Screening of molecular biomarkers ASPN and LBH and construction of a prediction nomogram for the progression of esophagogastric junction adenocarcinoma

Introduction

Esophagogastric junction adenocarcinoma (EGJA) is a malignant tumor of the gastrointestinal tract with high incidence and aggressive behavior. In recent years, its incidence has shown a significant upward trend globally (1,2). Due to its unique anatomical location and complex biological characteristics, the diagnosis and treatment of EGJA remain formidable challenges (3). Tumor staging at the time of diagnosis, the selection of individualized treatment strategies, and the optimization of therapeutic approaches are critical factors influencing patient prognosis. However, for the advanced stage EGJA, even after rigorous preoperative staging and potentially curative surgery, many patients experience tumor recurrence within two years postoperatively (4,5), and the 5-year overall survival rate remains below 25% (6,7). Therefore, improving the capability for early diagnosis, accurately predicting tumor progression, and optimizing treatment strategies have become key research priorities in recent years.

The development of nomogram models has provided significant support for the individualized treatment of EGJA. By integrating multiple biomarkers, clinical features, and imaging data, nomograms enable more accurate predictions of patient survival, lymph node metastasis risk, and treatment response, thus offering a more comprehensive and evidence-based approach to prognostic management. Nomograms effectively combine various clinical, pathological, and imaging indicators to provide personalized prognostic predictions and therapeutic guidance, demonstrating superior predictive performance compared to the traditional tumor-node-metastasis (TNM) staging system. For instance, Wei et al. (8) developed a nomogram based on the neutrophil-to-lymphocyte ratio (NLR), body mass index (BMI), and blood biomarkers to predict overall survival after curative resection of EGJA. This model showed superior predictive performance compared to the TNM staging system, offering an important tool for individualized survival prediction. Similarly, Xu et al. (9) constructed a nomogram incorporating nutrition-related blood biomarkers and computed tomography (CT) imaging features to preoperatively assess the risk of lymph node metastasis in EGJA patients, significantly improving the accuracy of prognostic predictions.

For EGJA patients undergoing neoadjuvant therapy, nomograms have also been widely used to predict disease-free survival (DFS) and chemotherapy response. For example, nomograms integrating factors, such as preoperative clinical staging, tumor pathological characteristics, and postoperative lymph node infiltration, have demonstrated high accuracy in predicting DFS for Siewert type II and III EGJA patients, exhibiting robust predictive performance (10). Additionally, a nomogram based on CT imaging features successfully predicted patient responses to chemotherapy regimens involving docetaxel, oxaliplatin, and S-1, providing precise guidance for the development of individualized treatment plans (11). The application of nomograms in predicting surgical prognosis and long-term survival has also garnered considerable attention. For instance, a study utilizing data from the Surveillance, Epidemiology, and End Results (SEER) database developed a nomogram to predict 1-, 3-, and 5-year survival rates for EGJA patients, identifying independent risk factors such as tumor size, stage, and surgery. The findings demonstrated that the model significantly outperformed the traditional TNM staging system in survival prediction (12). Moreover, nomogram models have been employed to evaluate prognostic differences between patients receiving neoadjuvant radiotherapy and those who did not, providing clinicians with more targeted, individualized decision-making support for radiotherapy planning (13).

However, for endoscopists, confidently and accurately distinguishing between early and advanced-stage EGJA remains a significant challenge in clinical practice. Against this backdrop, developing a preoperative diagnostic and prognostic model that integrates clinical information, endoscopic findings, pathological features, and molecular biology data to assess the progression of EGJA is a promising research direction.

Therefore, this study experimentally screened and validated the differential expression of molecular biomarkers ASPN and LBH in tumors at different stages, using them as important factors for predicting tumor progression. This provided strong support for the accuracy of the model. The constructed diagnostic prediction model integrates patients’ basic clinical information (such as age, gender), endoscopic features (such as lesion size), pathological results (such as histological differentiation), and the protein expression levels of ASPN and LBH. It can assist clinicians in making more objective and accurate judgments regarding tumor progression, thereby optimizing treatment strategies. We present this article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0361/rc).

