Abstract
Background
Despite the rising global incidence of adenocarcinoma of the esophagogastric junction (AEG), treatment for Siewert type II AEG remains controversial, particularly regarding whether neoadjuvant therapy confers differential benefit according to stage. Locally advanced patients may benefit from neoadjuvant therapy, yet the benefit of additional adjuvant therapy remains unclear. This study therefore aimed to evaluate the stage-specific efficacy of neoadjuvant therapy and to explore the value of adjuvant therapy following neoadjuvant treatment in locally advanced disease.
Methods
Patients diagnosed with Siewert type II AEG were retrieved from the Surveillance, Epidemiology, and End Results (SEER) database. Survival differences among patients were assessed using the Kaplan-Meier method and log‑rank test. To minimize baseline imbalances, propensity score matching (PSM) was performed, and multivariable Cox regression analyses were conducted before and after PSM to adjust for potential confounders. Univariate and multivariate Cox regression analyses were employed to identify independent prognostic factors. A nomogram predicting overall survival (OS) was constructed based on the variables selected by Cox regression. The predictive performance of the nomogram was evaluated by the concordance index (C‑index), receiver operating characteristic (ROC) curves with area under the curve (AUC) values, calibration plots, and decision curve analysis (DCA).
Results
A total of 1,847 patients were enrolled. In the overall cohort, no significant OS difference was observed between surgery alone (SA) and neoadjuvant-based therapy combined with surgery (NS) groups after PSM (P=0.14). Stage‑stratified analysis showed that neoadjuvant therapy significantly improved OS in locally advanced patients (P<0.001), but not in early‑stage patients (P=0.06). Among locally advanced patients receiving neoadjuvant therapy, adding adjuvant therapy provided no additional OS benefit (P=0.87). Prognostic nomograms were constructed for early‑stage and locally advanced patients. For the early‑stage model, the C‑index was 0.696 in the training set and 0.669 in the testing set, with AUC values for 1‑, 3‑, and 5‑year OS of 0.653, 0.730, and 0.748 in the training set, and 0.698, 0.702, and 0.692 in the testing set. For the locally advanced model, the C‑index was 0.671 and 0.635 in the training and testing sets, respectively, with AUC values of 0.741, 0.723, and 0.725 in the training set, and 0.693, 0.684, and 0.690 in the testing set. Calibration and DCA confirmed satisfactory predictive performance and clinical utility for both models.
Conclusions
Neoadjuvant therapy, predominantly chemoradiotherapy, significantly improved OS in locally advanced Siewert type II AEG but not in early-stage disease. In locally advanced patients, adding adjuvant therapy after neoadjuvant therapy did not provide additional OS benefit. These findings suggest that treatment strategies may differ by stage, though this observation warrants further validation through prospective studies.
Keywords: Siewert type II adenocarcinoma of the esophagogastric junction (Siewert type II AEG), neoadjuvant therapy, early-stage, locally advanced
Highlight box.
Key findings
• This study reveals that neoadjuvant therapy, predominantly chemoradiotherapy, improves overall survival (OS) exclusively in locally advanced Siewert type II adenocarcinoma of the esophagogastric junction (AEG), with no benefit in early-stage disease. Adjuvant therapy after neoadjuvant therapy did not confer a significant OS benefit in locally advanced patients.
What is known and what is new?
• It is well established that no unified treatment consensus has been reached for Siewert type II AEG. Although neoadjuvant therapy or perioperative therapy is currently recommended for AEG based on trials involving mixed populations, stage-specific and subtype-specific evidence remains insufficient.
• This study indicates that survival benefits of neoadjuvant therapy, predominantly neoadjuvant chemoradiotherapy, are confined to locally advanced Siewert type II AEG. Our findings indicate that adjuvant therapy following neoadjuvant treatment is not associated with a statistically significant survival benefit, suggesting that the use of perioperative regimens for this subtype remains to be further validated. Furthermore, we developed prognostic nomograms separately for early-stage and locally advanced patients using OS as the endpoint.
What is the implication, and what should change now?
• These findings suggest that early-stage and locally advanced stage may serve as important factors influencing the therapeutic effect of neoadjuvant therapy in Siewert type II AEG. Among locally advanced patients, adjuvant therapy following neoadjuvant therapy showed no statistically significant survival benefit compared with neoadjuvant therapy alone. Further prospective studies are required to corroborate these findings.
Introduction
The global incidence of adenocarcinoma of the esophagogastric junction (AEG) has risen markedly over recent decades, contributing to an increasing global health burden (1,2). AEG is defined as adenocarcinoma with the tumor epicenter located within 5 cm proximal or distal to the esophagogastric junction (EGJ) (3). According to the Siewert classification, it is further subdivided into three types. Among these, type II, with its epicenter situated between 1 cm above and 2 cm below the EGJ, is widely recognized as the true carcinoma of the cardia (4,5).
Owing to its distinctive anatomical location, biological behavior, and lymphatic drainage pattern, which all differ markedly from those of either esophageal or gastric adenocarcinoma, and because it also shares certain complex features of both, Siewert type II AEG has long lacked a unified consensus on treatment strategies (6). Currently, surgery remains the mainstay of treatment for this subtype (7). However, the indications for and survival benefit of neoadjuvant therapy compared with upfront surgery remain a matter of debate. The CROSS trial has demonstrated that neoadjuvant chemoradiotherapy significantly improves overall survival (OS) in patients with AEG (6). Nevertheless, evidence specific to Siewert type II AEG remains limited. To date, few studies have evaluated the benefit of neoadjuvant therapy in this subtype according to clinical stage. Given that early-stage and locally advanced patients differ considerably in clinicopathological features and prognosis, stage‑specific treatment strategies may be warranted (8). Multiple randomized controlled trials have demonstrated that perioperative chemotherapy improves survival in patients with AEG (9,10). However, these studies evaluated neoadjuvant and adjuvant therapy as a combined regimen, and the independent contribution and necessity of adjuvant therapy remain unclear, with few studies having addressed this question.
