Abstract
Background
The prognostic relevance of residual tumor deposits after neoadjuvant therapy (ypTD) in gastric cancer remains insufficiently defined. This study evaluated associations between ypTD positivity and 3-year disease-free survival (DFS) and overall survival (OS) after curative-intent gastrectomy.
Methods
This single-center retrospective cohort study included 336 patients with locally advanced, clinically lymph node-positive gastric adenocarcinoma who underwent neoadjuvant therapy followed by R0 gastrectomy. Patients were classified by postoperative ypTD status. Survival analyses were restricted to the postoperative 36-month observation window. Kaplan–Meier analysis, log-rank tests, Cox models, exploratory ypN-stratified analyses, and within-cohort model performance comparisons were performed.
Results
Fifty-one patients (15.2%) were ypTD-positive. ypTD-positive patients had higher proportions of ypT3–4 disease, ypN2–3 disease, and lymphovascular invasion. The 3-year DFS rate was lower in the ypTD-positive group than in the ypTD-negative group (41.2% vs. 69.8%; log-rank P < 0.001), whereas 3-year OS did not differ significantly (68.6% vs. 77.9%; P = 0.063). In the fully adjusted model, ypTD positivity was associated with increased DFS event risk (HR 1.79, 95% CI 1.08–2.95; P = 0.023), but not OS (HR 0.87, 95% CI 0.46–1.65; P = 0.676). Adding ypTD increased the 36-month time-dependent AUC from 0.598 to 0.645 in within-cohort exploratory comparisons.
Conclusions
ypTD positivity was associated with an increased risk of DFS events within 36 months after surgery, whereas no statistically significant association with OS was observed. ypTD may provide potential supplementary prognostic information for recurrence-risk assessment; however, its biological interpretation, potential clinical applicability, and long-term prognostic value require validation in larger multicenter cohorts.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12957-026-04500-6.
Keywords: Gastric cancer, Neoadjuvant therapy, Residual tumor deposits, Disease-free survival, Overall survival, Pathological risk stratification, Retrospective cohort study
Background
Gastric cancer remains a gastrointestinal malignancy with a substantial global disease burden [1]. For locally advanced gastric cancer, neoadjuvant therapy followed by curative-intent gastrectomy has become an important treatment strategy [2]. However, recurrence still occurs in a considerable proportion of patients after neoadjuvant therapy and curative-intent surgery [3, 4]. Therefore, identifying patients at high risk of recurrence based on residual pathological features after surgery is a key issue in postoperative management after neoadjuvant therapy [5].
Postoperative risk stratification is generally based on pathological factors such as ypT category, ypN category, resection margin status, lymphovascular invasion (LVI), and perineural invasion (PNI) [6]. Tumor deposits (TDs) are defined as discrete cancer nodules located within the lymphatic drainage area of the primary tumor, discontinuous from the primary lesion, and lacking identifiable lymph node structures [7]. In colorectal cancer, TDs have been incorporated into the AJCC/TNM staging system under defined conditions [8]. In contrast, their significance in gastric cancer remains controversial [7, 9]. Previous studies and systematic reviews have suggested that TD positivity is associated with more advanced T or N category, LVI or PNI positivity, higher tumor grade, and unfavorable survival outcomes; however, whether TDs should be considered an independent staging or risk-stratification factor in gastric cancer remains unsettled [9–12]. Most existing gastric cancer studies of TDs have been based on upfront surgery cohorts or mixed treatment strategies, and evidence specifically addressing residual tumor deposits after neoadjuvant therapy remains limited [13–15].
Residual tumor deposits after neoadjuvant therapy (ypTD) may have distinct clinical and pathological implications. Conventional TDs and ypTDs should not be assumed to be biologically equivalent. Conventional TDs are typically assessed in untreated surgical specimens, whereas ypTDs are identified after neoadjuvant therapy in a pathological setting affected by tumor regression, fibrosis, necrosis, and tissue remodeling [16]. Accordingly, ypTDs may represent heterogeneous pathological entities, including residual tumor within obliterated lymph nodes, discontinuous tumor spread, lymphovascular dissemination, or treatment-associated remodeling around residual cancer foci. Meanwhile, ypTD status may correlate with conventional postoperative pathological factors such as ypT, ypN, LVI, and PNI. Therefore, whether ypTD provides potential supplementary prognostic information beyond existing ypTNM staging and pathological risk factors remains to be clarified [6, 17]. This issue may be particularly relevant in patients with ypN0 disease, in whom conventional lymph node staging may not fully capture residual post-treatment risk. Previous studies have shown that LVI, PNI, and total lymph node yield remain prognostically relevant in patients with gastric cancer who achieve ypN0 status after neoadjuvant therapy, and TDs have also been associated with adverse outcomes in lymph node-negative gastric cancer without neoadjuvant therapy [18, 19]. Therefore, the prognostic significance of ypTD in patients with ypN0 gastric cancer after neoadjuvant therapy remains insufficiently defined.
Based on this background, this study used a single-center cohort of patients with gastric cancer who underwent curative-intent gastrectomy after neoadjuvant therapy to evaluate the clinicopathological characteristics of ypTD positivity and its association with postoperative 3-year disease-free survival (DFS) and overall survival (OS). We further explored the association between ypTD and DFS according to ypN status and assessed whether ypTD could provide potential incremental prognostic information for 36-month DFS event risk beyond conventional pathological factors.
Methods
Study design and patients
This was a single-center retrospective cohort study reported in accordance with the STROBE statement [20]. Eligible patients were those with pathologically confirmed gastric adenocarcinoma who underwent curative-intent gastrectomy after neoadjuvant therapy at the Affiliated Hospital of Qingdao University between March 2020 and March 2022. The main inclusion criteria were locally advanced, clinically lymph node-positive (cN-positive) disease without distant metastasis before treatment, as assessed by imaging and a multidisciplinary team; receipt of neoadjuvant therapy; and subsequent R0 gastrectomy with curative intent. Patients were excluded if they had distant metastasis before treatment or intraoperatively, underwent non-curative-intent surgery or non-R0 resection, had missing or indeterminate postoperative ypTD assessment, lacked key pathological covariates required for the primary adjusted analyses, or had unavailable follow-up information required to ascertain 36-month DFS or OS status. Key pathological covariates were ypT category, ypN category, lymphovascular invasion, and perineural invasion.