Methods

Bioinformatics analysis and gene screening

Transcriptomic and clinical data were obtained from The Cancer Genome Atlas Stomach Adenocarcinoma (TCGA-STAD) cohort. EGJA samples were strictly defined as those with anatomical site annotated as cardia in the original clinical metadata. The data were log-transformed (log(X+1)) and organized into an expression matrix using a Perl script. Differential gene expressions between cancerous and adjacent non-cancerous tissues were analyzed using the limma package in R version 4.1.3 (R Foundation for Statistical Computing, Vienna, Austria), with selection criteria set to logFC absolute value >0.5 and P value <0.05 to identify differentially expressed genes (DEGs). Samples were grouped based on clinical staging information, and gene expression differences between stages were analyzed again using the limma package, applying the same selection criteria (logFC absolute value >0.5 and P value <0.05).

The intersection of the genes identified in the two steps was taken to obtain a set of DEGs that met both conditions. Expression levels of the intersected genes were extracted and normalized by log₂(X+1) transformation. Univariate Cox proportional hazards regression analysis was performed using the survival package in R, with gene expression set as a continuous variable. No additional covariates or stratification procedures were applied. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated to identify key genes significantly associated with patient survival.

Quantitative real-time polymerase chain reaction (qRT-PCR) analysis of key gene expression differences between cancerous and adjacent non-cancerous tissues

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of Tongji Hospital, Tongji University (approval number 2023-025). Informed consent to participate was obtained from all participants before their inclusion in the study. Four patients diagnosed with EGJA from the hospital’s gastric cancer biobank were selected, and both cancerous and adjacent non-cancerous tissues were preserved during surgery. Total RNA was extracted from the tissue samples and reverse transcribed to synthesize cDNA. The qPCR reaction system was based on SYBR Green dye, with 18S RNA used as the internal reference gene (primer sequences are shown in Table 1). The qPCR amplification program included pre-denaturation, annealing, extension, and melt curve analysis to ensure amplification specificity. The relative expression levels of target genes were calculated using the ΔΔCt method, and data were analyzed statistically using GraphPad Prism 8.4.3 (GraphPad Software, San Diego, CA, USA).

Immunohistochemical (IHC) staining analysis of key gene-related protein expression differences between cancerous and adjacent non-cancerous tissues

In this part, 70 EGJA patients who underwent surgery at the hospital were enrolled. IHC staining was used to quantitatively analyze the protein expression in cancerous and adjacent non-cancerous tissues. Tissue sections were labeled with primary and secondary antibodies. Antigen retrieval was performed using an automatic immunohistochemistry pretreatment system. The system was preheated to 65 ℃ with sodium citrate-EDTA antigen retrieval buffer. Tissue sections were placed into the system to execute the retrieval program. Upon completion, the sections were cooled to room temperature and rinsed with PBS buffer to remove residual impurities. A hydrophobic barrier pen was used to circle the tissue area, which confines the coverage of reagents, reduces consumption and improves experimental accuracy. The detailed information of primary antibodies used in this study is as follows: Anti-ASPN antibody (Abcepta, Cat. No. AP6604C), working dilution: 1:150; Anti-LBH antibody (Zenbio, Cat. No. 125562), working dilution: 1:125; Anti-SFRP4 antibody (Proteintech, Cat. No. 15328-1-AP), working dilution: 1:330. And protein localization was visualized using DAB staining. Digital scanning of the tissue sections was performed using a high-resolution image acquisition system, capturing images of the sections, and representative fields were selected for analysis at different magnifications. The staining images were quantitatively analyzed using ImageJ software, measuring the average optical density (AOD) to infer the expression levels of the target proteins.

Construction of the EGJA prediction nomogram based on molecular biomarkers ASPN and LBH and multidimensional information

This study included baseline information, clinical-pathological data, and IHC results from 70 EGJA patients to develop an effective predictive model. In the data organization phase, descriptive statistical analysis was performed on the baseline characteristics of all patients to comprehensively understand their demographic and clinical features. Based on the research objectives, factors associated with predicting early or advanced tumor stages were selected, including age, gender, tumor size, differentiation type, and the expression levels of ASPN and LBH. To simplify the model and improve its operability, numerical and categorical variables were transformed into binary variables. Specifically, the differentiation type was categorized into well/moderately differentiated and poorly differentiated groups, age was divided into ≤66 and >66 years groups based on the mean value, and tumor size, ASPN, and LBH AOD expression values were evaluated using receiver operating characteristic (ROC) curves to determine the optimal cutoff points, using the Youden index (sensitivity + specificity − 1). The results showed that the optimal cutoff for tumor size was 2 cm, for ASPN AOD was 0.237, and for LBH AOD was 0.316, which were used to categorize the variables into binary groups.

