Molecular characteristics of proximal and distal esophagogastric junction adenocarcinoma

Introduction

Esophagogastric junction adenocarcinomas (EGJA), arising at the junction of the lower esophagus and proximal stomach, have increased in incidence over recent decades, which has been accompanied by persistently poor clinical outcomes (1). The Siewert system is used to standardize the management of EGJA, classifying it into the following three subtypes according to anatomical the tumor location: type I, lower esophagus—1–5 cm above the anatomical cardia; type II, true cardia—≤2 cm from the anatomical cardia; and type III, gastric cardia—2–5 cm below the anatomical cardia (2,3). Despite its widespread adoption, the Siewert system faces challenges in guiding therapeutic decisions due to overlapping molecular and clinical features across subtypes.

The eighth edition of the tumor-node-metastasis (TNM) staging system for esophageal cancer and gastric cancer issued by American Joint Committee on Cancer (AJCC) and Union for International Cancer Control (UICC) further complicates this landscape by determining EGJA staging based on tumor epicenter location as follows: tumors within 2 cm of the cardia are classified according to esophageal cancer staging, while those beyond 2 cm, even when invading the cardia, are classified according to gastric cancer staging (4,5). Despite being implemented globally in 2018, this dual-classification approach lacks robust molecular validation, and thus its clinical relevance remains controversial.

Advances in genomic technologies, particularly next-generation sequencing (NGS), have begun to reveal the molecular underpinnings of EGJA (6,7). However, critical gaps persist in three areas: (I) genetic heterogeneity—distinct genomic profiles of Siewert subtypes (proximal vs. distal EGJA) remain poorly defined; (II) cross-cancer comparisons—the overlap in molecular characteristics and differences between EGJA, gastric adenocarcinoma (GCA), and esophageal squamous cell carcinoma (ESCC) remain underclarified; and (III) clinical translation—prognostic biomarkers and evidence supporting a unified staging system for EGJA are lacking. To address these deficiencies, we performed targeted NGS of 450 cancer-related genes recurrently mutated in solid tumors, analyzing 198 cases of EGJA (58 proximal and 140 distal), 42 cases of GCA, and 36 cases of ESCC. The objectives of the studies were as follows: (I) define high-frequency mutations by identifying common genomic alterations in EGJA and subtype-specific drivers (proximal vs. distal); (II) characterize molecular landscapes by systematically comparing EGJA with GCA and ESCC to clarify their genomic relationships; (III) identify prognostic biomarkers by correlating mutations with survival outcomes to guide risk stratification; and (IV) assess the staging rationale by determining whether EGJA’s molecular distinctiveness justifies a dedicated TNM system. By integrating genomic profiling with clinical data, this study brings molecular insights to the practice of oncology, offering a framework to refine the classification, prognostication, and therapeutic precision related to EGJA. We present this article in accordance with the REMARK reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0027/rc).

Methods

Patients and ethical considerations

In this retrospective cohort study, patients with histologically confirmed EGJA, ESCC, or GCA were consecutively enrolled from Shandong Provincial Hospital Affiliated to Shandong First Medical University (Jinan, China) between January 2013 and December 2015. EGJA cases were subclassified based on tumor anatomical location: (I) proximal EGJA was considered present if the tumor epicenter was located within 2 cm of the EGJ; (II) distal EGJA was considered present if the tumor epicenter extended >2 cm beyond the EGJ. Notably, ESCC cases were included as a molecular comparator from the adjacent anatomical region to better delineate the genomic features specific to EGJA. Tumor tissues were prospectively archived according to these criteria. Postoperative follow-up was systematically conducted for the collection of clinicopathological data, with overall survival (OS) serving as the primary endpoint. OS was defined as the interval from histopathological diagnosis to death from any cause; data from patients lost to follow-up or alive at the study cutoff date (December 2023) were censored. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study protocol was approved by the Institutional Ethics Review Committee of Shandong Provincial Hospital Affiliated to Shandong First Medical University (LCYJ: No. 2019-148), and written informed consent was obtained from all participants.

