BackgroundGastrectomy with lymphadenectomy is the standard of care for noncurative endoscopic resection (ER) (endoscopic curability C2; eCuraC2) for early gastric cancer (EGC); however, this strategy may be excessive for most patients because few patients have lymph node metastasis (LNM). In addition, whether EGCs with undifferentiated components (undiff-components) in the eCuraA group are suitable for ER remains controversial. The aim of this study was to stratify the eCuraC2 group according to LNM risk, and to verify the safety of conservative treatment for patients with undiff-components of the eCuraA group.MethodsWe retrospectively analyzed 272 patients with submucosal invasion or with undiff-components who underwent initial ER for EGC. These specimens were classified into eCura A, B, C1, and C2 groups according to the Japanese Gastric Cancer Association (JGCA) guidelines, and the rate of LNM in each group was analyzed. The key risk factors were identified by analyzing the correlation between different combinations of risk factors and LNM, and according to the LNM risk, further graded the eCuraC2 group.ResultsAmong the 162 eCuraC2 patients, 9 (5.6%) had LNM. But no patients in the other groups, including all 57 patients of the eCuraA group (T1a, < 2 cm in diameter and no ulceration) with undiff-components, had LNM. A tumor diameter >3 cm (1.7% vs 12.2%, P = .005), positive for vertical margins (1.6% vs 20.0%, P < .001), submucosal invasion (≥500 μm) (0.7% vs 6.6%, P = .012), undiff-components type dominance (0% vs 11.9%, P < .001) and lymphovascular infiltration (LVI) (1.3% vs 16.7%, P < .001) were significantly correlated with LNM. When the patients in the eCuraC2 group were divided into 4 groups based on the presence of LVI and undiff-components, the LNM rate in each group was 0/81 patients (0%) in the LVI (-) undiff-components (-) group, 3/45 patients (6.7%) in the LVI (-) undiff-components (+) group, 0/15 patients (0%) in the LVI (+) undiff-components (-) group, and 6/21 patients (28.6%) in the LVI (+) undiff-components (+) group. Finally, based on these 2 factors, eCura C2 patients were classified into 3 LNM risk grades: low (LVI (-) undiff-components (-), LNM 0%), intermediate (LVI (+) or undiff-components (+), LNM 5%), and high (LVI (+) undiff-components (+), LNM 28.6%).ConclusionBased on LVI and histological differentiation, eCuraC2 patients were classified into 3 LNM risk grades, and approximately half of the eCuraC2 patients were reclassified into the low-risk group. No LNM was found in patients in the eCuraA group with undiff-components.

Self-supervised exceptional prototypical network for few-shot grading of gastric intestinal metaplasia.

Aberrant expression of p53 and elevated Ki-67 proliferation index have been associated with tumor progression and recurrence; however, their prognostic value in postoperative early gastric cancer (EGC) remains to be fully established. This study aimed to develop and internally validate a nomogram integrating these biomarkers for individualized recurrence risk prediction. We retrospectively analyzed 536 EGC patients from a single institution between January 2017 and October 2018, with follow-up through June 2025. Patients were randomly divided into training (n = 375) and validation (n = 161) cohorts. Univariate and multivariate Cox proportional hazards regression analyses were performed to identify independent prognostic predictors. Model performance was assessed using time-dependent AUC, integrated Brier score, calibration curves, and decision curve analysis. Five independent factors were identified: Helicobacter pylori infection (HR = 1.83), Ki-67 > 30% (HR = 4.28), lymphovascular invasion (HR = 1.91), perineural invasion (HR = 4.37), and p53 (reference: mutant, HR = 0.37). The nomogram achieved good discriminative ability (1-year AUC: 0.861 training, 0.878 validation; 5-year AUC: 0.808, 0.788). Nomogram-based risk stratification demonstrated significant differences in survival outcomes (log-rank P < 0.001). This nomogram demonstrates potential utility for recurrence risk stratification in postoperative EGC patients. However, the model requires external validation in independent, multi-center cohorts before clinical implementation. The single-center retrospective design and relatively short follow-up represent important limitations.

