Prognosis comparison and prognostic model construction in patients with Siewert type II adenocarcinoma of the esophagogastric junction after neoadjuvant therapy between proximal gastrectomy and total gastrectomy Authors: Guoyi Shen, Pubo Shi, Pengjie Yang, Benben Zhu, Yong Li, Shuchen Chen Published: 2026 ABSTRACT AbstractBackgroundThe incidence of adenocarcinoma of the esophagogastric junction (AEG) is rising globally. As the most common subtype, Siewert type II AEG still lacks consensus on surgical strategy. For patients treated with neoadjuvant therapy, long-term survival differences between proximal gastrectomy (PG) and total gastrectomy (TG) remain unclear, and surgical-stratified prognostic models are currently unavailable. This study aimed to compare long-term outcomes of PG versus TG in Siewert type II AEG patients after neoadjuvant therapy, and to construct stratified prognostic models for individualized clinical decision-making and prognostic assessment.MethodsA total of 1,364 patients with Siewert type II AEG who received neoadjuvant therapy followed by PG or TG between 2004 and 2015 were enrolled from the Surveillance, Epidemiology, and End Results (SEER) database. Propensity score matching (PSM) was performed at a 2:1 ratio with a caliper width of 0.01 to balance confounding factors between groups. Survival analyses for overall survival (OS) and cancer-specific survival (CSS) were conducted both before and after PSM using the Kaplan-Meier analysis, followed by multivariate Cox proportional hazards regression to verify the independent association between surgical approach and long-term prognosis. Prognostic models for OS and CSS were separately constructed for the PG and TG groups. Variables were first screened by univariate Cox regression, and those with P<0.05 were entered into multivariate analysis; independent factors with P<0.05 were retained to establish nomograms. Model performance was validated using the concordance index (C-index), receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).ResultsSurvival analyses before and after PSM consistently demonstrated that patients undergoing PG exhibited significantly superior OS and CSS compared with those receiving TG, with all comparisons achieving statistical significance (P<0.05). Multivariate Cox proportional hazards regression further identified TG as an independent risk factor associated with adverse long-term prognosis. To enable individualized risk stratification, we constructed four prognostic models stratified by surgical approach: the OS model for PG patients incorporated gender, age, tumor (T) stage, node (N) stage, and metastasis (M) stage; the OS model for TG patients comprised age, N stage, M stage, and the number of examined lymph nodes (ELNs); the CSS model for PG patients included gender, T stage, N stage, and M stage; and the CSS model for TG patients consisted of N stage and M stage. All models prioritized tumor-node-metastasis (TNM) stage-related indices as core prognostic predictors and exhibited robust discriminatory capacity, satisfactory calibration performance, and favorable clinical utility.ConclusionsIn Siewert type II AEG patients receiving neoadjuvant therapy, PG confers better long-term survival. For patients with marked tumor regression, PG achieves an optimal balance between oncological safety and survival benefit. Prognostic models stratified by surgical approach aid precise clinical risk stratification, individualized surgical decision-making and postoperative management.Keywords: Esophagogastric junction and upper third gastric adenocarcinoma (EGJ and upper third gastric adenocarcinoma), proximal gastrectomy (PG), total gastrectomy (TG), neoadjuvant therapy Highlight box Key findings • For Siewert type II adenocarcinoma of the esophagogastric junction (AEG) patients after neoadjuvant therapy, proximal gastrectomy (PG) was associated with better overall survival and cancer-specific survival than total gastrectomy (TG). For patients with obvious tumor regression after neoadjuvant therapy, PG could balance oncological safety and survival benefits without extended resection. Prognostic models were constructed for different surgical approaches and survival endpoints. What is known and what is new? • Siewert type II AEG is commonly encountered, yet controversy persists regarding the optimal surgical strategy between PG and TG. Neoadjuvant therapy plays a vital role in the management of Siewert type II AEG. However, no consensus has been reached regarding the prognostic differences between PG and TG among patients receiving neoadjuvant therapy, and individualized stratified prognostic models remain lacking. • Among patients with Siewert type II AEG who received neoadjuvant therapy, PG was associated with better survival prognosis than TG and could avoid unnecessary extended resection. In this study, prediction models were separately established according to different surgical procedures, which better fitted the prognostic characteristics of patients undergoing different operations and improved the rationality and clinical applicability of the models. What is the implication, and what should change now? • This study suggests that PG can be recommended as a preferred surgical approach for eligible patients receiving neoadjuvant therapy. The stratified models constructed according to surgical procedures enable accurate