Sample preparation and targeted NGS

Histologic sections from all enrolled patients underwent independent pathological review by two board-certified pathologists to confirm diagnostic accuracy prior to sample processing. Tissue samples were formalin fixed paraffin-embedded (FFPE) following standardized protocols (8,9) at accredited clinical hospitals and subsequently transferred under controlled conditions to a College of American Pathologists (CAP)/Clinical Laboratory Improvement Amendments (CLIA)-certified laboratory in OrigiMed (Shanghai, China) for genomic analysis. Samples consisting of more than 20% tumor cells were selected for further DNA extraction. Genomic DNA was isolated with the QIAamp DNA FFPE Tissue Kit and QIAamp DNA Blood Midi Kit (Qiagen, Hilden, Germany) in strict adherence to the manufacturer’s protocols. DNA concentration was normalized to 20–50 ng/µL to ensure optimal input for subsequent library preparation and targeted NGS.

Identification of genomic alterations

Genomic profiling was performed with the YuanSu450 gene panel (OrigiMed), with all the coding exons of the 450 cancer-related genes frequently mutated in solid tumors being targeted. DNA libraries were sequenced on the NextSeq 500 platform (Illumina, San Diego, CA, USA) with an average depth of 800×. Bioinformatic analyses were conducted according to previously described procedures (10,11): (I) for variant detection, single-nucleotide variants (SNVs) were identified via MuTect (v. 1.7) under the default parameters, insertion-deletion mutations (Indels) were identified via PINDEL (v. 0.2.5), and copy number variations (CNVs) were analyzed via Control-FREEC (v. 9.7), with window =50,000 and step =10,000. Gene fusions were identified with an in-house pipeline, validated by manual inspection in Integrative Genomics Viewer (Broad Institute of MIT and Harvard, Cambridge, MA, USA). (II) For functional annotation, all variants (e.g., missense, nonsense, and splice site) were annotated for functional impact via SnpEff 3.0 and cross-referenced with the Catalogue of Somatic Mutations in Cancer (COSMIC) and ClinVar databases.

Statistical analysis

Statistical analyses were performed with SPSS v. 22.0 statistical software (IBM Corp., Armonk, NY, USA). Associations of categorical variable were examined with the Fisher exact test, which is particularly suitable for small sample sizes or sparse data configurations. Survival analysis for OS outcome included the use of Kaplan-Meier methodology to estimate time-to-event probabilities, with between-group differences assessed with the two-sided log-rank test. An a priori significance threshold was established at α=0.05, with all reported P values representing two-tailed probabilities.

Results

Patient characteristics

This retrospective cohort enrolled 198 patients with EGJA, with a median age of 63 years (range, 39–88 years). The cohort was predominantly male, with 168 males (84.8%) and 30 females (15.2%). (I) Clinicopathological stratification revealed the following distributions for tumor stage: stage I, 5.6% (n=11); stage II, 21.7%; (n=43); stage III, 48.5% (n=96); and stage IV, 24.2% (n=48). (II) The distribution for histological differentiation was as follows: poorly differentiated, 32.3% (n=64); moderately differentiated, 48.0% (n=95); well-differentiated, 1.0% (n=2); and undetermined differentiation, 18.7% (n=37). (III) Finally, the distribution of Lauren classification was as follows: diffuse-type, 9.1% (n=18); intestinal-type, 21.7% (n=43), indeterminate/mixed type, 12.1% (n=24); and unclassified, 57.1% (n=113).

Anatomic categorization followed standardized criteria: distal EGJA was defined as tumors with epicenters >2 cm from the esophagogastric junction, while proximal EGJA included tumors within 2 cm of the gastric cardia. Accordingly, 70.7% (n=140) were classified as distal EGJA and 29.3% (n=58) as proximal EGJA. Additionally, 42 patients with GCA (median age 63 years; range, 25–80 years) and 36 patients with ESCC (median age 62 years; range, 43–78 years) were included as comparative cohorts. The comprehensive clinical and pathological data are summarized in Table 1.