In the model construction phase, the operations were performed using the R software (version 4.2.2), along with MSTATA software (www.mstata.com). Univariate analysis was performed to assess the independent effect of each variable on early tumor prediction. Multicollinearity among variables was evaluated using variance inflation factor (VIF) and tolerance analysis. Multivariate screening was conducted using the least absolute shrinkage and selection operator (LASSO) logistic regression analysis to select independent risk factors for constructing the predictive model. The model’s performance was assessed using ROC curves and calibration curves, and decision curve analysis (DCA) was further employed to evaluate the clinical applicability of the model.

Statistical analysis

Statistical analyses were performed using R software version 4.1.3 (R Foundation for Statistical Computing, Vienna, Austria), GraphPad Prism version 8.4.3 (GraphPad Software, San Diego, CA, USA), and MSTATA software. Continuous variables were summarized as the mean and standard deviation, whereas categorical variables were presented as frequencies and percentages. Paired comparisons between cancerous and matched adjacent non-cancerous tissues were performed using paired Student’s t-tests. Comparisons among three or more groups were conducted using one-way analysis of variance followed by Tukey’s multiple-comparison test. Univariate Cox proportional hazards regression was used to evaluate associations between gene expression and overall survival, with hazard ratios and 95% confidence intervals reported. For construction of the tumor-stage prediction model, univariate logistic regression was initially performed to evaluate candidate predictors. Multicollinearity was assessed using the VIF and tolerance, and predictor selection was subsequently conducted using LASSO logistic regression. ROC curves and the Youden index were used to determine optimal cutoff values and evaluate model discrimination. Model calibration and potential clinical utility were assessed using calibration plots and DCA, respectively. All statistical tests were two-sided, and a P value <0.05 was considered statistically significant.

Results

Bioinformatics analysis results

All HR values and survival analysis results were derived from the TCGA-STAD cohort data. A total of 4,638 DEGs were identified through differential gene analysis. Further analysis of gene expression differences across different tumor stages revealed 424 DEGs, and corresponding volcano plots were generated (Figure 1A,1B). Upregulated DEGs were intersected, resulting in 72 common genes, and a Venn diagram was generated (Figure 1C).

Figure 1 Bioinformatics analysis results. Data were derived from the TCGA-STAD cohort. (A) Significantly differentially expressed genes between cancer and adjacent normal tissue. (B) Differentially expressed genes across different tumor stages. (C) Venn diagram of differential genes intersection. (D) Forest plot of the top 6 genes with the highest hazard ratios. FC, fold change; TCGA-STAD, The Cancer Genome Atlas Stomach Adenocarcinoma.

The expression levels of the common genes were extracted and combined with the survival data of the patients for univariate survival analysis. The results indicated that high expression of SFRP4 (HR =1.260, 95% CI: 1.023 to 1.551, P=0.03), APOE (HR =1.264, 95% CI: 1.012 to 1.580, P=0.04), ASPN (HR =1.628, 95% CI: 1.141 to 2.323, P=0.007), LBH (HR =1.634, 95% CI: 1.020 to 2.616, P=0.041), FAM19A5 (HR =1.608, 95% CI: 1.063 to 2.432, P=0.02), and VCAM1 (HR =1.713, 95% CI: 1.068 to 2.747, P=0.03) was significantly associated with an increased risk of patient mortality, with all P values less than 0.05, indicating statistical significance (Figure 1D). The high expression of these genes may serve as potential prognostic markers for EGJA patients.

qRT-PCR results

The qRT-PCR results indicated significant differences in the expression of LBH, ASPN, and SFRP4 between EGJA tissues and adjacent normal tissues. Specifically, the mean expression of SFRP4 was significantly higher in the cancer group compared to the adjacent normal group (P=0.049). ASPN also showed a significant difference (P=0.02), with higher RNA expression levels in the cancer group than in the adjacent normal tissues. Furthermore, LBH expression was significantly higher in cancer tissues than in adjacent normal tissues (P=0.007), suggesting that it may play an important role in tumorigenesis and progression. Although other genes such as APOE, VCAM-1, and FAM19A5 showed some expression differences, they did not reach statistical significance (P>0.05), which may be attributed to sample size or biological variability (Figure 2).