Current evidence remains insufficient to guide neoadjuvant therapy decisions for Siewert type II AEG, particularly regarding the individualized benefit across different clinical stages. To address this gap, this study aims to compare the prognostic outcomes of neoadjuvant-based therapy combined with surgery (NS) versus surgery alone (SA) in Siewert type II AEG using a large-scale real-world cohort, to determine whether this treatment effect varies by clinical stage, and to evaluate the additional survival benefit of adjuvant therapy following neoadjuvant therapy. Ultimately, our study aims to provide evidence to inform individualized treatment strategies for patients with Siewert type II AEG. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-2069/rc).
Methods
Study population
The present study extracted data on Siewert type II AEG from the Surveillance, Epidemiology, and End Results (SEER) database (https://seer.cancer.gov/) using SEER*Stat (version 8.4.3). Inclusion criteria were defined as follows: (I) year of diagnosis between 2004 and 2015; (II) classification as Siewert type II AEG, ascertained by the “Primary Site” code of 16.0 (cardia) and the “Collaborative Stage (CS) site-specific factor 25” coded as 982 (esophagus, gastroesophageal junction); (III) unique primary malignancy; (IV) histologically confirmed AEG according to ICD-O-3 codes 8140–8147, 8160–8162, 8180–8221, 8245, 8250–8507, 8514–8551, 8571–8574, 8576, and 8940–8941; and (V) receipt of either radical SA or NS. Exclusion criteria comprised: (I) presence of distant metastases (M); (II) fewer than one examined lymph node (ELN); (III) missing essential clinicopathologic information; and (IV) survival of one month or less. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.
For patients diagnosed with Siewert type II AEG between 2004 and 2015, staging was performed within the CS framework. Pathological stage was prioritized for patients who underwent upfront radical SA, whereas staging information was predominantly derived from clinical staging for those receiving neoadjuvant therapy followed by surgical resection. The eighth edition of the American Joint Committee on Cancer (AJCC) TNM classification for gastric adenocarcinoma was used to stratify the overall patient cohort into early stage (T1–2N0M0) and locally advanced stage (T3–4 any N M0, or any TN+M0). The inclusion and exclusion process is shown in Figure 1.
Figure 1.
Flowchart of patient inclusion. AEG, adenocarcinoma of the esophagogastric junction; M, metastasis; N, node; SEER, Surveillance, Epidemiology, and End Results; T, tumor.
For treatment comparison, enrolled patients were divided into two groups: SA and NS. The NS group was further split into neoadjuvant only (NO) and perioperative therapy (PT). Specifically, neoadjuvant treatments included radiotherapy, chemotherapy or chemoradiotherapy.
This study included the following variables: age, sex, race, marital status, tumor (T) stage, node (N) stage, grade, tumor size, surgical type, and ELNs. The endpoint was OS, which was defined as the time from diagnosis to all-cause death or the date of last follow-up.
Statistical analysis
Categorical variables were presented as frequencies and percentages. Continuous variables were presented as means with standard deviations (SDs). Propensity score matching (PSM) with a 1:1 ratio was performed to balance covariates including age, sex, race, and tumor characteristics, using a caliper width of 0.1 on the logit scale. Survival outcomes before and after PSM were compared using Kaplan-Meier curves with the log-rank test, and corresponding survival curves were plotted. To address residual confounding indicated by standardized mean differences (SMDs) >0.1 for certain covariates after PSM, we further constructed multivariable Cox regression models before and after matching, specifically targeting the subpopulation with such imbalances. Three models were fitted: Model I was unadjusted; Model II adjusted for age, sex, race, and marital status; and Model III additionally adjusted for T stage, N stage, grade, tumor size, surgical procedure, and ELN. Patients with early-stage and locally advanced disease were randomly divided into a training set and a testing set at a 7:3 ratio. Univariate Cox regression was first conducted in the training set to identify variables associated with OS. Variables with P<0.05 in the univariate analysis were entered into the multivariate Cox regression model. Independent prognostic factors with P<0.05 in the multivariate analysis were selected to construct the final OS prediction model. The discriminative performance of the model was assessed using the concordance index (C-index) and the area under the curve (AUC). Calibration curves were plotted to evaluate the agreement between predicted and observed outcomes. Decision curve analysis (DCA) was performed to verify the clinical utility of the nomogram. All analyses were performed using R software (version 4.2.0). A two-sided P<0.05 was considered statistically significant.
Results
Participant characteristics
A total of 1,847 eligible patients with Siewert type II AEG were enrolled from the SEER database, including 788 patients in the SA group and 1,059 patients in the NS group. The mean age of all participants was 63.8 years. Most patients were male (1,499/18,47, 81.2%), white (1,621/1,847, 87.8%) and married (1,333/1,847, 72.2%). T3 stage (738/1,847, 40.0%), N0 stage (796/1,847, 43.1%), grade III (956/1,847, 51.8%) and tumor size of 2–4 cm (536/1,847, 29.0%) were dominant. In the NS group, chemoradiotherapy was the predominant neoadjuvant regimen (852/1,059, 80.5%). In the PT group, chemotherapy was the predominant adjuvant regimen (115/176, 65.3%), followed by radiotherapy (32/176, 18.2%) and chemoradiotherapy (29/176, 16.5%). Detailed baseline characteristics are shown in Table 1.