Genomic landscape of EGJA

Among the 198 patients with EGJA who underwent comprehensive genomic profiling, high-quality sequencing data were successfully analyzed in 197 patients. Notably, one case (0.5%) exhibited no detectable pathogenic variants despite meeting all quality control thresholds, potentially reflecting biological characteristics such as ultra-low tumor mutational burden or chromosomally stable phenotypes. A total of 1,788 somatic alterations from 431 genes were characterized, comprising the following mutation subtypes: SNVs/Indels (n=937, 52.4%), gene amplifications (n=485, 27.1%), truncating mutations (n=266, 14.9%), fusion/rearrangement events (n=77, 4.3%), and homozygous deletions (n=23, 1.3%). The predominant driver alterations occurred included in TP53 (74%), CCNE1 (14%), ERBB2 (12%), FAT3 (11%), ARID1A (11%), PIK3CA (10%), and SPTA1 (10%) (Figure 1). As expected, gene amplification was the primary alteration mechanism for CCNE1 and ERBB2, consistent with the known biology of these genes. Further analysis revealed amplification-driven oncogenic activation in multiple pathways, including cell cycle regulators (CDK6 and CDKN2A), growth factor signaling (EGFR, MET, and MDM2), angiogenesis mediators (VEGFA), 11q13 amplification genes (CCND1, FGF3, FGF4, and FGF19), and immune modulation (BLK). The structural variant distribution and allele-specific frequencies are detailed in Figure 1.

Figure 1 Significantly mutated genes in EGJA. The distal and proximal groups are shown, with the matrix of mutations being colored according to the type of mutation. Each column denotes an individual tumor, and each row represents a gene. Blue represents the distal EGJA, and green represents the proximal EGJA. EGJA, esophagogastric junction adenocarcinoma.

Genomic analysis comparing EGJA with GCA/ESCC

To determine molecular distinctions between EGJA and GCA/ESCC, we performed integrative genomic profiling of the EGJA cohort, along with 40 GCA cases and 36 ESCC cases.

In contrast to both GCA and ESCC, EGJA exhibited significantly lower mutation frequencies for the genes listed in Figure 2A,2B, respectively (P<0.05 in both comparisons).

Figure 2 Comparative analysis of the frequently mutated genes (mutation frequency >10%) in EGJA, GCA, and ESCC. (A) Significantly lower mutation frequencies in EGJA vs. GCA (P<0.05); (B) significantly lower mutation frequencies in EGJA vs. ESCC (P<0.05). EGJA, esophagogastric junction adenocarcinoma; ESCC, esophageal squamous cell carcinoma; GCA, gastric adenocarcinoma.

Site-specific molecular profiling of EGJA subtypes

To characterize the anatomic heterogeneity of EGJA, we performed comparative genomic characterization of the proximal (n=58) and distal (n=140) subtypes.

Both subtypes exhibited conserved alterations in the following core tumorigenic genes: TP53 (distal: 76%; proximal: 71%), ARID1A (distal: 10%; proximal: 12%), CCNE1 (distal: 16%; proximal: 10%), and PIK3CA (distal: 9%; proximal: 12%).

Distal tumors demonstrated enrichment in FGFR2 (15.8%), FAT3 (12%), ERBB2 (11%), SPTA1 (11%), CDK6 (10%), LRP1B (10%), and GATA6 (9%) (Figure 3A). Statistical analysis indicated that distal EGJA, as compared to proximal EGJA, had significantly higher mutational frequencies of FGFR2 (15.79% vs. 5.04%; P=0.02) and ZEN217 (7.02% vs. 0.72%; P=0.03) (Table 2).