Of 352 initially screened patients, 16 were excluded because of missing or indeterminate postoperative ypTD assessment (n = 6), missing key pathological covariates (n = 5), or unavailable information to ascertain 36-month DFS or OS status (n = 5), leaving 336 patients in the analytic cohort. The study protocol was approved by the Ethics Committee of the Affiliated Hospital of Qingdao University (approval no. QYFYWZLL50115), and the requirement for informed consent was waived because anonymized retrospective clinical data were used. The follow-up cutoff date was March 2025. For each outcome, all included patients had either at least 36 months of follow-up after surgery or experienced the corresponding endpoint event within 36 months; therefore, DFS and OS analyses were restricted to the postoperative 36-month observation window. The study flow diagram is shown in Fig. 1.
Fig. 1.
Study flow diagram . A total of 352 patients with locally advanced, clinically lymph node-positive (cN-positive) gastric adenocarcinoma who underwent neoadjuvant therapy followed by gastrectomy were screened. After excluding 16 patients with missing or indeterminate postoperative ypTD assessment (n = 6), missing key pathological covariates (n = 5), or unavailable 36-month DFS or OS status (n = 5), 336 patients who underwent curative-intent gastrectomy with R0 resection were included in the analytic cohort and classified as ypTD-negative (n = 285) or ypTD-positive (n = 51). cN, clinical lymph node category; DFS, disease-free survival; OS, overall survival; R0, microscopically margin-negative resection; ypTD, residual tumor deposits after neoadjuvant therapy
Variable definitions
The primary exposure was the presence of residual tumor deposits after neoadjuvant therapy (ypTD) in the postoperative pathology report. ypTD was defined as residual cancer nodules located in the primary tumor bed or its lymphatic drainage area after neoadjuvant therapy that were discontinuous from the primary lesion and lacked identifiable lymph node structures. ypTD status was extracted from the final postoperative pathology report, and patients were categorized as ypTD-negative or ypTD-positive.
Available clinical, operative, pathological, and treatment records were reviewed to determine whether variables related to treatment intensity, surgical details, postoperative complications, postoperative treatment, histology, and pathological response could be consistently extracted; only consistently documented variables were included in the relevant analyses.
Variables included age, sex, Eastern Cooperative Oncology Group (ECOG) performance status before neoadjuvant therapy [21], Charlson Comorbidity Index (CCI) [22], body mass index (BMI), clinical T category, pretreatment clinical N subcategory, neoadjuvant treatment regimen, use of immunotherapy and/or targeted therapy during neoadjuvant treatment, postoperative pathological ypT and ypN categories, total lymph node yield, Lauren classification, signet-ring cell component, lymphovascular invasion (LVI), perineural invasion (PNI), and postoperative adjuvant therapy. Surgical variables included surgical approach, extent of gastrectomy, and extent of lymphadenectomy. Resection margin status was not evaluated as a between-group variable because all patients in the analytic cohort underwent R0 resection by study design. Neoadjuvant treatment regimens were categorized according to the main chemotherapy regimen as SOX, XELOX/CAPOX, FLOT, DOS, FOLFOX, or other regimens. Clinical and postoperative pathological categories were determined according to the eighth edition of the AJCC TNM staging system [23]. Because all patients were clinically lymph node-positive before treatment and no patient was categorized as cN1, pretreatment clinical N subcategory was presented descriptively as cN2 versus cN3 within the cN-positive cohort. For descriptive analyses and regression models, age, BMI, ECOG performance status, CCI, clinical T category, postoperative pathological T category, and postoperative pathological N category were categorized as < 70 versus ≥ 70 years, < 25 versus ≥ 25 kg/m², 0–1 versus ≥ 2, 0 versus ≥ 1, cT1–3 versus cT4, ypT0–2 versus ypT3–4, and ypN0–1 versus ypN2–3, respectively. Neoadjuvant immunotherapy and/or targeted therapy was defined as any use during the neoadjuvant treatment phase. Postoperative adjuvant therapy was defined as receipt of at least one cycle of postoperative systemic adjuvant treatment after curative-intent gastrectomy, including chemotherapy with or without immunotherapy or targeted therapy. This binary definition was used because detailed postoperative adjuvant-treatment intensity, including administered cycles, treatment completion, dose reduction or discontinuation, and postoperative immunotherapy and/or targeted therapy exposure, was not consistently documented in extractable retrospective records.
Outcome definitions
The primary outcome was DFS, defined as the time from the date of surgery to the first documented tumor recurrence or death from any cause, whichever occurred first. The secondary outcome was OS, defined as the time from the date of surgery to death from any cause. Tumor recurrence and death status were ascertained from institutional medical records, postoperative surveillance records, imaging and/or endoscopic follow-up documentation, readmission records, and follow-up information available at the study center. The primary survival analysis window was the first 36 months after surgery. All DFS and OS analyses were restricted to this window; patients who did not experience the corresponding event were administratively censored at 36 months. Three-year DFS and 3-year OS were estimated using the Kaplan–Meier method. Exploratory model performance comparisons for 36-month DFS events used postoperative month 36 as the statistical time point.
Statistical analysis
Categorical variables were compared using Pearson’s χ² test without continuity correction or Fisher’s exact test, and continuous variables were compared using the Wilcoxon rank-sum test. For multicategory variables with small expected cell counts, the Fisher–Freeman–Halton exact test or Monte Carlo exact method was used. Two-sided P values < 0.05 were considered statistically significant. Kaplan–Meier curves were generated for DFS and OS [24], and between-group differences were compared using the log-rank test. All survival analyses were based on survival time and event status within the postoperative 36-month observation window. Cox proportional hazards models were used to evaluate the association of ypTD with DFS and OS, and results were expressed as hazard ratios (HRs) with 95% confidence intervals (CIs) [25]. HRs compared ypTD-positive with ypTD-negative patients within the postoperative 0–36-month window. The proportional hazards assumption was assessed using Schoenfeld residuals [26].