risk stratification and individualized treatment, challenging the standardized uniform surgical strategy of TG. Introduction In recent decades, the overall incidence of esophageal and gastric cancers has shown a significant downward trend, whereas the incidence of adenocarcinoma of the esophagogastric junction (AEG) has continued to increase rapidly worldwide due to the rising prevalence of risk factors including gastroesophageal reflux disease, obesity, smoking, and Helicobacter pylori infection (1,2). AEG is a distinct type of adenocarcinoma that crosses the esophagogastric junction (EGJ), with the tumor center located within 5 cm above or below the EGJ line. The Siewert classification is currently widely used in clinical practice and research for its categorization: type I AEG refers to distal esophageal adenocarcinoma, with the tumor center located 1–5 cm above the EGJ; type II represents true cardiac carcinoma, with the tumor center positioned from 1 cm above to 2 cm below the EGJ; and type III is located 2–5 cm below the EGJ (3,4). Among these three types, Siewert type II AEG is the most common and clinically controversial, with no consensus on its surgical approach, resection extent, or prognostic evaluation due to its unique location and complex anatomy (5). Neoadjuvant therapy plays an important role in the comprehensive treatment of Siewert type II AEG (6). It can reduce tumor volume, improve the radical resection rate, and decrease the risk of recurrence and metastasis, thus becoming an important preoperative strategy in clinical practice (5,7). However, neoadjuvant therapy may alter tumor stage and lymph node metastasis status, which could further affect the survival benefits of different surgical procedures. Current research is still unable to provide precise evidence for surgical selection in patients treated with neoadjuvant therapy. The common surgical procedures for Siewert type II AEG include proximal gastrectomy (PG) and total gastrectomy (TG) (8). PG preserves partial gastric function and helps improve postoperative quality of life, but carries concerns regarding insufficient surgical margin safety and limited lymph node dissection (9,10). Given the requirement for oncologically safe distal resection margins and radical lymph node dissection, TG is recommended as the standard procedure (11). Although TG allows more thorough lymph node dissection and safer resection margins, thereby reducing the risk of local recurrence, it is associated with postoperative complications such as malnutrition and reflux, which adversely affect quality of life (12,13). For patients with Siewert type II AEG who have received neoadjuvant therapy, consensus has not yet been reached regarding the long-term survival benefits of the two procedures in terms of overall survival (OS) and cancer-specific survival (CSS). Moreover, prognostic prediction models based on different surgical modalities have not been established. To address these clinical controversies and research gaps, the present study included patients with Siewert type II AEG who received neoadjuvant therapy to compare the effects of PG and TG on OS and CSS, clarify the differences in survival benefits between the two procedures, and provide evidence‑based support for clinical surgical decision-making; furthermore, we developed individualized prognostic prediction models for patients undergoing PG and TG, respectively, to assess postoperative survival risk and offer a precise tool for optimizing treatment strategies and improving patient prognosis. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0833/rc). Methods Patients A total of 1,364 patients with Siewert type II AEG who received neoadjuvant therapy followed by TG or PG were enrolled in this study from the Surveillance, Epidemiology, and End Results (SEER) database (https://seer.cancer.gov/). Inclusion criteria: year of diagnosis from 2004 to 2015; confirmed diagnosis of Siewert type II AEG, with the classification and coding requirements met simultaneously: screened by the “EsophagusGEJunction” code in the Collaborative Staging (CS) Schema V0204+entry, with a primary site code of 16.0 (a primary site entry of “C16.0-Cardia, NOS”); patients with a primary malignancy of the Siewert type II AEG; pathologically confirmed AEG, with histology consistent with the ICD-O-3 codes (8140–8147, 8160–8162, 8180–8221, 8250–8507, 8514–8551, 8571–8574, 8576, 8940, and 8941); having received neoadjuvant therapy and undergone TG (surgical codes 40–42, 52) or PG (surgical codes 33, 51). A total of 1,486 patients initially met the above inclusion criteria. Exclusion criteria: missing marital status (n=35), missing race information (n=1), missing tumor (T) stage (n=40), missing node (N) stage (n=5), missing metastasis (M) stage (n=5), and missing number of examined lymph nodes (ELNs) (n=30). The variables collected in this study included age, gender, race, marital status, T stage, N stage, M stage, G grade, surgical approach, type of neoadjuvant therapy, tumor size, number of lymph node regions removed (NLR), number of ELNs, rural-urban continuum, and median household income. The study endpoints were OS and CSS. OS was defined as the time from tumor diagnosis to death from any cause, and CSS was defined as the time from tumor diagnosis to death due to the primary tumor under investigation. Statistical analysis