment strategies for patients, such as perioperative systemic therapy. In our study, the survival prognosis of patients in the NAC group was superior to that of the SA group, although the difference was not statistically significant, it still suggested a trend of improved patient outcomes with perioperative systemic therapy.

The present study has some limitations. Firstly, the retrospective design and relatively small samples would introduce selection bias. Furthermore, only 4 patients (3.7%) had positive PRM status, which resulted in wide confidence intervals and potential instability of hazard ratio estimates. Our findings regarding the prognostic impact of PRM status should be interpreted as hypothesis-generating rather than definitive conclusions. Secondly, all patients were treated at a single Chinese center with CapeOX/SOX regimens. Previous studies have suggested potential differences in tumor biology and treatment response between Eastern and Western populations. Whether our findings can be generalized to patients treated with different regimens (e.g., FLOT, ECF, et al.) warrants further investigation. Thirdly, with the advanced in clinical oncology, neoadjuvant therapy no longer consists solely of chemotherapy, but also includes neoadjuvant chemoradiotherapy, immunotherapy, and targeted therapy. The impact of these newer modalities on tumor regression patterns and optimal margin length may differ from traditional chemotherapy. Future studies should explore PRM length-prognosis relationship in patients with locally advanced Siewert type II AEG receiving these emerging therapies.

A positive status of PRM appears to be associated with adverse survival outcomes of patients with locally advanced Siewert type II AEG after surgery. However, for patients with negative status, the length of the PRM does not influence survival, regardless of whether they undergo surgical resection alone or NAC followed by surgery.