Figure 3 Significantly mutated genes in (A) distal EGJA and (B) proximal EGJA. (Left panel) The matrix of mutations colored according to type of mutation. Each column denotes an individual tumor and each row represents a gene. (Right panel) The gene name of mutations. EGJA, esophagogastric junction adenocarcinoma; TMB, tumor mutation burden.

Meanwhile, proximal tumors were characterized by ERBB2 (15%) and FGFR2 (15%) mutations; 11q13 locus amplification (14% each for FGF3, FGF4, and FGF19); and CCND1 (10%), CDKN2A (10%), and MDM2 (10%) mutations (Figure 3B). MYC variants (6.5%) were exclusively detected in proximal cases and thus could be considered a unique alteration (P=0.06; Table 2).

Differential genetic alterations between proximal EGJA and ESCC and between distal EGJA and GCA

To clarify the molecular relationships between proximal EGJA and ESCC and between distal EGJA and GCA, we systematically compared the mutation frequencies between proximal EGJA, distal EGJA, ESCC and GCA. A pronounced anatomic gradient was observed for TP53 inactivation, with frequencies progressively declining from ESCC (94.4%) to proximal EGJA (75.5%), distal EGJA (70.2%), and GCA (50.0%). Conversely, FGFR2 alterations increased across this anatomic continuum, from ESCC (0%), to proximal EGJA (5.0%), distal EGJA (15.8%), and GCA (11.9%).

Comparative analysis revealed partial a molecular overlap between proximal EGJA and ESCC, with comparable mutation rates in TP53 (75.5% vs. 94.4%; P=0.12) and SPTA1 (10.3% vs. 8.3%; P=0.72), while distal EGJA shared conserved alterations with GCA in TP53 (70.2% vs. 50.0%; P=0.08), CCNE1 (16% vs. 14.3%; P=0.81), and FGFR2 (15.8% vs. 11.9%; P=0.67). Notably, distal EGJA and ESCC demonstrated similar mutations frequencies for ZNF217 (7.0% vs. 0.7%; P=0.03) and FAT3 (12.1% vs. 11.1%; P=0.85), whereas proximal EGJA aligned with GCA in terms of alterations to ERBB2 (15% vs. 14.3; P=0.91), MYC (6.5% vs. 4.8%; P=0.72), and FAT3 (12.1% vs. 9.5%; P=0.65).

These site-specific overlaps suggest a molecular continuum influenced by anatomic origin, although divergent patterns in other genes (CDH1 and NOTCH1) suggest subtype-unique oncogenic mechanisms. The observed FGFR2 gradient, with EGJA at the apex, may reflect location-dependent activation of fibroblast growth factor signaling, while the mutation frequency of TP53 declines across the spectrum, paralleling the transition from squamous (ESCC) to glandular (GCA) carcinogenesis pathways (Table 3).

Analysis of the OS for patients with EGJA

To identify the prognostic determinants in EGJA, we performed comprehensive survival analyses incorporating clinicopathological parameters and molecular biomarkers. Multivariate Cox regression revealed the independent predictors of OS to be tumor stage [hazard ratio (HR): 1.82, 95% confidence interval (CI): 1.34–2.47] and metastatic lymph node burden (HR 2.15, 95% CI: 1.56–2.96) (Figure 4A), while anatomical subsite (proximal vs. distal EGJA) showed no significant association with clinical outcomes (P=0.14; Figure 4B).

Figure 4 Overall survival analysis of EGJA. (A) Multivariate cox regression analysis. (B) Comparison of survival between proximal EGJA and distal EGJA. CI, confidence interval; EGJA, esophagogastric junction adenocarcinoma; N, node.