Table 1. Baseline characteristics of patients with Siewert type II AEG.
| Characteristics | Total (n=1,847) | SA (n=788) | NS (n=1,059) | P |
|---|---|---|---|---|
| Age, years | 63.8±11.3 | 67.4±11.5 | 61.1±10.3 | <0.001 |
| Sex | <0.001 | |||
| Male | 1,499 (81.2) | 594 (75.4) | 905 (85.5) | |
| Female | 348 (18.8) | 194 (24.6) | 154 (14.5) | |
| Race | 0.07 | |||
| White | 1,621 (87.8) | 676 (85.8) | 945 (89.2) | |
| Black | 80 (4.3) | 42 (5.3) | 38 (3.6) | |
| Other | 146 (7.9) | 70 (8.9) | 76 (7.2) | |
| Marital status | 0.02 | |||
| Unmarried | 514 (27.8) | 242 (30.7) | 272 (25.7) | |
| Married | 1,333 (72.2) | 546 (69.3) | 787 (74.3) | |
| T stage | <0.001 | |||
| T1 | 463 (25.1) | 362 (45.9) | 101 (9.5) | |
| T2 | 563 (30.5) | 283 (35.9) | 280 (26.4) | |
| T3 | 738 (40.0) | 118 (15.0) | 620 (58.5) | |
| T4 | 83 (4.5) | 25 (3.2) | 58 (5.5) | |
| N stage | <0.001 | |||
| N0 | 796 (43.1) | 492 (62.4) | 304 (28.7) | |
| N1 | 696 (37.7) | 186 (23.6) | 510 (48.2) | |
| N2 | 253 (13.7) | 72 (9.1) | 181 (17.1) | |
| N3 | 102 (5.5) | 38 (4.8) | 64 (6.0) | |
| Grade | <0.001 | |||
| I | 126 (6.8) | 78 (9.9) | 48 (4.5) | |
| II | 630 (34.1) | 290 (36.8) | 340 (32.1) | |
| III | 956 (51.8) | 384 (48.7) | 572 (54.0) | |
| IV | 34 (1.8) | 12 (1.5) | 22 (2.1) | |
| Unknown | 101 (5.5) | 24 (3.0) | 77 (7.3) | |
| Tumor size | <0.001 | |||
| <2 cm | 283 (15.3) | 180 (22.8) | 103 (9.7) | |
| ≥2, <4 cm | 536 (29.0) | 262 (33.2) | 274 (25.9) | |
| ≥4, <6 cm | 400 (21.7) | 150 (19.0) | 250 (23.6) | |
| ≥6 cm | 309 (16.7) | 119 (15.1) | 190 (17.9) | |
| Unknown | 319 (17.3) | 77 (9.8) | 242 (22.9) | |
| Surgical type | 0.01 | |||
| TG | 522 (28.3) | 198 (25.1) | 324 (30.6) | |
| PG | 1,325 (71.7) | 590 (74.9) | 735 (69.4) | |
| ELN | 16.8±10.9 | 15.8±10.6 | 17.5±11.2 | 0.001 |
| Neoadjuvant regimen | – | |||
| Radiotherapy | 3 (0.2) | – | 3 (0.3) | – |
| Chemotherapy | 204 (11.0) | – | 204 (19.3) | – |
| Chemoradiotherapy | 852 (46.1) | – | 852 (80.5) | – |
Data are presented as n (%) or mean ± standard deviation. AEG, adenocarcinoma of the esophagogastric junction; ELN, examined lymph nodes; N, node; NS, neoadjuvant-based therapy combined with surgery; PG, proximal gastrectomy; SA, surgery alone; TG, total gastrectomy; T, tumor.
PSM for confounder balancing
PSM was adopted to balance baseline confounders and mitigate selection bias. Matching was performed between SA and NS groups in the overall cohort, early-stage and locally advanced patients, and between NO and PT subgroups within the NS group. Baseline information before and after matching and SMD were provided to illustrate covariate distribution status (Tables S1-S4, Figures S1-S4).
Comparison of OS between SA and NS groups in the overall cohort
Before PSM, Kaplan-Meier survival analysis was performed on the overall population, showing no statistically significant difference in OS between the SA and NS groups (P=0.15). The 1-, 3-, and 5-year OS rates in the SA group were 73.6% [95% confidence interval (CI): 70.5–76.7%], 53.3% (95% CI: 49.9–56.9%), and 44.9% (95% CI: 41.5–48.5%), respectively. The corresponding rates in the NS group were 83.8% (95% CI: 81.6–86.1%), 49.0% (95% CI: 46.1–52.1%), and 37.0% (95% CI: 34.2–40.1%) (Figure 2A).
Figure 2.
Comparison of OS between the SA and the NS groups in the overall population before (A) and after (B) PSM. The blue solid line represents the SA group, and the red solid line represents the NS 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 interval; OS, overall survival; NS, neoadjuvant-based therapy combined with surgery; PSM, propensity score matching; SA, surgery alone.
After PSM, the OS between the two groups remained statistically insignificant (P=0.14). In the matched SA group, the 1-, 3-, and 5-year OS rates were 68.9% (95% CI: 64.5–73.6%), 45.5% (95% CI: 40.8–50.7%), and 36.5% (95% CI: 32.0–41.6%), respectively. In the matched NS group, the corresponding rates were 84.8% (95% CI: 81.4–88.4%), 52.7% (95% CI: 48.0–57.8%), and 40.2% (95% CI: 35.7–45.4%) (Figure 2B).
Comparison of OS between SA and NS groups in early-stage patients
Before PSM, Kaplan-Meier survival analysis showed that early-stage patients in the SA group had significantly better OS than those in the NS group (P=0.002). The 1-, 3- and 5-year OS rates were 87.3% (95% CI: 84.3–90.5%), 75.7% (95% CI: 71.9–79.8%) and 66.7% (95% CI: 62.5–71.3%) in the SA group, and 88.4% (95% CI: 83.4–93.8%), 66.7% (95% CI: 59.5–74.7%) and 55.1% (95% CI: 47.6–63.8%) in the NS group (Figure 3A).
Figure 3.
Comparison of OS between the SA and the NS groups in early-stage patients before (A) and after (B) PSM. The blue solid line represents the SA group, and the red solid line represents the NS 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 interval; OS, overall survival; NS, neoadjuvant-based therapy combined with surgery; PSM, propensity score matching; SA, surgery alone.