Fibroblasts associated to carcinomas express fibroblast activation protein (FAP). FAP-targeted imaging with Positron Emission Tomography (PET) with 68Ga-FAP inhibitors (FAPi-PET), is a pan-tumoral imaging that enables in-vivo visualization of the tumor stroma and detection of neoplastic foci in various cancers. The primary aim was to evaluate the diagnostic performance of FAPi-PET in detecting locally advanced gastric cancers (LAGC), including the evaluation of peritoneal disease (PD). Thirty consecutive patients with LAGC were included in this prospective study between Nov-2023 and Nov-2024 at the European Institute of Oncology in Milano. All patients had an indication to staging laparoscopy and underwent FAPi-PET within 1 week prior to surgery. Significant gastric expression of FAP (mean SUVmax 13,12 ± 5,9 SD) was detected in 29 FAPi-PET scans. The only GC without detectable gastric uptake was an early GC. The sensitivity of FAPi-PET for the identification of primary tumor was 96,7%. PD was suspected based on FAPi-PET imaging in 2 patients (mean SUVmax of 7.55 ±2.19 SD). At staging laparoscopy PD was observed in only 1 patient. Conversely, in 2 cases with a negative preoperative FAPi-PET, PD was observed during laparoscopy,. The accuracy of FAPi-PET for PD was 90% (sensitivity 33,3%, specificity 96,3%). FAPi-PET showed a high detection rate for LAGC, even in cases with diffuse histology. This technique appears to provide high specificity for PD. These promising preliminary findings, support further validation of FAPi-PET in upper-GI malignancies to evaluate its impact on clinical practice for staging and treating LAGC.

Gastric cancer (GC) remains a global health burden. While international guidelines share consensus, variations exist in disputed issues. The 2025 Taiwan Consensus and Management Guidelines for Gastric Cancer had just been released. We compared the key recommendations with established international guidelines. A multidisciplinary taskforce addressed key questions and recommendations using modified Delphi method and evidence-based approaches. The comparative review aimed to elucidate similarities and differences among guidelines from Taiwan, Japan, South Korea, China, Europe, and the US. Guidelines converge on absolute endoscopic resection criteria for early GC but differ in extended indications and perioperative approaches for locally advanced disease. Heterogeneity exists in biomarker assessment protocols, cutoff thresholds, and companion diagnostics. For metastatic disease, consensus exists on anti-HER2, anti-VEGF, immunotherapy, and biomarker-driven strategies, though oligo-metastatic definitions and intraperitoneal chemotherapy indications remain controversial. Optimal GC management requires integrating global evidence with regional contexts. The comparative review addresses heterogeneity among international guidelines, which helps harmonize management strategies as therapeutic standards evolve.

Ultrasensitive detection of gastric cancer biomarker using tetrahedral DNA nanostructure-functionalized floating-gate carbon nanotube field-effect transistor biosensor.

In this study, we investigated the diagnostic performance of magnifying endoscopy (ME) with third-generation narrow-band imaging (3G-NBI) for early gastric cancer (EGC), primarily in patients with Helicobacter pylori eradication. This was a post hoc analysis of a multicenter, randomized trial comparing 3G-NBI, texture and color enhancement imaging, and white-light imaging in gastric neoplasm (GN) detection. For all detected lesions, the endoscopic diagnosis of ME using 3G-NBI was compared with the pathological diagnosis. The primary analyses focused on the sensitivity and specificity of ME with 3G-NBI for EGC or non-EGC. The diagnostic performance was analyzed according to confidence level, macroscopic type, lesion size, and H. pylori infection status. This study included 901 patients; 228 suspected GN lesions in 187 patients were analyzed. The lesions were diagnosed with EGC in 62 (27 with high confidence) and non-EGC in 166 (91 with high confidence) patients using ME with 3G-NBI and pathologically diagnosed as EGC in 61 and non-EGC in 167 patients. The overall diagnostic performance was sensitivity and specificity of 70.5% and 88.6%, respectively. The diagnostic performance of each category was as follows: (1) confidence level (high/low); sensitivity 78.1%/62.1%, specificity 97.7%/79.0%; (2) macroscopic type (elevated/flat or depressed); sensitivity 84.6%/66.7%, specificity 96.7%/86.9%; (3) lesion diameter (< 10mm/ ≥ 10mm); sensitivity 65.7%/76.9%, specificity 88.6%/88.9%; (4) H. pylori infection status (uninfected/previously infected/currently infected); sensitivity 50.0%/71.0%/86.7%, specificity 91.3%/88.2%/88.9%. The diagnostic performance of ME with 3G-NBI for EGC was acceptable, primarily in patients with H. pylori eradication. This trial was registered in jRCT (Identifier jRCT1032210213).