Molecular characterization identified eight somatic mutations significantly correlated with survival deterioration (with hazard ratios exceeding 2.5 across all significant variants): LRP2 (P=0.05), PRDM1 (P<0.001), GRM3 (P=0.04), HMGA2 (P=0.01), MYB (P=0.01), RET (P<0.001), SF3B1 (P<0.001), and SPTA1 (P=0.04) (Table 4 and Figure 5). Stratification by tumor location indicated distinct molecular prognosticators: CDK6 (P=0.01) emerged as a specific predictor in proximal EGJA, whereas FAT4 (P=0.04) and PMS2 (P=0.01) mutations exhibited prognostic relevance exclusively in distal EGJA (Table 4). Notably, seven genetic alterations (CDK6, GRM3, HMGA2, LRP2, MYB, PRDM1, and SF3B1) showed stronger prognostic associations in proximal tumors than in distal tumors (Table 4), suggesting anatomical heterogeneity in molecular drivers of disease progression. These findings underscore the critical prognostic value of molecular profiling beyond conventional staging systems, revealing spatially distinct genetic determinants of survival related to EGJA pathogenesis.

Figure 5 Analysis of the survival-related mutated genes: (A) LRP2 and (B) SPTA1.

Discussion

In the eighth edition TNM classification for esophageal and gastric cancers (AJCC and UICC), EGJA is staged according to the criteria for esophageal or for gastric cancer (4,5). However, the classification and surgical management of EGJA remain controversial, underscoring the need for a dedicated TNM staging system.

This study investigated the genomic landscape of proximal and distal EGJA through targeted NGS of a 450-cancer gene panel. The specific aims were to identify putative driver mutations distinguishing proximal from distal EGJA, compare the genomic profiles of EGJA subtypes with those of GCA and ESCC, assess the differences in mutation frequency between these entities, and evaluate the prognostic associations of somatic mutations in EGJA. Notably, our analysis revealed that proximal and distal EGJA harbor distinct genomic signatures, and these molecular differences correlate with divergent clinicopathological features. Survival analysis further clarified the prognostic value of mutations across subgroups.

These findings align with prior reports identifying TP53, CCNE1, ERBB2, and LRP1B as frequently mutated genes in Western EGJA cohorts (12-15), although our cohort revealed distinct ethnic-specific patterns. We systematically characterized putative driver mutations across EGJA, GCA, and ESCC. In EGJA, the most prevalent somatic alterations included those for TP53 (74%), CCNE1 (14%), ERBB2 (12%), FAT3 (11%), ARID1A (11%), PIK3CA (10%), SPTA1 (10%), CDK6 (9%), FGF3 (9%), and LRP1B (9%). Furthermore, in line with a recent analysis of the MSK cohort, TP53 was the most common alteration in both studies, with highly comparable mutation rates (74% in our cohort vs. 74–75% in the MSK cohort) (16). In this study, our results indicated that EGJA, GCA, and ESCC had distinct genomic alterations. Notably, only TP53 and PIK3CA mutations were shared as high-frequency events across all three malignancies: First, the mutation frequency of TP53 in EGJA exceeded that of GCA but was significantly lower than that of in ESCC, suggesting an anatomical gradient from proximal to distal tumors (15). Second, PIK3CA mutations exhibited lower prevalence in EGJA than in both ESCC and GCA. Beyond these shared drivers, EGJA demonstrated enrichment of ERBB2 (17,18) and ARID1A (19,20) alterations. The genomic landscape of EGJA reveals targetable alterations in genes such as PIK3CA, ERBB2 (HER2), and ARID1A (implicated in the homologous recombination deficiency pathway). Such alterations point to potential subtype-specific therapies analogous to those deployed in. Consequently, molecular stratification may identify therapeutic vulnerabilities in EGJA subgroups (21): PI3K inhibitors for PIK3CA-mutant tumors (22,23), HER2 inhibitors for ERBB2-altered tumors (24), and agents targeting homologous recombination deficiency for ARID1A-deficient tumors. Comparison of the mutated genes between EGJA, GCA, and ESCC revealed that EGJA was more similar genomically to GCA than to ESCC, particularly distal EGJA. This pattern implies that distal EGJA may share oncogenic mechanisms with conventional GCA, warranting re-evaluation of current classification frameworks.