After PSM, the difference in OS rate between the SA group and NS group was no longer statistically significant among early-stage patients (P=0.06). The 1-, 3- and 5-year OS rates were 90.2% (95% CI: 85.1–95.6%), 72.2% (95% CI: 64.7–80.6%) and 59.1% (95% CI: 51.0–68.5%) in the SA group, while those in the NS group were 88.6% (95% CI: 83.2–94.4%), 65.9% (95% CI: 58.0–74.8%) and 52.8% (95% CI: 44.7–62.4%) (Figure 3B).
Table S5 presents the results of multivariable Cox regression analyses before and after PSM in patients with early-stage Siewert type II AEG. Before matching, NS group was associated with significantly worse OS in Model III after full adjustment for confounders [hazard ratio (HR) =1.31, 95% CI: 1.01–1.70, P=0.04]. However, this association was no longer statistically significant after PSM (HR =1.31, 95% CI: 0.94–1.83, P=0.11).
Comparison of OS between SA and NS groups in locally advanced patients
Kaplan-Meier survival analysis revealed that locally advanced patients in the NS group had markedly better OS than those in the SA group before PSM (P<0.001). The 1-year OS rate was 83.0% (95% CI: 80.6–85.5%) in the NS group versus 55.1% (95% CI: 50.0–60.7%) in the SA group. The 3- and 5-year OS rates were 46.1% (95% CI: 43.0–49.5%) versus 23.2% (95% CI: 19.1–28.2%) and 34.1% (95% CI: 31.1–37.3%) versus 15.6% (95% CI: 12.2–20.1%) (Figure 4A).
Figure 4.
Comparison of OS between the SA and the NS groups in locally advanced patients before (A) and after (B) PSM. The blue solid line represents the SA group, and the red solid line represents the NS 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 interval; OS, overall survival; NS, neoadjuvant-based therapy combined with surgery; PSM, propensity score matching; SA, surgery alone.
After PSM, the survival advantage of the NS group over the SA group remained statistically significant among locally advanced patients (P<0.001). The 1-year OS rate was 80.4% (95% CI: 75.4–85.6%) in the NS group and 57.7% (95% CI: 51.7–64.4%) in the SA group. The 3- and 5-year OS rates were 42.3% (95% CI: 36.4–49.1%) versus 25.7% (95% CI: 20.6–31.9%) and 31.2% (95% CI: 25.8–37.7%) versus 17.9% (95% CI: 13.6–23.6%) (Figure 4B).
Table S6 presents the results of multivariable Cox regression analyses before and after PSM in patients with locally advanced Siewert type II AEG. In Model III, NS group was consistently associated with significantly improved OS both before (HR =0.53, 95% CI: 0.45–0.62, P<0.001) and after matching (HR =0.60, 95% CI: 0.48–0.75, P<0.001), indicating a robust survival benefit of neoadjuvant therapy in these patients.
Comparison of OS between NO and PT subgroups in locally advanced patients receiving NS therapy
Kaplan-Meier analysis of OS among locally advanced patients receiving neoadjuvant therapy showed no significant difference between the NO and PT groups before PSM (P=0.46). The 1-year OS rate was 84.3% (95% CI: 81.7–87.0%) in the NO group and 86.3% (95% CI: 81.4–91.6%) in the PT group. The 3-year OS rates were 47.4% (95% CI: 43.9–51.2%) versus 40.7% (95% CI: 34.0–48.7%), and the 5-year OS rates were 35.0% (95% CI: 31.7–38.6%) versus 30.2% (95% CI: 24.0–37.9%) (Figure 5A).
Figure 5.
Comparison of OS between the NO and the PT groups in locally advanced patients before (A) and after (B) PSM. The blue solid line represents the NO group, and the red solid line represents the PT 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 interval; OS, overall survival; NO, neoadjuvant only; PT, perioperative therapy; PSM, propensity score matching.
After PSM, no statistically significant difference in OS was observed between the two groups (P=0.87). The 1-, 3-, and 5-year OS rates in the NO group were 82.6% (95% CI: 77.1–88.6%), 44.7% (95% CI: 37.8–53.0%), and 32.5% (95% CI: 26.1–40.5%), compared with 87.5% (95% CI: 82.6–92.6%), 41.5% (95% CI: 34.6–49.7%), and 30.4% (95% CI: 24.1–38.3%) in the PT group (Figure 5B).
Table S7 presents the results of multivariable Cox regression analyses before and after PSM in patients with locally advanced Siewert type II AEG receiving neoadjuvant therapy, comparing PT group with the NO group. In Model III, PT showed no significant OS benefit over NO either before (HR =1.00, 95% CI: 0.82–1.22, P=0.97) or after matching (HR =0.97, 95% CI: 0.75–1.26, P=0.84).
Prognostic model development and evaluation for early-stage and locally advanced patients
This study established prognostic models containing treatment modalities for early-stage and locally advanced patients separately. Patients in both subgroups were randomly divided into training and testing sets at a 7:3 ratio. Baseline characteristics of each group are shown in Tables S8,S9.
Based on training set data, univariate Cox regression analysis was performed on all clinical variables of early-stage patients. Four variables, including age, T stage, tumor size and treatment modality, with P<0.05, were enrolled into multivariate Cox regression analysis (Table 2). The results revealed that all indicators except treatment modality were independent prognostic factors for OS (P<0.05). Given that treatment modality was the core research focus of this study, it was still retained. These variables were finally used to construct the prognostic prediction model for early-stage patients (Figure 6A).