To sequentially adjust for potential confounders, the following Cox models were constructed: a univariable model; a core pathological model including ypT3–4, ypN2–3, LVI, and PNI; a clinical-treatment model including Eastern Cooperative Oncology Group performance status before neoadjuvant therapy, CCI, cT4, and postoperative adjuvant therapy; and a fully adjusted model including all variables in the clinical-treatment and core pathological models. Clinical T4 status was retained as a clinically relevant pretreatment local disease-burden variable. Pretreatment cN subcategory, neoadjuvant regimen, and the use of neoadjuvant immunotherapy and/or targeted therapy were summarized descriptively but were not included in the primary multivariable Cox models for parsimony, because all patients were clinically lymph node-positive, postoperative ypN category represented post-treatment nodal burden, and some regimen categories were sparse. To reduce overfitting and preserve model stability given the number of DFS events within the 36-month observation window, the primary multivariable Cox models were restricted to prespecified clinically relevant covariates rather than all variables reported in the descriptive tables. Exploratory stratified analyses were performed according to ypN status, with patients categorized as ypN0 or ypN-positive; ypN positivity was defined as ypN1–3. Stratum-specific Cox models were adjusted for ypT3–4, LVI, and PNI. Interaction testing was performed by including ypTD, ypN status, and the ypTD × ypN interaction term in a Cox model adjusted for ypT3–4, LVI, and PNI. No adjustment for multiple comparisons was performed for the exploratory stratified analyses or model performance comparisons, and these findings were interpreted as exploratory.
To explore the potential incremental prognostic information provided by ypTD for postoperative 36-month DFS event risk, a within-cohort exploratory model-performance comparison was conducted. The base pathological model included ypT3–4, ypN2–3, LVI, and PNI, whereas the ypTD-extended model additionally included ypTD. The two models were compared using Harrell’s C-index within the postoperative 36-month observation window, time-dependent AUC at 12, 24, and 36 months after surgery, and decision curve analysis (DCA) for predicting 36-month DFS events [27–29]. The 95% CIs for C-indices, C-index differences, and AUC differences were estimated using 1000 bootstrap resamples. These model performance metrics were not optimism-corrected or externally validated and were used primarily to describe model performance within the current cohort. In a complete-category sensitivity analysis, the same pathological model framework was used, with ypT and ypN treated as complete multicategory variables rather than binary variables while retaining LVI, PNI, and ypTD, to assess the robustness of the association between ypTD and DFS. The primary analyses used a complete-case approach for variables included in each model, and no statistical imputation was performed. All statistical analyses were performed using R software, version 4.5.2.
Results
Patient characteristics according to ypTD status
A total of 336 patients were included, of whom 51 (15.2%) were ypTD-positive and 285 (84.8%) were ypTD-negative. All patients were assessed as clinically lymph node-positive before treatment. Within this cN-positive cohort, cN3 disease was numerically more frequent in ypTD-positive patients than in ypTD-negative patients (66.7% vs. 52.6%, P = 0.064), although the difference did not reach statistical significance. There were no statistically significant between-group differences in age, sex, ECOG performance status before neoadjuvant therapy, CCI, BMI, cT category, neoadjuvant treatment regimen, use of neoadjuvant immunotherapy and/or targeted therapy, PNI positivity, or postoperative adjuvant therapy. In contrast, ypTD-positive patients had significantly higher proportions of ypT3–4 disease, ypN2–3 disease, and LVI positivity. Total lymph node yield also differed significantly between groups, although the absolute difference in medians was small and should be interpreted cautiously. Patient characteristics according to ypTD status are summarized in Table 1. Additional surgical and histological characteristics according to ypTD status are provided in Supplementary Tables S5 and S6 (Additional file 1).
Table 1.
Baseline, treatment-related, and clinicopathological characteristics according to ypTD status
| Variable | ypTD-negative (n = 285) | ypTD-positive (n = 51) | P value |
|---|---|---|---|
| Age group | 0.598 | ||
| < 70 years | 179/285 (62.8%) | 34/51 (66.7%) | |
| ≥ 70 years | 106/285 (37.2%) | 17/51 (33.3%) | |
| Sex | 0.126 | ||
| Female | 129/285 (45.3%) | 29/51 (56.9%) | |
| Male | 156/285 (54.7%) | 22/51 (43.1%) | |
| ECOG performance status before neoadjuvant therapy | 0.097 | ||
| 0–1 | 235/285 (82.5%) | 37/51 (72.5%) | |
| ≥ 2 | 50/285 (17.5%) | 14/51 (27.5%) | |
| CCI | 0.066 | ||
| 0 | 168/285 (58.9%) | 23/51 (45.1%) | |
| ≥ 1 | 117/285 (41.1%) | 28/51 (54.9%) | |
| BMI category, kg/m² | 0.244 | ||
| < 25 | 171/285 (60.0%) | 35/51 (68.6%) | |
| ≥ 25 | 114/285 (40.0%) | 16/51 (31.4%) | |
| cT category | 0.086 | ||
| cT1–3 | 160/285 (56.1%) | 22/51 (43.1%) | |
| cT4 | 125/285 (43.9%) | 29/51 (56.9%) | |
| Pretreatment cN subcategory | 0.064 | ||
| cN2 | 135/285 (47.4%) | 17/51 (33.3%) | |
| cN3 | 150/285 (52.6%) | 34/51 (66.7%) | |
| Neoadjuvant treatment regimen | 0.399 | ||
| SOX | 155/285 (54.4%) | 23/51 (45.1%) | |
| XELOX/CAPOX | 66/285 (23.2%) | 12/51 (23.5%) | |
| FLOT | 21/285 (7.4%) | 7/51 (13.7%) | |
| DOS | 18/285 (6.3%) | 6/51 (11.8%) | |
| FOLFOX | 14/285 (4.9%) | 2/51 (3.9%) | |
| Other | 11/285 (3.9%) | 1/51 (2.0%) | |
| Neoadjuvant immunotherapy and/or targeted therapy | 0.718 | ||
| No | 251/285 (88.1%) | 44/51 (86.3%) | |
| Yes | 34/285 (11.9%) | 7/51 (13.7%) | |
| ypT category | < 0.001 | ||
| ypT0–2 | 195/285 (68.4%) | 21/51 (41.2%) | |
| ypT3–4 | 90/285 (31.6%) | 30/51 (58.8%) | |
| ypN category | 0.002 | ||
| ypN0–1 | 217/285 (76.1%) | 28/51 (54.9%) | |
| ypN2–3 | 68/285 (23.9%) | 23/51 (45.1%) | |