To clarify the molecular distinctions between proximal and distal EGJA, GCA, and ESCC, we conducted comparative analysis of the putative driver mutations in this study. Notably, FGFR2 emerged as a predominantly mutated gene in distal EGJA, demonstrating significantly elevated mutation frequencies as compared with proximal EGJA. This observation aligns with Klempner et al.’s characterization of FGFR2-altered gastroesophageal adenocarcinomas as a distinct clinicopathologic entity with a unique genomic profile (25). Furthermore, ZNF127 mutations were preferentially enriched in distal EGJA relative to proximal lesions. To our knowledge, this represents the first report identifying FRFR2 and ZNF127 mutations as molecular hallmarks of distal EGJA, suggesting their potential utility as differential diagnostic biomarkers for anatomical subtyping. Intriguingly, MYC mutations were exclusively absent in distal EGJA, providing an additional discriminatory feature. Despite these subtype-specific findings, comprehensive mutational analysis failed to reveal significant molecular commonalities or distinguishing patterns across the proximal/distal EGJA, GCA, and ESCC subgroups.

Prognostic analysis of EGJA identified 11 gene alterations showing significant associations with OS: ATR, CD1D, FGF7, GRM3, HMGA2, LRP2, MYB, PRMD1, RET, SF3B1, and SPTA1. Ataxia telangiectasia and Rad3-related (ATR) can be considered an attractive target for cancer treatment due to its deleterious effect on cancer cells harboring a homologous recombination defect. Inhibition of ATR could induce synthetic lethality with Ataxia Telangiectasia Mutated (ATM) deficiency in gastric cancer cells (26). In ESCC, ATR inhibitor VE-822 was identified as a therapeutic strategy for enhancing cisplatin chemosensitivity (27). Thus, ATR may be an attractive target for the treatment of EGJA. In a study on GC, HMGA2 expression was significantly associated with shorter recurrence-free survival, and was an independent prognostic factor for tumor recurrence (28). Additionally, high expression of HMGA2 has been correlated with a higher T stage, lower differentiation degree, lymph node metastasis, recurrence status, and TNM stage in ESCC (29). Similarly, HMGA2 gene mutations were significantly associated with OS in EGJA in our study. Overall, few studies have reported on the prognostic-related gene alterations examined in our analysis, and the role of these alterations in GCA or ESCC remains unknown. These gene alterations were not similar between EGJA, GCA, and ESCC, indicating that there are large molecular differences between these cancer types.

In the analysis of prognostic-related genomic alterations in distal EGJA and proximal EGJA, we found little similarity in the prognostic-related mutated genes between these subgroups, while the mutated genes in proximal EGJA were more similar to those in general EGJA, suggesting that the gene alterations in proximal EGJA might have a greater influence on prognosis. Moreover, proximal or distal EGJA was not significantly associated with OS, and there was no survival difference between proximal and distal EGJA. Given the limited sample size within key subgroups—including tumor location (distal vs. proximal) and the distribution of disease stages (I/II/III/IV and cases with unknown stage)—we were unable to perform multivariate analysis to adjust for stage at diagnosis in the assessment of OS. As disease stage is a well-established major prognostic factor in EGJA, the inability to adjust for this variable represents a potential confounder; therefore, the survival findings should be interpreted with caution. Future studies with larger, well-annotated, and stage-homogeneous cohorts are warranted to enable robust multivariate analyses that adequately control for key clinical covariates.