Table 2. Univariable and multivariable Cox regression analysis of OS in patients with early-stage Siewert type II AEG.
| Characteristic | Univariable | Multivariable | ||
|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | |
| Age, years | 1.06 (1.05–1.07) | <0.001 | 1.06 (1.04–1.07) | <0.001 |
| Sex | ||||
| Male | Reference | |||
| Female | 0.90 (0.66–1.22) | 0.48 | ||
| Race | ||||
| White | Reference | |||
| Black | 0.84 (0.37–1.89) | 0.67 | ||
| Other | 0.88 (0.55–1.43) | 0.61 | ||
| Marital status | ||||
| Unmarried | Reference | |||
| Married | 0.79 (0.59–1.05) | 0.10 | ||
| T stage | ||||
| T1 | Reference | Reference | ||
| T2 | 2.16 (1.67–2.79) | <0.001 | 1.66 (1.25–2.22) | 0.001 |
| Grade | ||||
| I | Reference | |||
| II | 0.87 (0.59–1.27) | 0.47 | ||
| III | 1.22 (0.83–1.79) | 0.30 | ||
| IV | 1.07 (0.26–4.46) | 0.92 | ||
| Unknown | 0.70 (0.34–1.45) | 0.34 | ||
| Tumor size | ||||
| <2 cm | Reference | Reference | ||
| ≥2, <4 cm | 1.52 (1.10–2.12) | 0.01 | 1.32 (0.95–1.85) | 0.10 |
| ≥4, <6 cm | 2.22 (1.45–3.38) | <0.001 | 1.51 (0.98–2.35) | 0.06 |
| ≥6 cm | 4.67 (2.90–7.52) | <0.001 | 2.40 (1.44–3.99) | 0.001 |
| Unknown | 1.15 (0.77–1.73) | 0.49 | 1.25 (0.82–1.88) | 0.30 |
| Surgical type | ||||
| TG | Reference | |||
| PG | 1.09 (0.80–1.47) | 0.59 | ||
| Treatment modality | ||||
| SA | Reference | Reference | ||
| NS | 1.41 (1.06–1.87) | 0.02 | 1.19 (0.88–1.61) | 0.26 |
| ELN | 0.99 (0.98–1.01) | 0.39 | ||
AEG, adenocarcinoma of the esophagogastric junction; CI, confidence interval; ELN, examined lymph nodes; HR, hazard ratio; NS, neoadjuvant-based therapy combined with surgery; OS, overall survival; PG, proximal gastrectomy; SA, surgery alone; TG, total gastrectomy; T, tumor.
Figure 6.
Nomograms for predicting 1-, 3-, and 5-year OS in early-stage (A) and locally advanced (B) patients. ELN, examined lymph nodes; N, node; NS, neoadjuvant-based therapy combined with surgery; OS, overall survival; SA, surgery alone; T, tumor.
For locally advanced patients, univariate Cox regression analysis was conducted in the training set. Seven variables, including age, marital status, N stage, grade, tumor size, treatment modality and ELN, with P<0.05, were included in multivariate Cox regression analysis (Table 3). The results indicated that the above screened variables, except tumor size, were independent prognostic factors for OS (P<0.05). These effective variables were subsequently used to establish the prognostic prediction model for locally advanced patients (Figure 6B).
Table 3. Univariable and multivariable Cox regression analysis of OS in patients with locally advanced Siewert type II AEG.
| Characteristic | Univariable | Multivariable | ||
|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | |
| Age, years | 1.02 (1.01–1.03) | <0.001 | 1.01 (1.01–1.02) | <0.001 |
| Sex | ||||
| Male | Reference | |||
| Female | 0.93 (0.77–1.14) | 0.50 | ||
| Race | ||||
| White | Reference | |||
| Black | 0.87 (0.60–1.26) | 0.45 | ||
| Other | 0.84 (0.64–1.11) | 0.22 | ||
| Marital status | ||||
| Unmarried | Reference | Reference | ||
| Married | 0.76 (0.64–0.89) | <0.001 | 0.75 (0.64–0.88) | 0.001 |
| T stage | ||||
| T1 | Reference | |||
| T2 | 1.07 (0.77–1.49) | 0.68 | ||
| T3 | 0.77 (0.56–1.06) | 0.11 | ||
| T4 | 1.29 (0.86–1.94) | 0.22 | ||
| N stage | ||||
| N0 | Reference | Reference | ||
| N1 | 1.56 (1.24–1.96) | <0.001 | 1.56 (1.24–1.97) | <0.001 |
| N2 | 2.18 (1.69–2.82) | <0.001 | 2.55 (1.95–3.33) | <0.001 |
| N3 | 3.09 (2.24–4.26) | <0.001 | 3.70 (2.65–5.18) | <0.001 |
| Grade | ||||
| I | Reference | Reference | ||
| II | 1.05 (0.71–1.57) | 0.80 | 1.11 (0.74–1.65) | 0.62 |
| III | 1.56 (1.07–2.29) | 0.02 | 1.62 (1.10–2.38) | 0.01 |
| IV | 1.09 (0.55–2.15) | 0.80 | 1.12 (0.57–2.22) | 0.74 |
| Unknown | 0.79 (0.48–1.31) | 0.36 | 1.11 (0.66–1.84) | 0.70 |
| Tumor size | ||||
| <2 cm | Reference | Reference | ||
| ≥2, <4 cm | 1.22 (0.87–1.71) | 0.26 | 1.16 (0.82–1.63) | 0.41 |
| ≥4, <6 cm | 1.22 (0.87–1.72) | 0.25 | 1.20 (0.85–1.69) | 0.31 |
| ≥6 cm | 1.46 (1.03–2.06) | 0.03 | 1.30 (0.91–1.85) | 0.15 |
| Unknown | 1.12 (0.78–1.60) | 0.54 | 1.28 (0.89–1.83) | 0.19 |
| Surgical type | ||||
| TG | Reference | |||
| PG | 0.93 (0.79–1.10) | 0.40 | ||
| Treatment modality | ||||
| SA | Reference | Reference | ||
| NS | 0.48 (0.41–0.57) | <0.001 | 0.54 (0.45–0.65) | <0.001 |
| ELN | 0.99 (0.98–1.00) | 0.002 | 0.98 (0.97–0.98) | <0.001 |
AEG, adenocarcinoma of the esophagogastric junction; CI, confidence interval; ELN, examined lymph nodes; HR, hazard ratio; N, node; NS, neoadjuvant-based therapy combined with surgery; OS, overall survival; PG, proximal gastrectomy; SA, surgery alone; TG, total gastrectomy; T, tumor.