Figure 2 qRT-PCR results. Data were obtained from the qPCR validation cohort of EGJA tissues. *, P<0.05; **, P<0.01. EGJA, esophagogastric junction adenocarcinoma; N, qRT-PCR, quantitative real-time polymerase chain reaction; T.

These results suggest that the differential expression of LBH, ASPN, and SFRP4 in EGJA may hold clinical significance and could serve as potential biomarkers or regulatory factors for cancer. Building on these findings, this study investigated the protein expression of these three genes to validate their possible mechanisms in cancer initiation and progression.

IHC results

Expression of different markers in cancer and adjacent normal tissues

Immunofluorescence results showed that the expression of ASPN in cancer tissues was significantly higher than in adjacent normal tissues (95% CI: 0.0528 to 0.0898, P<0.0001) (Figure 3). In contrast, the expression of LBH was significantly lower in cancer tissues compared to adjacent normal tissues (95% CI: −0.1131 to −0.0653, P<0.0001) (Figure 4). The expression of SFRP4 was slightly higher in cancer tissues than in adjacent normal tissues (95% CI: 0.0022 to 0.0498, P=0.03) (Figure 5). These results suggest that the expression differences of ASPN, LBH, and SFRP4 may be related to the occurrence and progression of cancer.

Figure 3 ASPN expression is significantly higher in cancer tissue than in adjacent normal tissue. Data were obtained from 70 clinical EGJA patients (IHC cohort). ****, P<0.0001. AOD, average optical density; EGJA, esophagogastric junction adenocarcinoma; IHC, immunohistochemical; IOD.

Figure 4 LBH expression is significantly lower in cancer tissue than in adjacent normal tissue. Data were obtained from 70 clinical EGJA patients (IHC cohort). ****, P<0.0001. AOD, average optical density; EGJA, esophagogastric junction adenocarcinoma; IHC, immunohistochemical; IOD.

Figure 5 SFRP4 expression is higher in cancer tissue than in adjacent normal tissue. Data were obtained from 70 clinical EGJA patients (IHC cohort). *, P<0.05. AOD, average optical density; EGJA, esophagogastric junction adenocarcinoma; IHC, immunohistochemical; IOD.

Expression of different markers at different tumor stages

Immunofluorescence results revealed significant differences in ASPN expression across different tumor stages, particularly between stage I and stage II (P<0.05), as well as between stage I and stage III–IV (P<0.001). Further analysis indicated that the expression of ASPN was significantly lower in early-stage tumors compared to advanced-stage tumors (early-stage cancers widely accepted as lesions limited to the mucosa and submucosa, regardless of lymph node metastasis) suggesting that ASPN may be closely associated with EGJA progression and could serve as an important biomarker for tumor staging (Figures 6,7). Similarly, LBH expression also showed significant differences across stages, particularly between stage I and stage III–IV (P<0.05). The expression of LBH was significantly lower in advanced-stage tumors compared to early-stage tumors, indicating that LBH may play a negative regulatory role in tumor progression (Figures 8,9). In contrast, the expression of SFRP4 did not show significant differences between stages, with no apparent change between early and advanced-stage tumors (P>0.05), suggesting that SFRP4 may not be suitable as a biomarker for tumor staging (Figure S1).

Figure 6 Immunohistochemistry of ASPN in tumors at different stages. (A) ASPN expression in early-stage tumors (4×). (B) ASPN expression in early-stage tumors (20×). (C) ASPN expression in advanced-stage tumors (4×). (D) ASPN expression in advanced-stage tumors (20×). Data were obtained from 70 clinical EGJA patients (IHC cohort). EGJA, esophagogastric junction adenocarcinoma; IHC, immunohistochemical.

Figure 7 ASPN expressions across different tumor stages. Data were obtained from 70 clinical EGJA patients (IHC cohort). ns, not significant; *, P<0.05; ***, P<0.001; ****, P<0.0001. AOD, average optical density; EGJA, esophagogastric junction adenocarcinoma; IHC, immunohistochemical; IOD.