Conclusions

In conclusion, based on NGS technology, we identified ten genes with a high frequency of mutations in EGJA. The high frequency of PI3K-AKT pathway alterations in EGJA, compared with GCA and ESCC, suggests its potential susceptibility to PI3K/AKT‑targeted agents such as PIK3CA inhibitors, a hypothesis that needs future clinical validation. In addition, FRFR2, ZNF127, and MYC mutations may be biomarkers for distinguishing distal EGJA from proximal EGJA. Furthermore, there are significant differences in the molecular characteristics between EGJA, GCA, and ESCC, suggesting a molecular basis for the distinct TNM staging of EGJA.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the REMARK reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0027/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0027/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0027/prf

Funding: This work was supported by the Shandong Provincial Natural Science Foundation, China (No. ZR2020MH249).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1-0027/coif). All authors report funding from the Shandong Provincial Natural Science Foundation, China (No. ZR2020MH249). The authors have no other conflicts of interest to declare.

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. The study was approved by Institutional Ethics Review Committee of Shandong Provincial Hospital Affiliated to Shandong First Medical University (LCYJ: No. 2019-148) and informed consent was taken from all the patients.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. 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. [Crossref] [PubMed]
2. Siewert JR, Stein HJ. Classification of adenocarcinoma of the oesophagogastric junction. Br J Surg 1998;85:1457-9. [Crossref] [PubMed]
3. Rüdiger Siewert J, Feith M, Werner M, et al. Adenocarcinoma of the esophagogastric junction: results of surgical therapy based on anatomical/topographic classification in 1,002 consecutive patients. Ann Surg 2000;232:353-61. [Crossref] [PubMed]
4. Padilla-Leal KE, Medina-Franco H. Eighth edition of the American Joint Committee on Cancer staging system: are we getting closer to the ideal classification for gastric cancer? Ann Transl Med 2019;7:S52.
5. Marano L, D'Ignazio A, Cammillini F, et al. Comparison between 7th and 8th edition of AJCC TNM staging system for gastric cancer: old problems and new perspectives. Transl Gastroenterol Hepatol 2019;4:22.
6. Gao YB, Chen ZL, Li JG, et al. Genetic landscape of esophageal squamous cell carcinoma. Nat Genet 2014;46:1097-102. [Crossref] [PubMed]
7. Song Y, Li L, Ou Y, et al. Identification of genomic alterations in oesophageal squamous cell cancer. Nature 2014;509:91-5. [Crossref] [PubMed]
8. Paynter RA, Skibola DR, Skibola CF, et al. Accuracy of multiplexed Illumina platform-based single-nucleotide polymorphism genotyping compared between genomic and whole genome amplified DNA collected from multiple sources. Cancer Epidemiol Biomarkers Prev 2006;15:2533-6. [Crossref] [PubMed]
9. Press OA, Zhang W, Gordon MA, et al. Gender-related survival differences associated with EGFR polymorphisms in metastatic colon cancer. Cancer Res 2008;68:3037-42. [Crossref] [PubMed]
10. Frampton GM, Fichtenholtz A, Otto GA, et al. Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing. Nat Biotechnol 2013;31:1023-31. [Crossref] [PubMed]
11. Cao J, Chen L, Li H, et al. An Accurate and Comprehensive Clinical Sequencing Assay for Cancer Targeted and Immunotherapies. Oncologist 2019;24:e1294-302. [Crossref] [PubMed]
12. Rustgi N, Wu S, Samec T, et al. Molecular Landscape and Clinical Implication of CCNE1-amplified Esophagogastric Cancer. Cancer Res Commun 2024;4:1399-409. [Crossref] [PubMed]