The C-index of the prognostic model for early-stage patients was 0.696 in the training set and 0.669 in the testing set. The AUC values for predicting 1-, 3-, and 5-year OS were 0.653, 0.730, and 0.748 in the training set, and 0.698, 0.702, and 0.692 in the testing set (Figure 7A,7B). Calibration curves showed good agreement between predicted and observed 1-, 3-, and 5-year OS probabilities in both the training and testing sets (Figure 7C-7H). DCA confirmed that the model provided stable net benefit across a wide range of thresholds, indicating favorable clinical utility (Figure 7I-7N).
Figure 7.
Evaluation of the nomogram for predicting 1-, 3-, and 5-year OS in early-stage patients. (A) ROC curves in the training set; (B) ROC curves in the testing 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 testing set; (I-K) DCA in the training set; (L-N) DCA in the testing set. AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis; OS, overall survival; M, metastasis; N, node; ROC, receiver operating characteristic; T, tumor.
The prognostic model for locally advanced patients achieved a C-index of 0.671 in the training set and 0.635 in the testing set. AUC values for 1-, 3-, and 5-year OS prediction were 0.741, 0.723, and 0.725 in the training cohort, and 0.693, 0.684, and 0.690 in the testing cohort, respectively (Figure 8A,8B). Calibration curves demonstrated favorable concordance between predicted and observed survival probabilities across all three time points in both cohorts (Figure 8C-8H). DCA further supported the model’s clinical applicability, showing consistent net benefit over a broad range of threshold probabilities (Figure 8I-8N).
Figure 8.
Evaluation of the nomogram for predicting 1-, 3-, and 5-year OS in locally advanced patients. (A) ROC curves in the training set; (B) ROC curves in the testing 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 testing set; (I-K) DCA in the training set; (L-N) DCA in the testing set. AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis; OS, overall survival; M, metastasis; N, node; ROC, receiver operating characteristic; T, tumor.
Discussion
The optimal therapeutic strategy for Siewert type II AEG has long been a subject of debate. Although neoadjuvant chemoradiotherapy or perioperative chemotherapy is currently recommended as a treatment option for AEG, the evidence specifically targeting this subtype remains limited (11,12). Importantly, it remains unclear whether the association between neoadjuvant therapy and patient prognosis varies across clinical stages (13). Furthermore, for patients who receive neoadjuvant therapy, whether the addition of adjuvant therapy is associated with further prognostic improvement is equally unknown (14). To address these uncertainties, the present study was designed to evaluate the stage-specific prognostic impact of neoadjuvant therapy, which predominantly consisted of chemoradiotherapy in this cohort, in patients with Siewert type II AEG. Additionally, this study assessed the prognostic value of adjuvant therapy as part of the perioperative regimen in locally advanced patients.
In our study, no significant survival difference was observed between the NS and SA groups in the overall cohort either before (P=0.15) or after (P=0.14) PSM, with 5-year OS rates of 36.5% and 40.2% in the matched NS and SA groups, respectively. This finding contrasts with the CROSS trial, which demonstrated that neoadjuvant chemoradiotherapy significantly improved OS in a mixed population of esophageal and esophagogastric-junction cancers (15). This discrepancy may be explained by the possibility that the treatment effect of neoadjuvant therapy varies across clinical stages, and that a pooled analysis of all stages may not adequately reflect the heterogeneity of treatment effects across stage subgroups. Indeed, Zhou et al. reported that neoadjuvant radiotherapy was associated with improved survival in locally advanced Siewert II AEG patients but not in early-stage patients (16). Given this consideration, we performed stage-stratified analyses to investigate whether the prognostic association of neoadjuvant therapy differs by clinical stage in Siewert type II AEG.
When stratified by clinical stage, the direction of survival benefit was completely opposite between early-stage and locally advanced patients. In early-stage patients, the SA group was associated with a significant OS advantage over the NS group before PSM (P=0.002), but this difference lost statistical significance after PSM (P=0.06). In contrast, among locally advanced patients, the NS group consistently correlated with significantly better OS than the SA group both before (P<0.001) and after (P<0.001) PSM, with matched 5-year OS rates of 31.2% vs. 17.9%.
The apparent survival benefit of the SA group in early-stage patients before PSM likely reflects selection bias and inherent limitations of clinical staging. Clinicians may have preferentially assigned younger patients with better performance status, fewer comorbidities, and lower tumor burden to the SA group, while referring those with suspicious or high-risk features to the NS group (17,18). Additionally, some patients in the NS group might have experienced disease progression, significant treatment-related toxicity, or deterioration in performance status during preoperative therapy, ultimately precluding curative resection, a well-recognized drawback of neoadjuvant approaches that could adversely affect survival outcomes in this subgroup (19). However, the disappearance of this survival advantage after PSM suggests that the apparent benefit of the SA group in early-stage disease is largely driven by confounding factors rather than a genuine therapeutic effect. Our results are consistent with an analysis of cT2N0 esophageal and gastroesophageal junction adenocarcinoma that found no survival benefit with neoadjuvant therapy versus SA (HR =1.132, P=0.11) (20). Another study of clinical stage IB gastric cancer reported comparable 5-year OS between the neoadjuvant therapy and SA groups after PSM (58.8% vs. 50.3%, P=0.51) (21). Collectively, these findings indicate that the added value of neoadjuvant therapy is limited in early-stage Siewert type II AEG, and that SA may be sufficient for this population.