Figure 8 Immunohistochemistry of LBH in tumors at different stages. (A) LBH expression in early-stage tumors (4×). (B) LBH expression in early-stage tumors (20×). (C) LBH expression in advanced-stage tumors (4×). (D) LBH expression in advanced-stage tumors (20×). Data were obtained from 70 clinical EGJA patients (IHC cohort). EGJA, esophagogastric junction adenocarcinoma; IHC, immunohistochemical.

Figure 9 LBH expression across different tumor stages. Data were obtained from 70 clinical EGJA patients (IHC cohort). ns, not significant; *, P<0.05. AOD, average optical density; EGJA, esophagogastric junction adenocarcinoma; IHC, immunohistochemical; IOD.

Construction of a prediction nomogram for early/advanced stage EGJA

Nomogram construction

The nomogram provides a tool for assessing the likelihood of early-stage EGJA occurrence by displaying the scores and weights of multiple predictive factors. The score range for each factor is determined based on its variable value, with higher scores indicating a greater probability of early tumor detection. The scores for factors such as gender, tumor size, differentiation type, LBH, and ASPN are presented in Figure 10. Regarding gender, males have a higher score (approximately 39 points), highlighting its significance in predicting early tumor presence. Patients with tumors smaller than 2 cm score 55 points, suggesting a higher likelihood of early-stage tumors. In terms of differentiation, patients with well- to moderately differentiated tumors score 42 points, whereas poorly differentiated patients score 0, indicating that lower differentiation is associated with a reduced likelihood of early tumor development. High expression of LBH is associated with a higher score (approximately 37 points), while low expression scores 0, suggesting that elevated LBH expression may be indicative of early-stage tumors. In contrast, patients with high expression of ASPN score 0, indicating a very low risk of early tumor development, while patients with low ASPN expression score 100 points, implying that lower ASPN expression is associated with a higher likelihood of early tumor occurrence. The linear prediction value is calculated based on the total score (ranging from 0 to 300), which is further used to estimate the probability of early tumor onset.

Figure 10 Nomogram. Data were obtained from 70 clinical EGJA patients. EGJA, esophagogastric junction adenocarcinoma.

ROC curve analysis

The ROC curve was used to evaluate the diagnostic performance of the model, with an AUC value of 0.977 (95% CI: 0.945 to 1.000), indicating high predictive accuracy. The inclusion of ASPN and LBH enhanced the model’s predictive capacity, particularly ASPN, which holds significant weight in the nomogram, underscoring its crucial role in predicting early-stage EGJA (Figure 11A). Compared to the model without ASPN and LBH [area under the curve (AUC) =0.897], the model incorporating these two factors showed a substantial improvement in AUC, confirming the diagnostic significance of these two factors (Figure 11B).

Figure 11 Model evaluation. (A) ROC curve of the model including ASPN and LBH. (B) ROC curve of the model excluding ASPN and LBH. (C) Decision curve analysis of the model. (D) Calibration plot of the model. Data were obtained from 70 clinical EGJA patients. AUC, area under the curve; EGJA, esophagogastric junction adenocarcinoma; ROC, receiver operating characteristic.

DCA

DCA was used to assess the clinical net benefit of the model across different threshold probabilities. Within the threshold range of 0.1 to 0.7, the DCA curve consistently exceeded the “None” and “All” baseline models, indicating that the model provides a positive clinical net benefit within this range and can offer valuable support for clinical decision-making (Figure 11C).

Calibration plot

The calibration plot was used to assess the consistency between the predicted values and the actual observed outcomes. Ideally, the 45° diagonal line represents perfect calibration, while the bias-corrected line should closely follow the ideal line, indicating that after calibration, the model’s predictions align closely with the actual results. This outcome suggests that the model demonstrates good predictive performance and can provide reliable early risk assessment for EGJA in clinical settings (Figure 11D).

Discussion

This study comprehensively explored the expression and potential functional mechanisms of various molecular biomarkers associated with EGJA through multiple analytical approaches, including bioinformatics analysis, qRT-PCR validation, and IHC experiments. The results of this study indicate that ASPN and LBH may play crucial roles in tumor progression and could serve as important molecular biomarkers for predicting the occurrence and progression of EGJA.