ObjectiveThis study aimed to propose and validate a new and practical lymph node (LN) staging strategy to mitigate staging migration due to examined LN (ELN) retrieval and improve the accuracy of prognostic evaluation for early gastric cancer (EGC) patients.BackgroundEGC patients often face staging inaccuracies due to inadequate ELNs, as there is no clear standard for ELN retrieval requirements.MethodsFrom an initial cohort of 7001 EGC patients across fifteen large institutions in China, 6566 eligible patients were included as the training and validation cohort after applying predefined exclusion criteria. We analyzed these data to determine the optimal cutoff value for the number of ELNs, construct a prediction model, and propose a new LN staging method distinct from that of the latest AJCC guidelines. Subsequently, 2094 patients (from 2326 with survival data available in the multicenter dataset) and 1944 patients (from 5262 initially retrieved from the SEER database after exclusions) were included as independent test cohorts to evaluate model performance, including the correlation between ELNs and metastatic LNs (MLNs) and survival differences.ResultsThis study found that ≤20 ELNs were inadequate for accurate LN evaluation in EGC patients. In the test cohort, MLNs was positively associated with ELNs. A model was constructed, and accurate MLNs could be displayed after correction for patients with inadequate ELNs. Survival analysis revealed significant differences between patients with different pNM (pN modified) stages in both the multicenter test cohort and the SEER test cohort.ConclusionModel-corrected pNM staging may improve staging migration and survival prediction than the AJCC staging, showing great clinical applicability for EGC patients with ELN ≤ 20.

Identification of potential prognostic biomarkers and therapeutic targets for gastric cancer (GC) remains crucial. We conducted an integrative analysis, utilizing single-cell transcriptome datasets concerning early GC from GEO database, along with DNA methylation data from UCSC Data Center, to screen for dysregulated genes in early GC. Our findings revealed that IFITM2 expression was up-regulated in early GC epithelial cells and its promoter methylation was down-regulated. Elevated IFITM2 levels were further confirmed in GC cell lines and GC tissues from TCGA patients. According to the TCGA database, a high level of IFITM2 was correlated with a poor prognosis in GC patients. Both Depmap single-gene CRISPR knockout screens across 35 GC cell lines and in house siRNA-mediated IFITM2 knockdown experiments demonstrated that IFITM2 silencing inhibited cell proliferation. GSEA analysis indicated that high levels of IFITM2 was enriched in TNFα signaling via NF-κB. Furthermore, STRING protein-protein interaction analysis suggested a functional link between IFITM2 and NF-κB. Finally, IFITM2 silencing significantly suppressed NF-κB activity and the expression of NF-κB targeted genes, including the anti-apoptotic genes BCL-2, BCL-XL, the pro-proliferation genes TNFα, CCL2, CDK1, and the pro-metastasis gene COX-2. Therefore, our study suggests that IFITM2 may promote the initiation and proliferation of GC by activating NF-κB signaling pathway, positioning IFITM2 as a prognostic biomarker and therapeutic target for GC.