| Total lymph node yield, median (IQR) | 33 (31–35) | 32 (30.5–33.5) | 0.006 |
| LVI | 0.012 | ||
| Negative | 211/285 (74.0%) | 29/51 (56.9%) | |
| Positive | 74/285 (26.0%) | 22/51 (43.1%) | |
| PNI | 0.762 | ||
| Negative | 218/285 (76.5%) | 40/51 (78.4%) | |
| Positive | 67/285 (23.5%) | 11/51 (21.6%) | |
| Postoperative adjuvant therapy | 0.782 | ||
| No | 25/285 (8.8%) | 3/51 (5.9%) | |
| Yes | 260/285 (91.2%) | 48/51 (94.1%) | |
Data are presented as n/N (%) or median (interquartile range). Categorical and continuous variables were compared as described in the Statistical analysis section. ECOG refers to Eastern Cooperative Oncology Group performance status before neoadjuvant therapy; no patient had ECOG performance status 4, and ECOG performance status ≥ 2 comprised ECOG performance status 2 and 3. All patients were clinically lymph node-positive before treatment; therefore, binary cN status was not compared between groups. Pretreatment cN subcategory was presented as cN2 versus cN3 within the cN-positive cohort, and no patient was categorized as cN1
CCI Charlson Comorbidity Index, BMI body mass index, IQR interquartile range, cN clinical lymph node category, cT clinical tumor category, LVI lymphovascular invasion, PNI perineural invasion, ypN postoperative pathological lymph node category after neoadjuvant therapy, ypT postoperative pathological tumor category after neoadjuvant therapy, ypTD residual tumor deposits after neoadjuvant therapy, SOX S-1 plus oxaliplatin, XELOX/CAPOX capecitabine plus oxaliplatin, FLOT fluorouracil/leucovorin, oxaliplatin, and docetaxel, DOS docetaxel, oxaliplatin, and S-1, FOLFOX fluorouracil/leucovorin plus oxaliplatin. Neoadjuvant immunotherapy and/or targeted therapy was defined as any use during the neoadjuvant treatment phase
ypTD status and 3-year DFS and OS
Within the postoperative 36-month observation window, 116 DFS events and 79 deaths were observed. The 3-year DFS rate was significantly lower in the ypTD-positive group than in the ypTD-negative group (41.2% vs. 69.8%, log-rank P < 0.001). The 3-year OS rate was numerically lower in the ypTD-positive group than in the ypTD-negative group, but the between-group difference did not reach statistical significance (68.6% vs. 77.9%, log-rank P = 0.063). Kaplan–Meier curves with number-at-risk tables are shown in Fig. 2, and detailed 3-year estimates are provided in Supplementary Table S1 [see Additional file 1].
Fig. 2.
Kaplan–Meier curves for DFS and OS according to ypTD status. Kaplan–Meier curves with number-at-risk tables are shown for DFS (A) and OS (B) according to postoperative ypTD status within the postoperative 36-month observation window. DFS, disease-free survival; OS, overall survival; ypTD, residual tumor deposits after neoadjuvant therapy
Cox analysis of the association between ypTD and DFS or OS
In the univariable Cox analysis within the postoperative 36-month observation window, ypTD positivity was significantly associated with an increased hazard of DFS events (HR 2.74, 95% CI 1.81–4.15, P < 0.001). This association remained significant in the core pathological model (HR 2.38, 95% CI 1.52–3.73, P < 0.001), the clinical-treatment model (HR 2.11, 95% CI 1.32–3.37, P = 0.002), and the fully adjusted model (HR 1.79, 95% CI 1.08–2.95, P = 0.023).
The association between ypTD positivity and OS did not reach statistical significance within the postoperative 36-month observation window. In the univariable model, ypTD-positive patients showed a numerically higher hazard of all-cause mortality, but this association was not statistically significant (HR 1.67, 95% CI 0.96–2.89, P = 0.067). The association also did not reach statistical significance in the core pathological, clinical-treatment, or fully adjusted models. The associations of ypTD with DFS and OS across Cox models are summarized in Table 2, and complete covariate results from the fully adjusted Cox models are provided in Supplementary Table S2 [see Additional file 1].
Table 2.
Association of ypTD with DFS and OS across Cox models
| Outcome | Model | HR (95% CI) | P value |
|---|---|---|---|
| DFS | Univariable model | 2.74 (1.81–4.15) | < 0.001 |
| Core pathological model | 2.38 (1.52–3.73) | < 0.001 | |
| Clinical-treatment model | 2.11 (1.32–3.37) | 0.002 | |
| Fully adjusted clinical-treatment-pathological model | 1.79 (1.08–2.95) | 0.023 | |
| OS | Univariable model | 1.67 (0.96–2.89) | 0.067 |
| Core pathological model | 1.32 (0.74–2.36) | 0.350 | |
| Clinical-treatment model | 1.14 (0.62–2.09) | 0.669 | |
| Fully adjusted clinical-treatment-pathological model | 0.87 (0.46–1.65) | 0.676 |
HRs are shown for ypTD-positive versus ypTD-negative patients. For DFS, HRs indicate the relative hazard of DFS events; for OS, HRs indicate the relative hazard of all-cause mortality. The core pathological model included ypT3–4, ypN2–3, LVI, and PNI. The clinical-treatment model included Eastern Cooperative Oncology Group performance status before neoadjuvant therapy, CCI, cT4, and postoperative adjuvant therapy. The fully adjusted clinical-treatment-pathological model included all variables in the clinical-treatment and core pathological models. Complete covariate results from the fully adjusted models are provided in Supplementary Table S2. HRs should be interpreted within the postoperative 0–36-month observation window
HR hazard ratio, CI confidence interval, DFS disease-free survival, OS overall survival, CCI Charlson Comorbidity Index, cT clinical tumor category, LVI lymphovascular invasion, PNI perineural invasion, ypN postoperative pathological lymph node category after neoadjuvant therapy, ypT postoperative pathological tumor category after neoadjuvant therapy, ypTD residual tumor deposits after neoadjuvant therapy
Schoenfeld residual tests showed no evidence of a substantial violation of the proportional hazards assumption in the fully adjusted DFS and OS models fitted within the postoperative 36-month observation window (global P = 0.339 and 0.449, respectively). The DFS-specific hazard ratio estimates for ypTD positivity across Cox adjustment models are graphically summarized in Fig. 3.