This study conducted a large-scale bioinformatics analysis using the TCGA database and identified six genes—SFRP4, APOE, ASPN, LBH, FAM19A5, and VCAM1—whose expression was significantly associated with patient prognosis. High expressions of genes such as ASPN, LBH, and SFRP4 were closely linked to poorer prognosis, suggesting that they may play a crucial role in tumor development and progression. To validate the expression of these genes, the mRNA levels of ASPN, LBH, and SFRP4 were examined in both cancerous and adjacent non-cancerous tissues. The results showed that mRNA expression of ASPN, LBH, and SFRP4 were significantly elevated in cancer tissues. IHC analysis further demonstrated that ASPN protein expression was significantly upregulated in cancer tissues, particularly in advanced-stage tumors. In contrast, LBH protein expression was lower in cancer tissues compared to adjacent non-cancerous tissues, and its expression significantly decreased with tumor progression, suggesting the involvement of complex post-transcriptional regulatory mechanisms. Although there was a difference in SFRP4 protein expression between cancer and adjacent tissues, no significant differences were observed across different tumor stages, sizes, or differentiation grades, indicating that SFRP4 may not serve as an effective predictor of tumor progression.

The specific mechanisms by which ASPN and LBH contribute to tumorigenesis and progression have been preliminarily elucidated. ASPN, an extracellular matrix protein, plays a crucial role in tumor microenvironment remodeling, as well as in the invasion and metastasis of cancer cells. Through Metascape analysis, ASPN has been found to participate in cancer-related signaling pathways and proteoglycan-associated pathways, regulating cell migration and epithelial-mesenchymal transition, and promoting tumor growth and expansion. Studies have shown that ASPN expression is elevated in various cancers, including gastric cancer, breast cancer, colon cancer, and pancreatic cancer, driving the proliferation, migration, and metastasis of tumor cells (14-23). In this study, high ASPN expressions in EGJA were positively correlated with tumor size, differentiation, and other pathological features, suggesting its potential oncogenic role in the tumor’s development and progression. Future research should further validate the specific functions of ASPN in tumor invasion, migration, and metastasis using animal models and molecular biology experiments, and explore its potential as a therapeutic target.

LBH exhibits complex and diverse roles in cancer, and its function is closely related to the progression of various tumors. LBH plays a crucial role in regulating the WNT signaling pathway and, by influencing epithelial-mesenchymal transition, cell proliferation, and differentiation, it limits cancer cell migration and proliferation (24,25). However, high expression of LBH has been associated with the progression of certain cancer types, such as prostate cancer, liver cancer, and breast cancer (26-29). In this study, although LBH mRNA levels were higher in EGJA compared to adjacent non-cancerous tissue, IHC results showed that LBH protein expression was higher in the adjacent tissue, and its protein expression decreased progressively with tumor progression. This discrepancy may be related to specific post-transcriptional regulatory mechanisms, including microRNA-mediated translational suppression that inhibits protein synthesis without altering mRNA abundance, as well as ubiquitin-proteasome pathway activation that accelerates the degradation of LBH protein in cancerous tissues. Other potential mechanisms such as dysregulated RNA-binding proteins or long non-coding RNAs may also contribute to the uncoupling between mRNA and protein levels. Future research should further investigate the molecular mechanisms underlying this inconsistency between mRNA and protein expression and explore the specific roles of LBH in different cancer types and its potential prognostic and therapeutic implications.

Based on the expression differences of ASPN and LBH, this study incorporated these two genes into the construction of a tumor prediction model, integrating clinical features, pathological results, and molecular biomarker expression. This approach provides crucial evidence for the early diagnosis and differentiation of advanced tumors. The expression differences of ASPN and LBH form an important molecular foundation for the model, further enhancing its predictive accuracy for clinical applications. This study not only experimentally validated the expression patterns of key genes but also developed risk prediction models for early and advanced tumors through a nomogram. By comprehensively analyzing the expression of proteins such as ASPN and LBH, along with other clinical variables, the proposed prediction model offers an individualized diagnostic tool for early EGJA. The nomogram integrates patient gender, tumor size, differentiation type, and ASPN/LBH expression, providing a visually intuitive and user-friendly tool for clinicians to predict tumor progression risks. ROC curve analysis of the model showed an AUC value of 0.977, indicating high accuracy in distinguishing early and advanced tumors. This result suggests that the model, incorporating ASPN and LBH, holds significant clinical potential and can assist clinicians in formulating more precise treatment plans, particularly in determining whether patients require more aggressive treatment. DCA further suggests that the proposed prediction model provides strong support for clinical decision-making and may help strike a balance between reducing overtreatment and avoiding missed diagnoses (30-34).