All variables in this study were treated as categorical data and presented as numbers and percentages (%). To investigate the effect of surgical approach on patient prognosis, Kaplan-Meier survival analysis was first performed to compare the prognostic differences in OS and CSS between the PG group and TG group before propensity score matching (PSM). Subsequently, to balance baseline confounding factors between the two groups, PSM was conducted at a ratio of 2:1 (PG:TG) with a caliper width of 0.01. After matching, Kaplan-Meier survival analysis was performed again to compare OS and CSS between the two groups for a more objective evaluation of the surgical approach. Multivariate regression analysis was then applied in the entire overall population to assess the independent effect of surgical approach on prognosis, with results reported for unadjusted, minimally adjusted, and fully adjusted models. Prognostic models for OS and CSS were separately constructed for patients undergoing PG and TG. All patients were randomly divided into a training cohort and a validation cohort at a ratio of 7:3; the training cohort was used to develop nomograms, while the validation cohort was used for internal validation. In the training cohort, variables with P<0.05 in univariate regression analysis were entered into multivariate regression analysis, and variables with P<0.05 in multivariate analysis were retained as predictors to establish the final prognostic models. Model performance was evaluated as follows: discrimination was assessed using Harrell’s concordance index (C-index) and receiver operating characteristic (ROC) curves; calibration was assessed using calibration plots; and clinical utility was evaluated using decision curve analysis (DCA). All statistical analyses were performed using R software (version 4.2.2; R Foundation for Statistical Computing, Vienna, Austria) and FreeStatisticsV2.4.0. A two-sided P<0.05 was considered statistically significant. Ethics statement The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The data were obtained from the public de-identified SEER database; therefore, informed consent and local ethical approval were not required. Results Baseline characteristics A total of 1,364 patients with Siewert type II AEG who received neoadjuvant therapy were enrolled from the SEER database, among whom 941 patients underwent PG (69.0%) and 423 patients underwent TG (31.0%). Most patients were younger than 65 years (830, 60.9%), and the cohort was predominantly male (1,159, 85.0%), White (1,227, 90.0%), and married (1,020, 74.8%). For tumor staging, T2 stage was most prevalent (773, 56.7%), N1 stage was predominant (800, 58.7%), and 1,298 (95.2%) were at M0 stage, and most had G Ⅲ (738, 54.1%). Neoadjuvant chemoradiotherapy was the most common treatment (1,054, 77.3%). Additionally, most patients had a tumor size of 2–4 cm (334, 24.5%), NLR ≥3 (1,273, 93.3%), and ≥15 ELNs (699, 51.2%). After PSM, all variables were evenly distributed (Figure S1). Baseline characteristics are shown in Table 1. Table 1Baseline characteristics of patients with Siewert type II adenocarcinoma of esophagogastric junction receiving neoadjuvant therapy before and after PSMCharacteristicsUnmatchedPSMTotal (n=1,364)PG (n=941)TG (n=423)PTotal (n=1,008)PG (n=672)TG (n=336)PAge, years0.080.65 <65830 (60.9)587 (62.4)243 (57.4)620 (61.5)410 (61.0)210 (62.5) ≥65534 (39.1)354 (37.6)180 (42.6)388 (38.5)262 (39.0)126 (37.5)Sex0.460.90 Male1159 (85.0)804 (85.4)355 (83.9)857 (85.0)572 (85.1)285 (84.8) Female205 (15.0)137 (14.6)68 (16.1)151 (15.0)100 (14.9)51 (15.2)Race0.0080.50 Black45 (3.3)23 (2.4)22 (5.2)26 (2.6)19 (2.8)7 (2.1) White1,227 (90.0)861 (91.5)366 (86.5)915 (90.8)612 (91.1)303 (90.2) Others92 (6.7)57 (6.1)35 (8.3)67 (6.6)41 (6.1)26 (7.7)Marital status0.320.96 Unmarried344 (25.2)230 (24.4)114 (27.0)262 (26.0)175 (26.0)87 (25.9) Married1020 (74.8)711 (75.6)309 (73.0)746 (74.0)497 (74.0)249 (74.1)T stage0.180.58 T1141 (10.3)96 (10.2)45 (10.6)111 (11.0)68 (10.1)43 (12.8) T2773 (56.7)521 (55.4)252 (59.6)579 (57.4)390 (58.0)189 (56.2) T3398 (29.2)291 (30.9)107 (25.3)273 (27.1)182 (27.1)91 (27.1) T452 (3.8)33 (3.5)19 (4.5)45 (4.5)32 (4.8)13 (3.9)N stage0.950.99 N0397 (29.1)273 (29)124 (29.3)287 (28.5)191 (28.4)96 (28.6) N1800 (58.7)553 (58.8)247 (58.4)596 (59.1)398 (59.2)198 (58.9) N2136 (10.0)95 (10.1)41 (9.7)99 (9.8)65 (9.7)34 (10.1) N331 (2.3)20 (2.1)11 (2.6)26 (2.6)18 (2.7)8 (2.4)M stage0.220.77 M01298 (95.2)900 (95.6)398 (94.1)954 (94.6)637 (94.8)317 (94.3) M166 (4.8)41 (4.4)25 (5.9)54 (5.4)35 (5.2)19 (5.7)Grade0.750.97 I56 (4.1)40 (4.3)16 (3.8)36 (3.6)23 (3.4)13 (3.9) II438 (32.1)306 (32.5)132 (31.2)337 (33.4)226 (33.6)111 (33) III738 (54.1)504 (53.6)234 (55.3)541 (53.7)358 (53.3)183 (54.5) IV24 (1.8)19 (2)5 (1.2)15 (1.5)10 (1.5)5 (1.5) Unknown108 (7.9)72 (7.7)36 (8.5)79 (7.8)55 (8.2)24 (7.1)Neoadjuvant therapy<0.0010.60 Chemotherapy310 (22.7)172 (18.3)138 (32.6)239 (23.7)156 (23.2)83 (24.7) Chemoradiotherapy1054 (77.3)769 (81.7)285 (67.4)769 (76.3)516 (76.8)253 (75.3)Tumor size0.020.96 <2 cm131 (9.6)89 (9.5)42 (9.9)98 (9.7)65 (9.7)33 (9.8) ≥2, <4 cm334 (24.5)243 (25.8)91 (21.5)238 (23.6)157 (23.4)81 (24.1) ≥4, <6 cm331 (24.3)244 (25.9)87 (20.6)209 (20.7)136 (20.2)73 (21.7) ≥6 cm244 (17.9)159 (16.9)85 (20.1)200 (19.8)137 (20.4)63 (18.8) Unknown324 (23.8)206 (21.9)118 (27.9)263 (26.1)177 (26.3)86 (25.6)NLR0.260.66 <391 (6.7)58 (6.2)33 (7.8)70 (6.9)45 (6.7)25 (7.4) ≥31,273 (93.3)883 (93.8)390 (92.2)938 (93.1)627 (93.3)311 (92.6)ELNs0.150.69 <15665 (48.8)471 (50.1)194 (45.9)468 (46.4)315 (46.9)153 (45.5) ≥15699 (51.2)470 (49.9)229 (54.1)540 (53.6)357 (53.1)183 (54.5)Rural-urban continuum0.070.81 <250,000 population113 (8.3)77 (8.2)36 (8.5)83 (8.2)59 (8.8)24 (7.1) ≥250,000, <1 million population304 (22.3)225 (23.9)79 (18.7)201 (19.9)135 (20.1)66 (19.6) ≥1 million population751 (55.1)516 (54.8)235 (55.6)563 (55.9)373 (55.5)190 (56.5) Others196 (14.4)123 (13.1)73 (17.3)161 (16.0)105 (15.6)56 (16.7)Median household income0.090.92 <$70,000365 (26.8)239 (25.4)126 (29.8)272 (27.0)182 (27.1)90 (26.8) ≥$70,000999 (73.2)702 (74.6)297 (70.2)736 (73.0)490 (72.9)246 (73.2)Status0.250.29 Alive367 (26.9)262 (27.8)105 (24.8)267 (26.5)185 (27.5)82 (24.4) Dead997 (73.1)679 (72.2)318 (75.2)741 (73.5)487 (72.5)254 (75.6)Cause0.190.08 Alive or dead of other cause545 (40.0)387 (41.1)158 (37.4)399 (39.6)279 (41.5)120 (35.7) Dead (attributable to this cancer dx)819 (60.0)554 (58.9)265 (62.6)609 (60.4)393 (58.5)216 (64.3)Data are presented as n (%). ELNs, examined lymph nodes; M, metastasis; N, node; NLR, number of lymph node regions removed; PG, proximal gastrectomy; PSM, propensity score matching; T, tumor; TG, total gastrectomy. Baseline characteristics of patients with Siewert type II adenocarcinoma of esophagogastric junction receiving neoadjuvant therapy before and after PSM Data are presented as n (%). ELNs, examined lymph nodes; M, metastasis; N, node; NLR, number of lymph node regions removed; PG, proximal gastrectomy; PSM, propensity score matching; T, tumor; TG, total gastrectomy. Comparison of OS and CSS before and after PSM between PG and TG Survival analysis without PSM showed that the prognosis of patients in the PG group was significantly better than that in the TG group (P=0.02). Specifically, the 1-, 3-, and 5-year OS rates in the PG group were 85.8% [95% confidence interval (CI): 83.6–88.1%], 53.8% (95% CI: 50.7–57.1%), and 40.7% (95% CI: 37.7–44.0%), respectively, while the corresponding OS rates in the TG group were 80.3% (95% CI: 76.5–84.2%), 46.5% (95% CI: 41.9–51.5%), and 35.6% (95% CI: 31.3–40.5%), respectively. After PSM, the 1-, 3-, and 5-year OS rates in the PG group were 85.7% (95% CI: 83.1–88.4%), 53.6% (95% CI: 50.0–57.5%), and 41.5% (95% CI: 37.9–45.4%), respectively, and those in the TG group were 80.9% (95% CI: 76.7–85.2%), 46.5% (95% CI: 41.5–52.2%), and 35.7% (95% CI: 30.9–41.2%), respectively. There was still a statistically significant difference between the two groups (P=0.04) (Figure 1). Comparison of OS between PG and TG patients before (A) and after (B) PSM. The blue solid line represents the PG group, and the red solid line represents the TG group; the light blue and light red shaded areas correspond to the 95% CIs of the survival curves for the two groups, respectively. CI, confidence intervals; OS, overall survival; PG, proximal gastrectomy; PSM, propensity score matching; TG, total gastrectomy. Without PSM, the 1-, 3-, and 5-year CSS rates were 87.4% (95% CI: 85.3–89.6%) vs. 82.8% (95% CI: 79.2–86.5%), 57.0% (95% CI: 53.8–60.3%) vs. 50.1% (95% CI: 45.4–55.2%), and 45.4% (95% CI: 42.3–48.8%) vs. 39.9% (95% CI: 35.3–45.0%) in the PG group and TG group, respectively, with a statistically significant difference between the two groups (P=0.02). After PSM, the 1-, 3-, and 5-year CSS rates in the PG group were 86.7% (95% CI: 84.1–89.3%), 56.4% (95% CI: 52.7–60.3%), and 45.9% (95% CI: 42.2–49.9%), respectively, while the corresponding CSS rates in the TG group were 83.0% (95% CI: 79.0–87.1%), 48.8% (95% CI: 43.6–54.6%), and 39.2% (95% CI: 34.2–44.9%), respectively. A statistically significant difference between the two groups remained (P=0.02) (Figure 2). Comparison of CSS between PG and TG patients before (A) and after (B) PSM. The blue solid line represents the PG group, and the red solid line represents the TG group; the light blue and light red shaded areas correspond to the 95% CIs of the survival curves for the two groups, respectively. CI, confidence intervals; CSS, cancer-specific survival; PG, proximal gastrectomy; PSM, propensity score matching; TG, total gastrectomy. Multivariate analysis of the impact of PG and TG on patient survival outcomes In this study, multivariate Cox regression analysis was performed to investigate the impact of surgical approach on patient survival outcomes (Table 2). In the unadjusted model, with PG as the reference group, patients who underwent TG exhibited a significantly elevated risk of all-cause death [hazard ratio (HR) =1.18, 95% CI: 1.03–1.35, P=0.02] and cancer-specific death (HR =1.19, 95% CI: 1.03–1.37, P=0.02), indicating that the overall prognosis of TG patients was worse than that of PG patients. In Model II, which was adjusted for age, sex, race, and marital status, TG patients still had significantly inferior OS (HR =1.17, 95% CI: 1.02–1.34, P=0.02) and CSS (HR =1.19, 95% CI: 1.03–1.38, P=0.02) compared with PG patients, and the differences remained statistically significant. In the fully adjusted model, which further incorporated T stage, N stage, M stage, G stage, NLR, ELNs, rural-urban continuum, and median household income, TG was still an independent predictor of worse OS (HR =1.20, 95% CI: 1.05–1.38, P=0.009) and CSS (HR =1.22, 95% CI: 1.05–1.42, P=0.01), further