Fig. 3.
Association between ypTD positivity and DFS across Cox models. Forest plot showing hazard ratios and 95% confidence intervals for DFS events associated with ypTD positivity across sequential Cox models within the postoperative 36-month observation window. Detailed model covariates are described in the Methods. CI, confidence interval; DFS, disease-free survival; HR, hazard ratio; ypTD, residual tumor deposits after neoadjuvant therapy
ypN-stratified analysis
In the exploratory ypN-stratified analysis, ypTD-positive patients had lower 3-year DFS rates in both the ypN0 and ypN-positive subgroups within the postoperative 36-month observation window. In the ypN0 subgroup, the 3-year DFS rate was lower in ypTD-positive than in ypTD-negative patients (41.7% vs. 67.2%, log-rank P = 0.047). In the stratum-specific adjusted Cox model, the hazard of DFS events was numerically higher in ypTD-positive patients but did not reach statistical significance (HR 2.20, 95% CI 0.98–4.91, P = 0.055). In the ypN-positive subgroup, the 3-year DFS rate was significantly lower in ypTD-positive than in ypTD-negative patients (41.0% vs. 71.8%, log-rank P < 0.001), and ypTD positivity was significantly associated with an increased hazard of DFS events in the adjusted Cox model (HR 3.04, 95% CI 1.84–5.02, P < 0.001). No statistically significant interaction was observed between ypTD and ypN status (P for interaction = 0.226). These exploratory stratified findings are summarized in Supplementary Table S3 [see Additional file 1] and presented in Supplementary Fig. S1 [see Additional file 2].
Exploratory model-performance comparison for 36-month DFS event risk
To explore whether ypTD provided potential incremental prognostic information for postoperative 36-month DFS event risk, we compared a base pathological model with a ypTD-extended model within the current cohort. The base model included ypT3–4, ypN2–3, LVI, and PNI, whereas the ypTD-extended model additionally included ypTD. Adding ypTD increased Harrell’s C-index within the 36-month observation window from 0.590 (95% CI 0.535–0.646) to 0.634 (95% CI 0.579–0.689), with a difference of 0.044 (bootstrap-derived 95% CI 0.011–0.072). Time-dependent AUC point estimates were higher for the ypTD-extended model than for the base model at 12, 24, and 36 months. Because earlier AUC estimates may be influenced by the number and timing of early events, we primarily interpreted the clinically relevant 36-month result. At 36 months, the AUC increased from 0.598 to 0.645 after adding ypTD, with a difference of 0.047 (bootstrap-derived 95% CI 0.008–0.080). Decision curve analysis suggested a higher net benefit for the ypTD-extended model across threshold probabilities of 25%–54% for predicting 36-month DFS events. These within-cohort exploratory model-performance comparisons are summarized in Supplementary Table S4 [see Additional file 1] and presented in Supplementary Fig. S2 [see Additional file 3].
To preserve pathological-stage granularity, we further fitted a DFS Cox model in which ypT and ypN were treated as complete multicategory variables while retaining ypTD, LVI, and PNI. In this complete-category model, ypTD positivity remained associated with an increased hazard of DFS events within the postoperative 36-month observation window (HR, 2.41; 95% CI, 1.53–3.79; P < 0.001; Table 3).
Table 3.
Cox model for DFS events using complete ypT and ypN categories
| Variable | HR | 95% CI | P value |
|---|---|---|---|
| ypTD status | |||
| Negative | Reference | ||
| Positive | 2.41 | 1.53–3.79 | < 0.001 |
| ypT category | |||
| ypT0 | Reference | ||
| ypT1 | 0.86 | 0.23–3.24 | 0.823 |
| ypT2 | 1.12 | 0.39–3.18 | 0.832 |
| ypT3 | 1.38 | 0.56–3.40 | 0.484 |
| ypT4 | 1.75 | 0.72–4.26 | 0.217 |
| ypN category | |||
| ypN0 | Reference | ||
| ypN1 | 1.08 | 0.57–2.06 | 0.814 |
| ypN2 | 1.26 | 0.66–2.40 | 0.483 |
| ypN3 | 1.58 | 0.82–3.03 | 0.170 |
| Lymphovascular invasion | |||
| Absent | Reference | ||
| Present | 1.14 | 0.75–1.73 | 0.539 |
| Perineural invasion | |||
| Absent | Reference | ||
| Present | 1.24 | 0.80–1.91 | 0.333 |
Hazard ratios were estimated using a Cox proportional hazards model for DFS events within the postoperative 36-month observation window. ypT and ypN categories were treated as complete multicategory variables, with ypT0 and ypN0 used as the reference categories, respectively. The model also retained ypTD status, lymphovascular invasion, and perineural invasion.
ypTD residual tumor deposits after neoadjuvant therapy, DFS disease-free survival, HR hazard ratio, CI confidence interval
Discussion
This study showed that ypTD-positive patients after neoadjuvant therapy for gastric cancer had a more adverse postoperative pathological profile, including higher proportions of ypT3–4 disease, ypN2–3 disease, and LVI positivity, and a significantly lower 3-year DFS rate. This 15.2% ypTD positivity rate should be interpreted in the context of the 17.9% tumor-deposit prevalence reported in a recent systematic review of resectable gastric cancer, although direct comparison is limited [12]. Although total lymph node yield differed statistically between groups, the absolute median difference was small and should not be overinterpreted as a major explanation for the observed association between ypTD and DFS. After adjustment for clinical, treatment-related, and pathological factors, ypTD positivity remained associated with an increased hazard of DFS events within the postoperative 36-month observation window, whereas its association with OS did not reach statistical significance. The within-cohort exploratory model-performance comparison further suggested that adding ypTD to conventional pathological factors modestly improved Harrell’s C-index and the 36-month time-dependent AUC, with DCA suggesting higher net benefit across an intermediate range of threshold probabilities. Overall, these findings suggest that ypTD may provide supplementary information for postoperative recurrence-risk assessment after neoadjuvant therapy for gastric cancer, but they should not be taken as evidence that ypTD is an established prognostic factor for OS.