Helicobacter pylori (H. pylori) is a globally prevalent gastric pathogen whose genetic diversity and virulence contribute to the varied clinical outcomes. Whole-genome sequencing has revolutionized microbial genomics by allowing high-resolution analyses to obtain insights into the evolution of pathogenic organisms; however, few studies have specifically focused on strains from Korea, where H. pylori infection and gastric cancer are highly prevalent. This study aimed to investigate the phylogenetic characteristics and virulence factor profiles of Korean H. pylori strains using whole-genome sequencing. Twelve H. pylori strains from Korean patients with chronic gastritis or early gastric cancer were subjected to whole-genome sequencing. Draft genomes were assembled de novo, and multi-locus sequence typing was conducted to determine the population structure. Pan-genome analysis and virulence factor profiling were also performed. All strains were clustered within the East Asian (hspEAsia) population in the phylogenetic tree. The pan-genome analysis identified 2574 gene clusters, with 1161 core genes (45.1%), 519 shell genes (20.2%), and 894 cloud genes (34.7%). The cytotoxin-associated gene pathogenicity island (cagPAI) was intact in 11/12 strains, and all strains harbored East Asian-type ABD Glu-Pro-Ile-Tyr-Ala (EPIYA) motifs. Notably, cag2 was not detected in all strains and cagY exhibited significant sequence variation. Vacuolating cytotoxin A (vacA) genotyping identified s1c/i1/m1 as the predominant type (66.7%), followed by s1a/i1/m1 (25.0%) and s1c/i1/m2 (8.3%). The Korean H. pylori strains demonstrated typical East Asian characteristics, and intact cagPAI and virulent vacA genotypes were highly prevalent, providing essential genomic insights into H. pylori pathogenesis in the Korean populations.

Association between additional operation timing and perioperative outcomes after noncurative endoscopic submucosal dissection for early gastric cancer.

Early gastric carcinogenesis progresses from gastritis through intestinal metaplasia to early gastric cancer, yet the spatial origin of malignant transformation and subtype divergence remains unclear. Here we show, using spatial transcriptomics across human gastric tissues spanning disease progression, that an OLFM4-positive transitional subset emerges at the interface between gastric isthmus/glands and proliferative intestinal metaplasia crypts, consistent with an early step in intestinalization. From these crypts, transcriptional trajectories diverge toward distinct intestinal differentiation programs. Notably, proliferative crypt regions display the closest transcriptional similarity to early gastric cancer and are enriched for proliferation and DNA-repair pathways, supporting their role as a premalignant niche. We further identify increased activity of USF2-associated regulatory networks shared between crypt and cancer regions. Further stratification by Lauren subtype reveals that intestinal-type gastric cancer is characterized by a CDX2-dominated program, whereas diffuse-type gastric cancer exhibits developmental and stemness-related networks with SPP1 and CD44 co-occurring. Together, these findings highlight the intestinal metaplasia crypt as a premalignant niche and provide spatially resolved candidate markers for early detection and mechanistic investigation.

Endoscopic submucosal dissection (ESD) is a minimally invasive treatment for early gastric cancer that enables the en bloc resection of large lesions. However, it is associated with a significant risk of delayed bleeding, particularly in patients receiving antithrombotic therapy. Delayed bleeding after ESD causes anemia and serious systemic complications, including cardiovascular events. Here, we report a rare case of Takotsubo syndrome triggered by delayed bleeding after gastric ESD that culminated in cardiac arrest. A man in his 70s who underwent gastric ESD experienced delayed bleeding six days after the treatment, requiring emergency endoscopic hemostasis. The following day, the patient developed sudden ventricular fibrillation without prior chest symptoms. After resuscitation with electrical defibrillation, left ventriculography revealed apical ballooning consistent with Takotsubo syndrome. With intensive care, cardiac function recovered, and the patient was discharged ambulatory. Takotsubo syndrome is typically induced by stress, and often triggers psychological and physical stimuli. Although most patients present with chest pain or discomfort, our patient showed no preceding symptoms and developed sudden arrhythmia and cardiac arrest. This case highlights the need for aggressive bleeding prevention and enhanced post-ESD cardiac surveillance to mitigate rare but potentially fatal complications.