confirming that the prognosis of TG patients was poorer than that of PG patients. Table 2Associations of surgical approach with OS and CSSCharacteristicsModel IModel IIModel IIIHR (95% CI)PHR (95% CI)PHR (95% CI)POS PGReferenceReferenceReference TG1.18 (1.03–1.35)0.021.17 (1.02–1.34)0.021.20 (1.05–1.38)0.009CSS PGReferenceReferenceReference TG1.19 (1.03–1.37)0.021.19 (1.03–1.38)0.021.22 (1.05–1.42)0.01Model I: crude model, unadjusted. Model II: adjusted for age, sex, race, and marital status. Model III: fully adjusted model, including age, sex, race, marital status, T stage, N stage, M stage, G stage, NLR, ELNs, rural-urban continuum and median household income. CI, confidence interval; CSS, cancer-specific survival; ELNs, examined lymph nodes; G, grade; HR, hazard ratio; M, metastasis; N, node; NLR, number of lymph node regions; OS, overall survival; PG, proximal gastrectomy; T, tumor; TG, total gastrectomy. Associations of surgical approach with OS and CSS Model I: crude model, unadjusted. Model II: adjusted for age, sex, race, and marital status. Model III: fully adjusted model, including age, sex, race, marital status, T stage, N stage, M stage, G stage, NLR, ELNs, rural-urban continuum and median household income. CI, confidence interval; CSS, cancer-specific survival; ELNs, examined lymph nodes; G, grade; HR, hazard ratio; M, metastasis; N, node; NLR, number of lymph node regions; OS, overall survival; PG, proximal gastrectomy; T, tumor; TG, total gastrectomy. Variable selection and prognostic model development for OS and CSS To develop and validate the prediction models, this study randomly divided patients in the PG group and TG group into training and validation sets at a ratio of 7:3, respectively. Corresponding prediction models were constructed separately for each group. The baseline and clinicopathological characteristics of patients in the PG and TG groups were presented in Table S1 and Table S2, respectively. In the training set of patients receiving PG, univariate Cox regression analysis was initially performed to screen prognostic variables. Results showed that age, sex, T stage, N stage, and M stage were significantly associated with OS (P<0.05). These statistically significant variables were then included in multivariate Cox regression analysis, which identified age, sex, T stage, N stage, and M stage as independent prognostic factors for OS (Table 3). Based on these findings, the above variables were ultimately incorporated to construct a nomogram for predicting OS (Figure 3A). For CSS, the same variable selection procedure was applied. Univariate and multivariate Cox regression analyses revealed that sex, T stage, N stage, and M stage were significantly associated with CSS (Table S3). These factors were subsequently used to establish a nomogram for CSS (Figure S2A). Table 3Univariate and multivariate analysis of OS in patients with Siewert type II esophagogastric junction adenocarcinoma undergoing PG after neoadjuvant therapyCharacteristicUnivariableMultivariableHR (95% CI)PHR (95% CI)PAge, years <65ReferenceReference ≥651.34 (1.12–1.61)0.0021.4 (1.16–1.68)<0.001Sex MaleReferenceReference Female0.74 (0.56–0.97)0.030.74 (0.56–0.98)0.03Race BlackReference White0.99 (0.55–1.80)0.98 Other0.71 (0.35–1.46)0.36Marital status UnmarriedReference Married0.86 (0.70–1.06)0.15T stage T1ReferenceReference T21.2 (0.86–1.67)0.281.09 (0.78–1.52)0.62 T31.24 (0.88–1.75)0.221.1 (0.77–1.57)0.59 T42.43 (1.43–4.12)<0.0012.17 (1.27–3.7)0.005N stage N0ReferenceReference N11.56 (1.27–1.93)<0.0011.52 (1.22–1.88)<0.001 N21.49 (1.05–2.11)0.031.46 (1.03–2.08)0.04 N32.69 (1.36–5.3)0.0042.2 (1.11–4.38)0.02M stage M0ReferenceReference M11.92 (1.30–2.83)0.0011.78 (1.20–2.65)0.004Grade IReference II0.97 (0.62–1.53)0.91 III1.17 (0.75–1.83)0.49 IV1.06 (0.47–2.39)0.89 Unknown0.61 (0.34–1.07)0.08Neoadjuvant therapy ChemotherapyReference Chemoradiotherapy1.05 (0.83–1.34)0.67Tumor size <2 cmReference ≥2, <4 cm1.24 (0.88–1.75)0.21 ≥4, <6 cm1.21 (0.85–1.71)0.29 >6 cm1.35 (0.94–1.95)0.11 Unknown1.16 (0.81–1.65)0.42NLR <3Reference ≥30.93 (0.64–1.35)0.71ELNs <15Reference ≥150.91 (0.76–1.09)0.33Rural-urban continuum <250,000 populationReference ≥250,000, <1 million population0.9 (0.63–1.3)0.59 ≥1 million population0.88 (0.63–1.24)0.47 Others1.04 (0.70–1.53)0.86Median household income <$70,000Reference ≥$70,0000.88 (0.72–1.07)0.20CI, confidence interval; ELNs, examined lymph nodes; HR, hazard ratio; M, metastasis; N, node; NLR, number of lymph node regions removed; OS, overall survival; PG, proximal gastrectomy; T, tumor. Univariate and multivariate analysis of OS in patients with Siewert type II esophagogastric junction adenocarcinoma undergoing PG after neoadjuvant therapy CI, confidence interval; ELNs, examined lymph nodes; HR, hazard ratio; M, metastasis; N, node; NLR, number of lymph node regions removed; OS, overall survival; PG, proximal gastrectomy; T, tumor. Nomograms for predicting 1-, 3-, 5-year OS in patients undergoing PG (A) and TG (B). ELNs, examined lymph nodes; M, metastasis; N, node; OS, overall survival; PG, proximal gastrectomy; T, tumor; TG, total gastrectomy. In the training set of patients undergoing TG, after screening by univariate and multivariate Cox regression analyses, age, N stage, M stage, and ELNs were identified as independent prognostic factors for OS (Table 4), based on which a prognostic