The attenuation of the association with DFS after adjustment is important for interpretation. The HR for ypTD positivity decreased from 2.74 in the univariable model to 1.79 in the fully adjusted model, suggesting that the association between ypTD and DFS may partly overlap with the broader adverse pathological context, including advanced ypT category, higher ypN category, and LVI. Therefore, ypTD should be interpreted as a marker that may provide supplementary prognostic information within this pathological context, rather than as an isolated or fully independent biological determinant.
ypTD may also be interpreted as a marker of incomplete pathological response to neoadjuvant therapy rather than as an independent causal determinant of recurrence. Patients with ypTD may represent a subgroup with residual viable tumor burden, incomplete pathological regression, or a treatment-resistant phenotype after neoadjuvant therapy. Therefore, the observed association between ypTD and DFS should be interpreted in the broader context of treatment response, postoperative pathological stage, LVI, and other adverse pathological features.
Histological background may also influence the interpretation of ypTD. Diffuse-type or poorly cohesive biology, signet-ring cell differentiation, tumor fragmentation, and lymphovascular spread may represent sources of histological or pathological heterogeneity that complicate the interpretation of discontinuous residual tumor foci after neoadjuvant therapy. In the present cohort, Lauren classification and signet-ring cell component were consistently extractable and were summarized in Supplementary Table S6; neither differed significantly according to ypTD status. However, signet-ring cell percentage, poorly cohesive histology, and tumor regression grade were not uniformly documented, precluding more granular analyses. Recent studies have suggested that the proportion of signet-ring cells, rather than the simple presence of signet-ring cells alone, may carry prognostic information in mixed-type and poorly cohesive gastric cancer [30, 31]. Therefore, the observed association between ypTD and DFS may partly reflect unmeasured histological or pathological-response heterogeneity, and future studies should incorporate standardized histological subtype, signet-ring cell percentage, and tumor regression grade assessment.
This study did not observe a statistically significant association between ypTD and OS within 36 months after surgery. DFS usually reflects postoperative recurrence risk earlier than OS, whereas OS can be influenced by post-recurrence treatment, subsequent lines of therapy, non-cancer-related death, follow-up duration, and the number of death events [32, 33]. Thus, the present findings should be interpreted primarily as supporting an association with recurrence risk rather than with OS.
The ypN-stratified analysis further showed that ypTD-positive patients had lower 3-year DFS rates in both the ypN0 and ypN-positive subgroups. In the ypN0 subgroup, this pattern suggested that ypTD may help identify, at an exploratory level, a subset of patients with ypN0 disease after neoadjuvant therapy who remain at higher recurrence risk [18, 19]. However, the number of ypTD-positive patients in the ypN0 subgroup was small, the adjusted Cox estimate approached but did not reach statistical significance, and no statistically significant interaction was observed between ypTD and ypN status. Therefore, these stratified findings should be regarded as exploratory and should not be interpreted as evidence that the prognostic association of ypTD differs according to ypN status.
From a clinical perspective, ypTD may complement existing postoperative pathological risk stratification. In the within-cohort exploratory model-performance comparison, adding ypTD to the base pathological model including ypT, ypN, LVI, and PNI improved model discrimination for 36-month DFS events, and DCA suggested higher net benefit for the extended model across an intermediate range of threshold probabilities. These findings suggest that ypTD may help inform future recurrence-risk assessment frameworks, but they should not be interpreted as evidence of established clinical utility. At this stage, ypTD positivity alone should not be used to determine postoperative treatment intensification, nor should these findings be used to justify immediate incorporation of ypTD into the gastric cancer staging system [11, 12]. Further external validation and prospective studies are needed.
This study has several limitations. First, this was a single-center retrospective cohort study, and heterogeneity may have existed in patient selection, neoadjuvant treatment, postoperative adjuvant therapy, and post-recurrence treatment. Because all patients were clinically lymph node-positive before treatment, the findings may not be directly generalizable to patients with clinically node-negative disease, patients treated with upfront surgery, or broader gastric cancer populations. Although neoadjuvant regimens and the use of immunotherapy and/or targeted therapy did not differ significantly between ypTD groups, we additionally reviewed available treatment, operative, postoperative, and pathology records, but the retrospective data were insufficient to comprehensively evaluate treatment cycles, treatment completion rates, dose intensity, specific immunotherapeutic or targeted therapy agents, the interval between neoadjuvant therapy and surgery, detailed postoperative treatment intensity, more granular surgical details, standardized details on postoperative complications, or some histological and pathological-response variables. Therefore, residual confounding related to treatment response, treatment intensity, postoperative complications, postoperative management, and other unmeasured factors could not be excluded despite adjustment for major clinical, treatment-related, and pathological factors.
Second, the number of ypTD-positive patients was limited, especially in the ypN0 subgroup, which reduced the stability of subgroup analyses. Sex-stratified survival analyses were not performed because subgroup event counts were limited. Third, the absence of a statistically significant association between ypTD and OS may be related to the limited number of OS events, the 36-month observation window, subsequent treatments, and limited stability of model estimates. Because DFS and OS analyses were restricted to 36 months after surgery, the Cox results reflect relative hazards within this period rather than long-term hazards over the entire follow-up. Fourth, pathological identification of ypTD may be affected by specimen sampling, slide evaluation, and pathologist judgment, and standardized assessment remains needed [12, 16]. Fifth, the main pathological models used binary ypT and ypN variables, although robustness was assessed using a multicategory ypT/ypN sensitivity model; complete ypTNM stage grouping was not further evaluated. Finally, model-performance metrics were calculated within the current cohort and were not optimism-corrected, externally validated, or assessed for calibration. Larger multicenter studies with standardized pathological assessment, longer follow-up, calibration assessment, and external validation are needed before ypTD can be used to refine risk stratification or inform postoperative treatment decisions [11, 12].