Despite the significant progress made in establishing a risk prediction model for EGJA and exploring the molecular mechanisms of genes such as ASPN and LBH, several limitations remain in this study. First, the sample size is relatively small, as this study is based on data from 70 patients, which may affect the generalizability of the model. The limited sample size also poses a risk of overfitting for the multivariable model. While the model shows good internal calibration, external validation using an independent cohort is mandatory before clinical adoption. Future studies should aim to expand the sample size through multi-center collaborations to further validate the expression of these genes and their clinical application value in larger populations. Second, the specific biological mechanisms of ASPN and LBH are not yet fully understood. Although this study has verified their expression differences through bioinformatics analysis, qRT-PCR, and immunohistochemistry, their roles in tumors still require further investigation, particularly how ASPN affects the tumor microenvironment through extracellular matrix remodeling and the signaling pathways through which LBH exerts its effects. Finally, although the effectiveness of the prediction model was demonstrated through ROC curve and DCA, its practical clinical application still requires prospective validation. Future research should conduct prospective clinical trials to assess its accuracy and effectiveness across different patient populations, particularly in patients with varying age, gender, and tumor staging.

Conclusions

This study revealed the significant roles of ASPN and LBH in both early-stage and advanced-stage EGJA, suggesting that these two molecular markers may serve as potential predictors for distinguishing between early and advanced tumors. Additionally, based on the selected molecular biomarkers, clinical information, and pathological characteristics, an effective nomogram model was constructed to predict early and advanced tumor stages. The model demonstrated an AUC value of 0.977, highlighting its excellent predictive performance and promising clinical applicability.

Acknowledgments

The authors thank the medical staff of the Endoscopy Center of Tongji Hospital, Shanghai, China, and all participants who were willing to take part in this study.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0361/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0361/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0361/prf

Funding: This work was supported by the Clinical Research Plan of SHDC (grant No. SHDC2022CRT004), Science and Technology Innovation Action Plan of STCSM (grant No. 22DZ2203900), Shanghai Municipal Science and Technology Project (grant No. 21Y11658500), Shanghai Informatization Development Special Project (grant No. 202001003), and National Natural Science Foundation of China (grant No. 81974067).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2026-0361/coif). The authors have no 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. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Tongji Hospital, Tongji University (approval number 2023-025). Written informed consent was obtained from all participants before their inclusion in the study.