nomogram was constructed (Figure 3B). For CSS, only N stage and M stage were independent prognostic factors (Table S4), and these two factors were used to construct a prognostic nomogram for CSS (Figure S2B). Table 4Univariate and multivariate analysis of OS in patients with Siewert type II esophagogastric junction adenocarcinoma undergoing TG after neoadjuvant therapyCharacteristicUnivariableMultivariableHR (95% CI)PHR (95% CI)PAge, years <65ReferenceReference ≥651.35 (1.03–1.76)0.031.33 (1.02–1.75)0.04Sex MaleReference Female1.22 (0.86–1.74)0.26Race BlackReference White1.53 (0.78–2.98)0.21 Other1.18 (0.53–2.65)0.69Marital status UnmarriedReference Married0.86 (0.64–1.15)0.31T stage T1ReferenceReference T21.62 (1.02–2.57)0.041.23 (0.76–2.00)0.39 T31.50 (0.91–2.48)0.111.04 (0.60–1.80)0.89 T41.65 (0.75–3.61)0.211.00 (0.44–2.25)0.99N stage N0ReferenceReference N11.64 (1.2–2.25)0.0021.51 (1.08–2.11)0.02 N22.99 (1.89–4.71)<0.0012.70 (1.65–4.42)<0.001 N33.38 (1.45–7.87)0.0053.37 (1.39–8.19)0.007M stage M0ReferenceReference M12.51 (1.52–4.15)<0.0011.91 (1.13–3.25)0.02Grade IReference II0.80 (0.40–1.6)0.53 III1.18 (0.60–2.32)0.63 IV0.32 (0.04–2.55)0.28 Unknown0.55 (0.22–1.36)0.20Neoadjuvant therapy ChemotherapyReference Chemoradiotherapy0.94 (0.70–1.25)0.66Tumor size <2 cmReferenceReference ≥2, <4 cm1.37 (0.82–2.29)0.231.31 (0.77–2.22)0.32 ≥4, <6 cm1.46 (0.85–2.49)0.171.41 (0.80–2.48)0.23 >6 cm1.83 (1.10–3.06)0.021.64 (0.96–2.80)0.07 Unknown1.49 (0.89–2.50)0.131.27 (0.75–2.17)0.37NLR <3Reference ≥30.69 (0.44–1.07)0.09ELNs <15Reference ≥150.76 (0.58–0.99)0.040.69 (0.52–0.92)0.01Rural-urban continuum <250,000 populationReference ≥250,000, <1 million population0.69 (0.38–1.25)0.22 ≥1 million population1.00 (0.60–1.66)0.99 Others1.14 (0.65–1.99)0.64Median household income <$70,000Reference ≥$70,0000.85 (0.64–1.13)0.26CI, confidence interval; ELNs, examined lymph nodes; HR, hazard ratio; M, metastasis; N, node; NLR, number of lymph node regions removed; OS, overall survival; T, tumor; TG, total gastrectomy. Univariate and multivariate analysis of OS in patients with Siewert type II esophagogastric junction adenocarcinoma undergoing TG after neoadjuvant therapy CI, confidence interval; ELNs, examined lymph nodes; HR, hazard ratio; M, metastasis; N, node; NLR, number of lymph node regions removed; OS, overall survival; T, tumor; TG, total gastrectomy. To comprehensively evaluate the prognostic predictive efficacy of the constructed nomogram, validation was performed in patients of the PG group from two aspects OS and CSS. The results showed that the C-index of the nomogram for OS prediction was 0.599 and 0.601 in the training set and internal test set, respectively. ROC analysis revealed that the area under the curve (AUC) values for 1-, 3-, and 5-year OS in the training set were 0.68, 0.64, and 0.62, respectively, while the corresponding values in the test set were 0.63, 0.62, and 0.66, suggesting that the model had good predictive ability. Calibration curves demonstrated a good agreement between the predicted survival rates and the actual observed results, and DCA also indicated that the model provided clinical net benefits (Figure 4). In terms of CSS prediction, the C-indexes in the training set and test set were 0.606 and 0.613, respectively, which were slightly better than those for OS, indicating that the nomogram also had good predictive efficacy for this endpoint. The relevant ROC curves, calibration curves, and DCA curves are shown in Figure S3. Evaluation of the nomogram for predicting 1-, 3-, and 5-year OS in patients undergoing PG. (A) ROC curves in the training set; (B) ROC curves in the validation set; (C-E) calibration curves of 1-, 3-, and 5-year OS in the training set; (F-H) calibration curves of 1-, 3-, and 5-year OS in the validation set; (I-K) DCA in the training set; (L-N) DCA in the validation set. AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis; OS, overall survival; PG, proximal gastrectomy; ROC, receiver operating characteristic. In patients of the TG group, the C-index of the nomogram for OS prediction was 0.625 in the training set and 0.621 in the internal test set, which was superior to that in the PG group, indicating better predictive consistency. ROC analysis showed that the AUC values for 1-, 3-, and 5-year OS in the training set were 0.70, 0.67, and 0.68, respectively, and the corresponding values in the test set were 0.69, 0.70, and 0.70. Among these, the AUC values for 3- and 5-year OS were both ≥0.70, indicating that the model had high accuracy in predicting long-term OS. The calibration curves showed good fit, and DCA analysis also supported its clinical application value (Figure 5). In CSS prediction, the C-indexes in the training set and test set were 0.613 and 0.629, respectively. The relevant ROC curves, calibration curves, and DCA curves are also shown in Figure S4. Evaluation of the nomogram for predicting 1-, 3-, and 5-year OS in patients undergoing TG. (A) ROC curves in the training set; (B) ROC curves in the validation set; (C-E) calibration curves of 1-, 3-, and 5-year OS in the training set; (F-H) calibration curves of 1-, 3-, and 5-year OS in the validation set; (I-K) DCA in the training set; (L-N) DCA in the validation set. AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis; OS, overall survival; ROC, receiver operating characteristic; TG, total