Endoscopic examination plays a crucial role in the secondary prevention of gastric cancer. However, its effectiveness is often limited by the subjective nature of assessment and the risk of missing neoplastic lesions. The use of deep learning (DL) and machine learning (ML) models allows for a significant improvement in the quality of endoscopic diagnosis of early gastric cancer. Real-time systems increase the sensitivity of the examination and reduce skill-based differences between endoscopists. They also enable precise assessment of tumor invasion depth, which is essential for qualifying patients for minimally invasive treatment. Explainable artificial intelligence (XAI) methods and ML algorithms are gaining increasing importance as they enhance the transparency of the diagnostic process. Despite promising data, the full implementation of AI systems in endoscopy faces challenges such as susceptibility to imaging artifacts, the risk of model overfitting, and regulatory hurdles. A key issue remains the development of representative, multicenter training image databases and the conduct of prospective clinical trials, which may revolutionize standards of care. The aim of this article is to discuss the role of AI-assisted gastroscopy in the prevention of early gastric cancer, with particular emphasis on its potential in population-based screening programs. The paper presents current clinical data and analyzes the benefits and challenges associated with implementing these technologies into routine clinical practice.

ASO Visual Abstract: Establishment and Validation of Prediction Models for Lymph Node Metastasis and Long-Term Survival in T1b Early Gastric Cancer Patients: A Retrospective Cohort Study with 10-Year Follow-up.

Non-invasive detection of gastric cancer biomarker 4-HBA via a ZIF/CDs-based fluorescent sensor.

T1b gastric cancer, characterized by tumor invasion into the submucosa, presents a therapeutic dilemma regarding the need for adjuvant therapy owing to varying rates of lymph node metastasis (LNM). This study aimed to develop and validate comprehensive nomogram models for predicting LNM risk and long-term survival outcomes in patients with T1b gastric cancer. A retrospective cohort study was conducted on 362 patients with pathologically confirmed T1b gastric cancer who underwent radical gastrectomy with D2 lymph node dissection at a single institution between 2014 and 2024. Patients were stratified into LN+ (lymph node-positive) and LN- (lymph node-negative) groups. Multivariate logistic regression identified independent risk factors for LNM, while Cox proportional hazards models assessed prognostic factors for overall survival (OS) and recurrence-free survival (RFS). Nomogram models were constructed and internally validated using bootstrap resampling. Among 362 patients, 92 (25.4%) had LNM. Independent predictors of LNM included tumor size ≥ 3 cm (odds ratio [OR] 2.84, 95% confidence interval [CI]: 1.62-4.98, p < 0.001), lymphovascular invasion (OR 4.21, 95% CI: 2.45-7.23, p < 0.001), poor differentiation (OR 3.12, 95% CI: 1.78-5.47, p < 0.001), and perineural invasion (OR 2.56, 95% CI: 1.23-5.32, p = 0.012). The LNM prediction nomogram showed excellent discrimination (area under the curve [AUC] 0.843, 95% CI: 0.801-0.885) and calibration. The integrated survival nomogram incorporating LNM risk demonstrated superior predictive performance for 5-year OS (C-index 0.782) compared with traditional staging (C-index 0.681). Decision curve analysis confirmed clinical utility across relevant threshold probabilities. Our validated nomogram models provide accurate individualized predictions for LNM risk and long-term survival in patients with T1b gastric cancer, potentially guiding personalized treatment decisions regarding adjuvant therapy and extent of lymphadenectomy.