By contrast, in locally advanced patients, the NS group demonstrated a consistent and robust survival benefit both before and after PSM, supporting its role as a cornerstone of multimodal treatment (22). This conclusion is consistent with a prospective randomized trial in locally advanced Siewert type II/III AEG, which reported that neoadjuvant chemoradiotherapy significantly improved 3year OS (63.4% vs. 52.2%, P=0.02) and DFS (52.4% vs. 39.1%, P=0.049) compared with SA (23). The starkly contrasting prognostic associations between early and locally advanced stages underscore the importance of clinical stage as a critical determinant of treatment response in Siewert type II AEG. From a clinical perspective, these results suggest that SA may be a reasonable initial option for early‑stage patients, whereas neoadjuvant therapy should be favored in locally advanced disease to optimize survival outcomes. Future investigations incorporating biomarkers or molecular subtypes may further refine patient selection and identify those most likely to benefit from neoadjuvant or adjuvant strategies.
This study showed no significant difference in OS between the PT group and the NO group in patients with locally advanced Siewert type II AEG (P=0.87). Similar negative findings were reported by Yan et al. in locally advanced esophageal squamous cell carcinoma, where perioperative chemotherapy did not improve survival over neoadjuvant chemotherapy alone (24). However, this negative result in the overall population does not negate the potential value of adjuvant therapy in specific subgroups. The prospective cohort study by Saunders et al. and the multicenter study by Airoldi et al. both demonstrated that the benefit of adjuvant chemotherapy depends on tumor response to neoadjuvant therapy, indicating that clinical decisions should be guided by post‑neoadjuvant pathological assessment to select appropriate candidates, thereby avoiding unnecessary treatment and optimizing perioperative strategies (25,26).
This study included 1,847 patients with Siewert type II AEG from the SEER database, providing a sufficient sample size. Stratified analyses by early and locally advanced stages allowed us to evaluate the differential efficacy of neoadjuvant therapy across stage subgroups, offering stage-specific evidence. In locally advanced patients receiving neoadjuvant therapy, we further compared neoadjuvant-only versus PT to assess the independent incremental value of adjuvant therapy, addressing the limitation of previous studies that evaluated neoadjuvant and adjuvant as a combined regimen. We also developed separate nomogram models for early and locally advanced patients, which demonstrated good internal performance and provided practical tools for individualized prognosis estimation.
Nevertheless, several limitations should be acknowledged. First, as a retrospective analysis of the SEER database, the potential for selection bias and unmeasured confounding cannot be entirely excluded. Second, there is inherent heterogeneity in staging criteria across groups. According to SEER coding rules, pathological information takes priority over clinical information. For patients receiving neoadjuvant therapy followed by surgical resection, pretreatment clinical stage is primarily documented. Given potential discrepancies between clinical and pathological assessment of tumour burden, different criteria are adopted when categorizing patients into early-stage and locally advanced stage subgroups. Third, key variables such as tumor regression grade, pathological complete response, R0 resection rate, postoperative complications, and adjuvant therapy completion were unavailable, limiting in-depth analysis of the mechanisms underlying perioperative benefit. Fourth, the nomograms were only internally validated; external validation in independent cohorts is needed to confirm their generalizability. Prospective studies are warranted to further validate our findings. In addition, the neoadjuvant therapy cohort in the present study was dominated by chemoradiotherapy. Sample sizes for neoadjuvant chemotherapy and radiotherapy were limited, and further prospective studies are required to validate our findings for these two neoadjuvant therapeutic regimens.
Conclusions
In this cohort, neoadjuvant therapy, predominantly chemoradiotherapy, significantly improved OS in locally advanced Siewert type II AEG but not in early-stage disease. The role of chemotherapy or radiotherapy alone warrants further validation due to limited sample sizes. Stratification by stage provided suggestive value for treatment guidance, though these findings require cautious interpretation. Among locally advanced patients receiving neoadjuvant therapy, the addition of adjuvant therapy showed no statistically significant OS difference. Whether adjuvant therapy provides additional OS benefit in this setting remains to be determined through prospective studies.
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-2026-2069/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), 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-2026-2069/coif). The authors have no conflicts of interest to declare.