gastrectomy. Discussion The optimal surgical resection extent for Siewert type II AEG remains controversial. Although TG has been widely used as the standard treatment for patients with AEG, there is no consensus on the prognostic superiority between PG and TG for the specific subtype of Siewert type II AEG (14-16). A study based on the SEER database showed no significant differences in prognosis between PG and TG in patients with Siewert type II AEG (17). Similarly, a systematic review and meta-analysis indicated comparable outcomes between the two surgical procedures in patients with Siewert type II/III AEG (16). Liao et al. reported that OS and CSS were generally similar between the PG and TG groups in matched cohorts. Further stratified analysis demonstrated that long-term survival was comparable between PG and TG in patients with tumor size ≤4 cm, whereas TG was associated with better OS and CSS than PG in patients with tumor size >4 cm (11). However, most of the above studies did not focus on patients with Siewert type II AEG who received neoadjuvant therapy, and their conclusions cannot be directly generalized to this specific population. In the present study, Kaplan-Meier survival analysis revealed that the PG group had significantly better OS and CSS than the TG group, both before and after matching. Multivariate Cox regression analysis further confirmed that TG was an independent risk factor for poor long-term prognosis compared with PG, indicating worse survival outcomes in patients undergoing TG. In terms of surgical characteristics, PG preserves part of the distal stomach and its digestive and storage functions, which helps reduce the risks of postoperative malnutrition, weight loss, and severe reflux, thereby facilitating postoperative recovery and longterm survival (18-20). Meanwhile, PG is associated with limited surgical trauma and mild disruption of the gastrointestinal anatomical structure (21). Postoperatively, it triggers a milder systemic inflammatory stress response, maintains cellular immune homeostasis, preserves endogenous antitumor immune surveillance, and attenuates the potential for tumor distant dissemination and abnormal proliferation of residual lesions (22,23). In contrast, although TG provides more extensive tumor resection, it is associated with greater surgical trauma and severe disruption of digestive function, which may lead to deteriorated nutritional status, impaired immune function, and ultimately unfavorable longterm survival (24,25). Severe surgical trauma can induce persistent systemic inflammation, suppress the activity of lymphocytes, natural killer (NK) cells and other immune cells, impair the body’s antitumor immune clearance capacity, and aggravate the immunosuppressive state (26,27). Meanwhile, the complete loss of gastric endocrine glands disrupts the homeostasis of gastrointestinal hormones, activates pro-tumor signaling pathways, and facilitates the proliferation, epithelial-mesenchymal transition, invasion and metastasis of residual cancer cells, thereby increasing the risk of postoperative tumor recurrence (28,29). Notably, all patients in this study were diagnosed with Siewert type II AEG and received neoadjuvant therapy. Neoadjuvant therapy can achieve tumor regression and downstaging, reduce tumor burden, and improve the R0 resection rate (30-32). Under these conditions, PG can achieve satisfactory radical resection, effectively compensating for its potential shortcomings in oncological safety. Meanwhile, the advantages of PG in preserving gastric function, reducing postoperative complications, and promoting rapid recovery are fully exerted, which ultimately translates into significant survival benefits. This study suggests that PG can achieve favorable local tumor control when combined with standardized lymph node dissection after tumor regression and clinical downstaging induced by neoadjuvant therapy (33). Patients with good tolerance to neoadjuvant therapy and marked tumor regression may be prioritized for PG, whereas those with a poor response to neoadjuvant therapy and more aggressive malignant features are recommended to undergo TG. This study provides evidence-based implications for the individualized selection of surgical strategies according to neoadjuvant treatment response. In conclusion, in the setting of effective neoadjuvant therapy, PG can achieve satisfactory oncological radicality while maximizing gastric function preservation and improving long‑term survival, thus presenting superior advantages over TG. The results of the present study suggest that for patients with obvious tumor regression after neoadjuvant therapy, extended resection is not always necessary, and appropriate selection of PG may achieve a better balance between oncological safety and survival benefits. Differential surgical procedures vary in resection scope, tumor biological characteristics and recurrence patterns, and distinct survival endpoints also differ in clinical implications and influencing factors. Therefore, this study constructed prognostic prediction models separately stratified by surgical approach and survival endpoint. The results demonstrated that tumor-node-metastasis (TNM) stage-related indicators were the common core prognostic factors across all models, which could collectively reflect the invasive and metastatic potential of tumors and represent the