13. Li-Chang HH, Kasaian K, Ng Y, et al. Retrospective review using targeted deep sequencing reveals mutational differences between gastroesophageal junction and gastric carcinomas. BMC Cancer 2015;15:32. [Crossref] [PubMed]
14. Geng Q, Lao J, Zuo X, et al. Identification of the distinct genomic features in gastroesophageal junction adenocarcinoma and its Siewert subtypes. J Pathol 2020;252:263-73. [Crossref] [PubMed]
15. Ireland AP, Shibata DK, Chandrasoma P, et al. Clinical significance of p53 mutations in adenocarcinoma of the esophagus and cardia. Ann Surg 2000;231:179-87. [Crossref] [PubMed]
16. Voutsadakis IA. The Landscape and Prognosis of Microsatellite Stable (MSS) Esophageal, Gastro-Esophageal Junction and Gastric Adenocarcinomas with High Tumor Mutation Burden (TMB). Cancer Invest 2024;42:697-709. [Crossref] [PubMed]
17. Kirouac DC, Du J, Lahdenranta J, et al. HER2+ Cancer Cell Dependence on PI3K vs. MAPK Signaling Axes Is Determined by Expression of EGFR, ERBB3 and CDKN1B. PLoS Comput Biol 2016;12:e1004827. [Crossref] [PubMed]
18. Li S, Ma YM, Zheng PS, et al. GDF15 promotes the proliferation of cervical cancer cells by phosphorylating AKT1 and Erk1/2 through the receptor ErbB2. J Exp Clin Cancer Res 2018;37:80. [Crossref] [PubMed]
19. Yang L, Yang G, Ding Y, et al. Inhibition of PI3K/AKT Signaling Pathway Radiosensitizes Pancreatic Cancer Cells with ARID1A Deficiency in Vitro. J Cancer 2018;9:890-900. [Crossref] [PubMed]
20. De P, Dey N. Mutation-Driven Signals of ARID1A and PI3K Pathways in Ovarian Carcinomas: Alteration Is An Opportunity. Int J Mol Sci 2019;20:5732. [Crossref] [PubMed]
21. Abe H, Urabe M, Yagi K, Yamashita H, Seto Y, Ushiku T. Expression of therapy target molecules in esophagogastric junction and Barrett's adenocarcinoma. Gastric Cancer 2025;28:264-74. [Crossref] [PubMed]
22. Janku F, Wheler JJ, Westin SN, et al. PI3K/AKT/mTOR inhibitors in patients with breast and gynecologic malignancies harboring PIK3CA mutations. J Clin Oncol 2012;30:777-82. [Crossref] [PubMed]
23. Vasan N, Razavi P, Johnson JL, et al. Double PIK3CA mutations in cis increase oncogenicity and sensitivity to PI3Kα inhibitors. Science 2019;366:714-23. [Crossref] [PubMed]
24. Li H, Ren C, Cui D, et al. Case report: Patients with positive HER-2 amplification locally advanced gastroesophageal junction cancer achieved pathologic complete response with the addition of pembrolizumab to chemotherapy plus trastuzumab as neoadjuvant therapy. Front Immunol 2025;16:1555074. [Crossref] [PubMed]
25. Klempner SJ, Madison R, Pujara V, et al. FGFR2-Altered Gastroesophageal Adenocarcinomas Are an Uncommon Clinicopathologic Entity with a Distinct Genomic Landscape. Oncologist 2019;24:1462-8. [Crossref] [PubMed]
26. Min A, Im SA, Jang H, et al. AZD6738, A Novel Oral Inhibitor of ATR, Induces Synthetic Lethality with ATM Deficiency in Gastric Cancer Cells. Mol Cancer Ther 2017;16:566-77. [Crossref] [PubMed]
27. Shi Q, Shen LY, Dong B, et al. The identification of the ATR inhibitor VE-822 as a therapeutic strategy for enhancing cisplatin chemosensitivity in esophageal squamous cell carcinoma. Cancer Lett 2018;432:56-68. [Crossref] [PubMed]
28. Jun KH, Jung JH, Choi HJ, et al. HMGA1/HMGA2 protein expression and prognostic implications in gastric cancer. Int J Surg 2015;24:39-44. [Crossref] [PubMed]
29. Zhang M, Hu D, Wang S, et al. Clinicopathologic significance of HMGA2 expression's correlation with prognosis of esophageal squamous cell carcinoma after Ivor Lewis esophagectomy. Minerva Chir 2016;71:239-44.

(English Language Editor: J. Gray)