Artificial intelligence (AI)-assisted endoscopy has been developed for the early detection of upper gastrointestinal cancer; however, its clinical effectiveness remains insufficiently evaluated. This study assessed its effectiveness in a health screening facility. This retrospective cohort study compared AI-assisted and non-AI-assisted upper gastrointestinal endoscopy at Omiya City Clinic, Japan (April 2021-March 2024). Participants who underwent endoscopy between April 2021 and March 2023 were classified as the non-AI group, while those examined between April 2023 and March 2024 comprised the AI group. The AI-assisted system was introduced in April 2023. The primary outcome was cancer detection rate (CDR), with secondary outcomes including biopsy rate and positive predictive value (PPV). Propensity score matching (PSM) was performed for age, sex, alcohol consumption, smoking, Helicobacter pylori infection history, endoscopist experience, and prior-year endoscopy to minimize bias. In total, 17,662 were included in the AI group and 32,318 in the non-AI group. PSM created 17,662 matched pairs. In the AI group, the CDR for gastric cancer (GC) was significantly higher compared to the non-AI group (0.10% vs. 0.03%, p < 0.05). The biopsy rate was slightly higher in the AI group, with no significant difference, whereas the PPV of biopsy for gastric cancer and esophageal cancer was significantly increased (4.84% vs. 2.16%, p < 0.05). In a clinical screening setting, AI-assisted endoscopy significantly improved the CDR of GC and enhanced the PPV of biopsies. These findings highlight AI-assisted endoscopy as a valuable tool for early GC diagnosis in screening environments.

The muscularis mucosae (MM) is a thin layer of smooth muscle in the gastrointestinal tract that may act as a barrier against carcinoma invasion; however, its characteristics in gastric cancer remain unclear. This study aimed to precisely measure MM thickness in early gastric cancer and assess its association with anatomical location, mucosal atrophy, Helicobacter pylori (H. pylori) infection status and cancer invasion. We analyzed 62 cases of early gastric cancer treated by endoscopic submucosal dissection (ESD), including 38 mucosa-confined (pT1a) and 24 submucosal-invasive (pT1b) cancers. The MM was visualized using desmin immunohistochemistry, and whole-slide imaging combined with digital image analysis enabled accurate quantification of MM thickness. The relationships between MM thickness and depth of invasion, tumor size, clinical parameters, anatomical location, mucosal atrophy and H. pylori infection were evaluated. No significant association was observed between MM thickness and cancer invasion, tumor size, or other clinical parameters. MM thickness was significantly greater in the pyloric antrum than in the gastric body and was thicker in the lesser curvature than in the greater curvature. In addition, MM thickness was significantly greater in cases with mucosal atrophy than in those without. Given the association between the presence of mucosal atrophy and increased MM thickness, together with the fact that the observed anatomical differences correspond to the distribution and progression of chronic gastritis, it is reasonable to speculate that chronic gastritis contributes to thickening of the MM.

Diagnosis of undifferentiated-type-predominant mixed-type early gastric cancer (UM-EGC) remains challenging due to its complex histological features and overlapping characteristics with other gastric cancer types. We aimed to develop a novel nomogram based on clinicopathological and endoscopic features and validate its performance in predicting UM-EGC prior to endoscopic treatment. In this retrospective single-center study, 808 patients with early gastric cancer (EGC) who underwent curative endoscopic submucosal dissection were included. Among them, 493 were assigned to the training cohort and 84 to the external validation cohort. Clinicopathological characteristics and endoscopic features were compared between differentiated EGC and UM-EGC using logistic regression analysis. A predictive nomogram was constructed and evaluated. Multivariable regression analysis identified open-type atrophic gastritis (O1-O3) (odds ratio [OR] 0.25, 95% confidence interval [CI] 0.08-0.82), IIb (OR 9.72, 95% CI 3.01-31.35), IIc (OR 7.75, 95% CI 2.81-21.39), discolored lesion (OR 4.12, 95% CI 1.57-10.80), horizontal location at the greater curvature (OR 2.98, 95% CI 1.15-7.75) or anterior wall (OR 2.91, 95% CI 1.26-6.74), and previous H. pylori eradication (OR 0.23, 95% CI 0.09-0.55) as independently associated with UM-EGC. UM-EGC was also more susceptible to metachronous cancer (OR 5.50, 95% CI 1.30-23.21). The nomogram demonstrated good discriminative ability with an area under the receiver operating characteristic curve of 0.83 (95% CI 0.77-0.87) in the training cohort and 0.82 (95% CI 0.69-0.98) in the external validation cohort. This nomogram comprising clinicopathological and endoscopic features may assist in the preoperative prediction of UM-EGC risk. The clinical trial registration number for patient source in this study is ChiCTR1800017117.