fundamental determinant of long-term survival in patients (34,35). These findings are consistent with the biological nature of AEG, further validating the central value of TNM staging in prognostic evaluation for AEG patients. Meanwhile, significant heterogeneity was observed in the predictive variables included in different models, suggesting distinct prognostic risk compositions among patients undergoing different surgical procedures. For patients who underwent PG, in addition to TNM stage-related indicators, gender and age were also incorporated into the model. This indicates that the prognosis of these patients is affected not only by tumor biological characteristics but also by individual baseline conditions. Under the premise of effective downstaging and reduced tumor burden following neoadjuvant therapy, PG can achieve satisfactory radical resection (36). In this context, individual patient differences exert a more pronounced influence on prognosis, whereas the weight of surgeryrelated factors is relatively lower (37,38). For patients who underwent TG, aside from TNM stage–related indicators, the ELNs emerged as an important predictive factor. This suggests that the prognosis of these patients depends not only on the invasive and metastatic potential of the tumor but also closely on the quality of lymph node dissection. Given that TG is associated with greater surgical trauma and more significant impacts on host immune and nutritional status, the adequacy of lymph node dissection plays a more critical role in prognostic assessment (39). These results indicate that the prognosis of patients with Siewert type II AEG after neoadjuvant therapy is centered on TNM stage as a common core factor, with distinct risk characteristics regulated by surgical approach. The coexistence of such commonalities and differences further confirms the rationality and necessity of constructing prognostic models stratified by surgical procedure, which not only captures the core rules of disease prognosis but also considers the clinical characteristics of different surgeries, providing a more reliable theoretical basis for clinical precise risk stratification, individualized surgical decision-making, and postoperative management. There are several limitations to this study. Clinically, TG is more frequently performed in patients with advanced tumor stages and complex conditions. Although PSM was applied to balance baseline clinicopathological characteristics, unmeasured residual confounding may still remain, and the influence of surgical selection bias on long-term survival cannot be fully eliminated. This study revealed that PG yields a better long-term prognosis than TG, and such survival benefits need further validation in large-sample, multicenter prospective studies and randomized controlled trials. In addition, the prognostic model constructed in this study has not undergone external validation. Future multicenter prospective studies are warranted to confirm our findings, refine the prognostic evaluation system, and provide high-level evidence for individualized management of patients with AEG following neoadjuvant therapy. Conclusions This study demonstrated that PG was associated with a more favorable prognosis than TG in patients with Siewert type II AEG after neoadjuvant therapy. For patients with obvious tumor regression following neoadjuvant therapy, extended resection may not always be necessary, and appropriate selection of PG can achieve a better balance between oncological safety and survival benefits. TNM stage-related indicators were identified as the common core prognostic factors across all models, while distinct prognostic factors were observed between different surgical procedures. The prognostic model stratified by surgical approach provides reliable evidence for precise risk stratification, individualized surgical decision-making, and postoperative management in these patients. As a single-center retrospective study, this research has certain limitations, and further multicenter and prospective studies are warranted for validation. Supplementary The article’s supplementary files as Acknowledgments None. Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Footnotes Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0833/rc Funding: This work was supported by the Science and Technology Program of the Joint Fund of Scientific Research for the Public Hospitals of the Inner Mongolia Academy of Medical Sciences (Nos. 2023GLLH0123 and 2024GLLH0384), the Inner Mongolia Medical University Joint Project (No. YKD2024LH021), the Beijing University Cancer Hospital Inner Mongolia Hospital “Qingmiao” Talent Plan (Nos. QM202325 and QM202312), Inner Mongolia Hospital of Peking University Cancer Hospital-Demonstration Project for the Reform and High-quality Development of Public Hospitals (Gastrointestinal Tumors + Thoracic Tumors) Construction of Clinical Key Specialties (No. 12024YNQN013), Natural Science Foundation of Inner Mongolia Autonomous Region (No. 2025LHMS08067), and Wu Jieping Medical Foundation Special Subsidy Fund for Clinical Research (No. 320.6750.2024-13-33). Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0833/coif). The authors have no conflicts of interest to declare. References Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0833/prf Supplementary Materials The article’s supplementary files as