References
- 1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63.. 10.3322/caac.21834 [DOI] [PubMed] [Google Scholar]
- 2. Arnold M, Ferlay J, van Berge Henegouwen MI, et al. Global burden of oesophageal and gastric cancer by histology and subsite in 2018. Gut 2020;69:1564-71.. 10.1136/gutjnl-2020-321600 [DOI] [PubMed] [Google Scholar]
- 3. Nagtegaal ID, Odze RD, Klimstra D, et al. The 2019 WHO classification of tumours of the digestive system. Histopathology 2020;76:182-8.. 10.1111/his.13975 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Siewert JR. Adenocarcinoma of the esophago-gastric junction. Gastric Cancer 1999;2:87-8.. 10.1007/s101200050028 [DOI] [PubMed] [Google Scholar]
- 5. Qin X, Xu L, Wang X, et al. Molecular characteristics of proximal and distal esophagogastric junction adenocarcinoma. J Thorac Dis 2026;18:237.. 10.21037/jtd-2026-1-0027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Zheng J, Li Y, Fan L, et al. Prospective evaluation of radical surgery for adenocarcinoma of oesophagogastric junction: real-world insights from the CLAEG study. Gut 2026;75:1726-38.. 10.1136/gutjnl-2025-337393 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Wei H, Wang M, Xu N, et al. Comparison of minimally invasive single-position left transthoracic and esophageal hiatal approach versus laparoscopic transesophageal hiatus approach for Siewert type II adenocarcinoma of the esophagogastric junction. Surg Endosc 2026;40:6829-37.. 10.1007/s00464-026-12981-5 [DOI] [PubMed] [Google Scholar]
- 8. Karstens KF, Ghadban T, Sawez S, et al. Comparison of the 8th UICC staging system for esophageal and gastric cancers in Siewert type II junctional adenocarcinomas. Eur J Surg Oncol 2020;46:638-43.. 10.1016/j.ejso.2019.12.011 [DOI] [PubMed] [Google Scholar]
- 9. Al-Batran SE, Homann N, Pauligk C, et al. Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro-oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial. Lancet 2019;393:1948-57.. 10.1016/S0140-6736(18)32557-1 [DOI] [PubMed] [Google Scholar]
- 10. Cunningham D, Allum WH, Stenning SP, et al. Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med 2006;355:11-20.. 10.1056/NEJMoa055531 [DOI] [PubMed] [Google Scholar]
- 11. Anjum MU, Naqvi SAA, Jajja SA, et al. Efficacy of perioperative and neoadjuvant therapies in gastric and gastroesophageal junction adenocarcinoma: a network meta-analysis. Oncologist 2025;30:oyaf157.. 10.1093/oncolo/oyaf157 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Abdalla TSA, Kist M, Arify R, et al. Effect of multimodal therapy in the treatment of adenocarcinoma of the esophagus and gastroesophageal junction : a population-based analysis from the German Clinical Cancer Registry Group. Langenbecks Arch Surg 2026;411:148.. 10.1007/s00423-026-04066-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Zhang J, Liu H, Yu H, et al. Development of a novel staging classification for Siewert II adenocarcinoma of the esophagogastric junction after neoadjuvant chemotherapy. World J Gastrointest Oncol 2024;16:2541-54.. 10.4251/wjgo.v16.i6.2541 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Heckl SM, Behrens HM, Ebert U, et al. Impact of adjuvant therapy on outcomes of cancer of the stomach and gastroesophageal junction in the real-world. Gastric Cancer 2025;28:982-92.. 10.1007/s10120-025-01624-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. van Hagen P, Hulshof MC, van Lanschot JJ, et al. Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med 2012;366:2074-84.. 10.1056/NEJMoa1112088 [DOI] [PubMed] [Google Scholar]
- 16. Zhou Y, Tian M, Güngör C, et al. Neoadjuvant radiotherapy for locoregional Siewert type II gastroesophageal junction adenocarcinoma: A propensity scores matching analysis. PLoS One 2021;16:e0251555.. 10.1371/journal.pone.0251555 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Rompen IF, Crnovrsanin N, Nienhüser H, et al. Age-dependent benefit of neoadjuvant treatment in adenocarcinoma of the esophagus and gastroesophageal junction: a multicenter retrospective observational study of young versus old patients. Int J Surg 2023;109:3804-14.. 10.1097/JS9.0000000000000713 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Janczewski LM, Buchheit J, Jacobs RC, et al. Utilization and survival outcomes of neoadjuvant chemotherapy for early-stage gastric cancer. J Surg Oncol 2024;130:249-56.. 10.1002/jso.27732 [DOI] [PubMed] [Google Scholar]
- 19. Borggreve AS, van Rossum PSN, Mook S, et al. Frequency of surgical resection after starting neoadjuvant chemoradiotherapy in patients with esophageal cancer: A population-based cohort study. Eur J Surg Oncol 2019;45:1919-25.. 10.1016/j.ejso.2019.03.031 [DOI] [PubMed] [Google Scholar]
- 20. Perez Holguin RA, Olecki EJ, Stahl KA, et al. Management of Clinical T2N0 Esophageal and Gastroesophageal Junction Adenocarcinoma: What Is the Optimal Treatment?. J Gastrointest Surg 2022;26:2050-60.. 10.1007/s11605-022-05441-7 [DOI] [PubMed] [Google Scholar]
- 21. Waller GC, Patel DJ, Baker MS. Early-stage gastric and gastroesophageal junction cancer: Is there a survival benefit to neoadjuvant therapy?. Surgery 2022;172:1753-8.. 10.1016/j.surg.2022.08.018 [DOI] [PubMed] [Google Scholar]
- 22. Wirsik NM, Schmidt T, Nienhüser H, et al. Impact of the Surgical Approach for Neoadjuvantly Treated Gastroesophageal Junction Type II Tumors: A Multinational, High-volume Center Retrospective Cohort Analysis. Ann Surg 2023;278:683-91.. 10.1097/SLA.0000000000006011 [DOI] [PubMed] [Google Scholar]
- 23. Tian Y, Wang Q, Wang J, et al. Neoadjuvant chemoradiotherapy combined with surgery versus direct surgery in the treatment of Siewert type II and III adenocarcinomas of the esophagogastric junction: long-term prognostic analysis of a prospective randomized controlled trial. Zhonghua Wei Chang Wai Ke Za Zhi 2021;24:128-37. [Article in Chinese].. 10.3760/cma.j.cn.441530-20201019-00565 [DOI] [PubMed] [Google Scholar]
- 24. Yan W, Zhao P, Fu H, et al. Survival After Induction Chemotherapy and Esophagectomy Is Not Improved by Adjuvant Chemotherapy. Ann Thorac Surg 2019;108:1505-13.. 10.1016/j.athoracsur.2019.04.106 [DOI] [PubMed] [Google Scholar]
- 25. Saunders JH, Bowman CR, Reece-Smith AM, et al. The role of adjuvant platinum-based chemotherapy in esophagogastric cancer patients who received neoadjuvant chemotherapy prior to definitive surgery. J Surg Oncol 2017;115:821-9.. 10.1002/jso.24601 [DOI] [PubMed] [Google Scholar]
- 26. Airoldi M, Cammarota A, Carbonell-Asins JA, et al. Prognostic impact of pathological complete response and response-adapted outcomes in patients with gastroesophageal adenocarcinoma treated with neoadjuvant or perioperative treatments: a multicentric cohort study. ESMO Open 2026;11:106084.. 10.1016/j.esmoop.2026.106084 [DOI] [PMC free article] [PubMed] [Google Scholar]