Conclusions
In this single-center cohort of patients with clinically lymph node-positive gastric cancer who underwent neoadjuvant therapy and curative-intent gastrectomy, ypTD positivity was associated with more adverse postoperative pathological features and a lower 3-year DFS rate. After adjustment for clinical, treatment-related, and pathological covariates, ypTD positivity remained associated with an increased hazard of DFS events within the postoperative 36-month observation window, whereas no statistically significant association with OS was observed. Exploratory ypN-stratified and within-cohort model-performance analyses suggested that ypTD may provide supplementary prognostic information for postoperative recurrence-risk assessment, but these findings should not be interpreted as evidence of established clinical utility. Larger multicenter studies with standardized pathological assessment, longer follow-up, calibration assessment, and external validation are needed before ypTD can be used to inform postoperative treatment intensification or staging refinement.
Supplementary Information
Additional File 1. Supplementary tables and supplementary figure legends (.docx). Supplementary Tables S1–S6 and legends for Supplementary Figs. S1–S2.
Additional File 2. Supplementary Fig. S1 (.png). Exploratory ypN-stratified analysis of ypTD and DFS.
Additional File 3. Supplementary Fig. S2 (.png). Within-cohort exploratory model-performance comparison between the base pathological model and the ypTD-extended model for 36-month DFS event risk.
Acknowledgements
Not applicable.
Declaration of AI-assisted language editing
ChatGPT (OpenAI) was used only for limited English-language polishing of selected sentences. The authors reviewed and approved all edited text and take full responsibility for the accuracy, integrity, and scientific content of the final manuscript. AI-assisted tools were not used to generate study data, perform statistical analyses, interpret results, or draw scientific conclusions.
Clinical trial registration
Not applicable.
Abbreviations
- AI
Artificial intelligence
- AJCC
American Joint Committee on Cancer
- AUC
Area under the curve
- BMI
Body mass index
- CAPOX
Capecitabine plus oxaliplatin
- CCI
Charlson Comorbidity Index
- CI
Confidence interval
- C-index
Concordance index
- cN
Clinical lymph node category
- cT
Clinical tumor category
- DCA
Decision curve analysis
- DFS
Disease-free survival
- DOS
Docetaxel, oxaliplatin, and S-1
- ECOG
Eastern Cooperative Oncology Group
- FLOT
Fluorouracil/leucovorin, oxaliplatin, and docetaxel
- FOLFOX
Fluorouracil/leucovorin plus oxaliplatin
- HR
Hazard ratio
- IQR
Interquartile range
- LVI
Lymphovascular invasion
- OS
Overall survival
- PNI
Perineural invasion
- R0
Microscopically margin-negative resection
- SOX
S-1 plus oxaliplatin
- STROBE
Strengthening the Reporting of Observational Studies in Epidemiology
- TD
Tumor deposit
- TNM
Tumor-node-metastasis
- XELOX
Capecitabine plus oxaliplatin
- ypN
Postoperative pathological lymph node category after neoadjuvant therapy
- ypT
Postoperative pathological tumor category after neoadjuvant therapy
- ypTD
Residual tumor deposits after neoadjuvant therapy
- ypTNM
Postoperative pathological tumor-node-metastasis staging after neoadjuvant therapy
Authors' contributions
YW conceived the study, developed the study design, curated the data, performed the statistical analyses, interpreted the results, and drafted the manuscript. YT contributed to study design, methodological guidance, data verification, interpretation of the results, and critical revision of the manuscript. XL, GL, QL, and KW contributed to clinical data collection, data checking, and case verification. YZ supervised the study, provided administrative and resource support, and revised the manuscript for important intellectual content. All authors reviewed and approved the final manuscript.
Funding
This work was supported by the Natural Science Foundation of Shandong Province (ZR2023QH103 and ZR2025MS1521), the China Postdoctoral Science Foundation (2024M751571), the Qingdao Postdoctoral Science Foundation (QDBSH20240101031), and the Qingdao Outstanding Young Medical Talents Training Program (2025_000214).
Role of the funder
The funders had no role in the study design, data collection, data analysis, data interpretation, manuscript writing, or the decision to submit the manuscript for publication.
Data availability
The data analyzed in this study were obtained from retrospective clinical records at the Affiliated Hospital of Qingdao University. Because the dataset contains patient-level clinical information and is subject to institutional privacy and data-governance requirements, it is not publicly available. De-identified data may be made available from the corresponding author upon reasonable request and after approval by the relevant ethics committee and institutional authorities. Supplementary tables and figures supporting the reported analyses are included as additional files.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committee of the Affiliated Hospital of Qingdao University (approval no. QYFYWZLL50115). The requirement for informed consent was waived by the ethics committee because this was a retrospective study using anonymized clinical data.
Ethics approval number
QYFYWZLL50115.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
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.Lordick F, Carneiro F, Cascinu S, et al. Gastric cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol. 2022;33:1005–20. 10.1016/j.annonc.2022.07.004. [DOI] [PubMed] [Google Scholar]
- 3.Nappo F, Fornaro L, Pompella L, et al. Pattern of recurrence and overall survival in esophagogastric cancer after perioperative FLOT and clinical outcomes in MSI-H population: the PROSECCO Study. J Cancer Res Clin Oncol. 2023;149:6601–11. 10.1007/s00432-023-04636-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.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]
- 5.Achilli P, De Martini P, Ceresoli M, et al. Tumor response evaluation after neoadjuvant chemotherapy in locally advanced gastric adenocarcinoma: a prospective, multi-center cohort study. J Gastrointest Oncol. 2017;8:1018–25. 10.21037/jgo.2017.08.13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Woodham BL, Chmelo J, Donohoe CL, Madhavan A, Phillips AW. Prognostic significance of lymphatic, venous and perineural invasion after neoadjuvant chemotherapy in patients with gastric adenocarcinoma. Ann Surg Oncol. 2020;27:3296–304. 10.1245/s10434-020-08389-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gu L, Chen P, Su H, et al. Clinical significance of tumor deposits in gastric cancer: a retrospective and propensity score-matched study at two institutions. J Gastrointest Surg. 2020;24:2482–90. 10.1007/s11605-019-04421-8. [DOI] [PubMed] [Google Scholar]
- 8.Weiser MR. AJCC 8th edition: colorectal cancer. Ann Surg Oncol. 2018;25:1454–5. 10.1245/s10434-018-6462-1. [DOI] [PubMed] [Google Scholar]
- 9.Tan J, Yang B, Xu Z, et al. Tumor deposit indicates worse prognosis than metastatic lymph node in gastric cancer: a propensity score matching study. Ann Transl Med. 2019;7:671. 10.21037/atm.2019.10.33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Yu J, Yao R, Han N, et al. Evaluating the prognostic significance of tumor deposits in gastric cancer and strategies for their integration into the TNM staging system: a single-center retrospective study. Cell Oncol (Dordr). 2025;48:761–73. 10.1007/s13402-025-01046-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Xu R, Li Y, Zhao H, et al. Tumor deposits in gastric cancer cannot be regarded as metastatic lymph nodes: a single-center retrospective study. Eur J Surg Oncol. 2025;51:109719. 10.1016/j.ejso.2025.109719. [DOI] [PubMed] [Google Scholar]
- 12.Seretis F, Panagaki A, Triantafyllou T, et al. Tumor deposits in gastric cancer: a systematic review of the literature with meta-analysis on prevalence and prognostic implications. Ann Surg Open. 2026;7(2):e629. 10.1097/AS9.0000000000000629. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Yao R, Gao X, Yu J, et al. Tumor deposits in colorectal and gastric cancers. Med Plus. 2024;1:100025. 10.1016/j.medp.2024.100025. [DOI] [Google Scholar]
- 14.Song X, Liu K, Liao X, et al. Clinical significance of tumor deposits in gastric cancer after radical gastrectomy: a propensity score matching study. World J Surg Oncol. 2023;21:325. 10.1186/s12957-023-03208-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhou M, Yang W, Zou W, et al. Prognostic significance of tumor deposits in radically resected gastric cancer: a retrospective study of a cohort of 1915 Chinese individuals. World J Surg Oncol. 2022;20:304. 10.1186/s12957-022-02773-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Pai RK, Pai RK. Pathologic assessment of gastrointestinal tract and pancreatic carcinoma after neoadjuvant therapy. Mod Pathol. 2018;31:4–23. 10.1038/modpathol.2017.87. [DOI] [PubMed] [Google Scholar]
- 17.Li Z, Wang Y, Shan F, et al. ypTNM staging after neoadjuvant chemotherapy in the Chinese gastric cancer population: an evaluation on the prognostic value of the AJCC eighth edition cancer staging system. Gastric Cancer. 2018;21:977–87. 10.1007/s10120-018-0830-1. [DOI] [PubMed] [Google Scholar]
- 18.Li W, Shao X, Ma F, et al. Total lymph node yield is associated with prolonged survival after neoadjuvant chemotherapy in ypN0 gastric cancer patients. Oncol Res Treat. 2023;46:287–95. 10.1159/000531436. [DOI] [PubMed] [Google Scholar]
- 19.Xu R, Zhang Y, Zhao J, Chen K, Wang Z. Prognostic value of tumor deposits in lymph node-negative gastric cancer: a propensity score matching study. Eur J Surg Oncol. 2023;49:845–52. 10.1016/j.ejso.2022.12.004. [DOI] [PubMed] [Google Scholar]
- 20.von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Ann Intern Med. 2007;147:573–7. 10.7326/0003-4819-147-8-200710160-00010. [DOI] [PubMed] [Google Scholar]
- 21.Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5:649–55. [PubMed] [Google Scholar]
- 22.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83. 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 23.Amin MB, Greene FL, Edge SB, et al. The eighth edition AJCC Cancer Staging Manual: continuing to build a bridge from a population-based to a more personalized approach to cancer staging. CA Cancer J Clin. 2017;67:93–9. 10.3322/caac.21388. [DOI] [PubMed] [Google Scholar]
- 24.Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–81. 10.1080/01621459.1958.10501452. [DOI] [Google Scholar]
- 25.Cox DR. Regression models and life-tables. J R Stat Soc Series B Stat Methodol. 1972;34:187–202. 10.1111/j.2517-6161.1972.tb00899.x. [DOI]
- 26.Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69:239–41. 10.1093/biomet/69.1.239. [DOI] [Google Scholar]
- 27.Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996;15:361–87. 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4. [DOI] [PubMed] [Google Scholar]
- 28.Heagerty PJ, Lumley T, Pepe MS. Time-dependent ROC curves for censored survival data and a diagnostic marker. Biometrics. 2000;56:337–44. 10.1111/j.0006-341x.2000.00337.x. [DOI] [PubMed] [Google Scholar]
- 29.Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Mak. 2006;26:565–74. 10.1177/0272989X06295361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Marano L, Ambrosio MR, Resca L, et al. The percentage of signet ring cells is inversely related to aggressive behavior and poor prognosis in mixed-type gastric cancer. Front Oncol. 2022;12:897218. 10.3389/fonc.2022.897218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Roviello F, Marano L, Ambrosio MR, et al. Signet ring cell percentage in poorly cohesive gastric cancer patients: a potential novel predictor of survival. Eur J Surg Oncol. 2022;48(3):561–9. 10.1016/j.ejso.2021.09.003. [DOI] [PubMed] [Google Scholar]
- 32.Wilson MK, Karakasis K, Oza AM. Outcomes and endpoints in trials of cancer treatment: the past, present, and future. Lancet Oncol. 2015;16(1):e32–42. 10.1016/S1470-2045(14)70375-4. [DOI] [PubMed] [Google Scholar]
- 33.Liu H, Wang Y, Qi C, et al. Evaluation of event-free survival surrogating overall survival as the endpoint in neoadjuvant clinical trials of gastroesophageal adenocarcinoma. Front Oncol. 2022;12:835389. 10.3389/fonc.2022.835389. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Additional File 1. Supplementary tables and supplementary figure legends (.docx). Supplementary Tables S1–S6 and legends for Supplementary Figs. S1–S2.
Additional File 2. Supplementary Fig. S1 (.png). Exploratory ypN-stratified analysis of ypTD and DFS.
Additional File 3. Supplementary Fig. S2 (.png). Within-cohort exploratory model-performance comparison between the base pathological model and the ypTD-extended model for 36-month DFS event risk.
Data Availability Statement
The data analyzed in this study were obtained from retrospective clinical records at the Affiliated Hospital of Qingdao University. Because the dataset contains patient-level clinical information and is subject to institutional privacy and data-governance requirements, it is not publicly available. De-identified data may be made available from the corresponding author upon reasonable request and after approval by the relevant ethics committee and institutional authorities. Supplementary tables and figures supporting the reported analyses are included as additional files.