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. Buas MF, Vaughan TL. Epidemiology and risk factors for gastroesophageal junction tumors: understanding the rising incidence of this disease. Semin Radiat Oncol 2013;23:3-9. [Crossref] [PubMed]
2. Sitarz R, Skierucha M, Mielko J, et al. Gastric cancer: epidemiology, prevention, classification, and treatment. Cancer Manag Res 2018;10:239-48. [Crossref] [PubMed]
3. Siewert JR, Stein HJ. Carcinoma of the gastroesophageal junction - classification, pathology and extent of resection. Diseases of the Esophagus 1996;9:173-82.
4. Mariette C, Balon JM, Piessen G, et al. Pattern of recurrence following complete resection of esophageal carcinoma and factors predictive of recurrent disease. Cancer 2003;97:1616-23. [Crossref] [PubMed]
5. de Manzoni G, Pedrazzani C, Pasini F, et al. Pattern of recurrence after surgery in adenocarcinoma of the gastro-oesophageal junction. Eur J Surg Oncol 2003;29:506-10. [Crossref] [PubMed]
6. Hulscher JB, Tijssen JG, Obertop H, et al. Transthoracic versus transhiatal resection for carcinoma of the esophagus: a meta-analysis. Ann Thorac Surg 2001;72:306-13. [Crossref] [PubMed]
7. van Sandick JW, van Lanschot JJ, ten Kate FJ, et al. Indicators of prognosis after transhiatal esophageal resection without thoracotomy for cancer. J Am Coll Surg 2002;194:28-36. [Crossref] [PubMed]
8. Wei ZJ, Qiao YT, Zhou BC, et al. Model established based on blood markers predicts overall survival in patients after radical resection of types II and III adenocarcinoma of the esophagogastric junction. World J Gastrointest Surg 2022;14:788-98. [Crossref] [PubMed]
9. Xu Y, Luo L, Feng X, et al. Nomogram for Predicting Risk of Esophagogastric Junction (EGJ) Resection During Laparoscopic Resection of Gastrointestinal Stromal Tumors in EGJ: A Retrospective Multicenter Study. Front Surg 2021;8:712984. [Crossref] [PubMed]
10. Guo Z, Guo H, Tian Y, et al. Nomograms for Predicting Disease-Free Survival in Patients With Siewert Type II/III Adenocarcinoma of the Esophagogastric Junction Receiving Neoadjuvant Therapy and Radical Surgery. Front Oncol 2022;12:908229. [Crossref] [PubMed]
11. Zhou CQ, Gao D, Gui Y, et al. Computed tomography-based nomogram of Siewert type II/III adenocarcinoma of esophagogastric junction to predict response to docetaxel, oxaliplatin and S-1. World J Radiol 2024;16:9-19. [Crossref] [PubMed]
12. Liu X, Jiang Q, Yue C, et al. Clinicopathological Characteristics and Survival Predictions for Adenocarcinoma of the Esophagogastric Junction: A SEER Population-Based Retrospective Study. Int J Gen Med 2021;14:10303-14. [Crossref] [PubMed]
13. Guo Z, Wang N, Liu F, et al. Prognostic nomogram for Siewert type II adenocarcinoma of the esophagogastric junction patients with and without neoadjuvant radiotherapy: a retrospective cohort study. Am J Transl Res 2022;14:135-49.
14. Xu L, Li Z, Liu SY, et al. Asporin and osteoarthritis. Osteoarthritis Cartilage 2015;23:933-9. [Crossref] [PubMed]
15. Fan R, Yan X, Zhang W. Relationship between asporin and extracellular matrix behavior: A literature review. Medicine (Baltimore) 2022;101:e32490. [Crossref] [PubMed]
16. Zhang Z, Li H, Zhao Y, et al. Asporin promotes cell proliferation via interacting with PSMD2 in gastric cancer. Front Biosci (Landmark Ed) 2019;24:1178-89. [Crossref] [PubMed]
17. Maris P, Blomme A, Palacios AP, et al. Asporin Is a Fibroblast-Derived TGF-β1 Inhibitor and a Tumor Suppressor Associated with Good Prognosis in Breast Cancer. PLoS Med 2015;12:e1001871. [Crossref] [PubMed]
18. Wu H, Jing X, Cheng X, et al. Asporin enhances colorectal cancer metastasis through activating the EGFR/src/cortactin signaling pathway. Oncotarget 2016;7:73402-13. [Crossref] [PubMed]
19. Wang L, Wu H, Wang L, et al. Asporin promotes pancreatic cancer cell invasion and migration by regulating the epithelial-to-mesenchymal transition (EMT) through both autocrine and paracrine mechanisms. Cancer Lett 2017;398:24-36. [Crossref] [PubMed]
20. Lall SP, Alsafwani ZW, Batra SK, et al. ASPORIN: A root of the matter in tumors and their host environment. Biochim Biophys Acta Rev Cancer 2024;1879:189029. [Crossref] [PubMed]
21. Ding Q, Zhang M, Liu C. Asporin participates in gastric cancer cell growth and migration by influencing EGF receptor signaling. Oncol Rep 2015;33:1783-90. [Crossref] [PubMed]
22. Zhang Z, Min L, Li H, et al. Asporin represses gastric cancer apoptosis via activating LEF1-mediated gene transcription independent of β-catenin. Oncogene 2021;40:4552-66. [Crossref] [PubMed]
23. Satoyoshi R, Kuriyama S, Aiba N, et al. Asporin activates coordinated invasion of scirrhous gastric cancer and cancer-associated fibroblasts. Oncogene 2015;34:650-60. [Crossref] [PubMed]
24. Rieger ME, Sims AH, Coats ER, et al. The embryonic transcription cofactor LBH is a direct target of the Wnt signaling pathway in epithelial development and in aggressive basal subtype breast cancers. Mol Cell Biol 2010;30:4267-79. [Crossref] [PubMed]
25. Wu A, Luo N, Xu Y, et al. Exosomal LBH inhibits epithelial-mesenchymal transition and angiogenesis in nasopharyngeal carcinoma via downregulating VEGFA signaling. Int J Biol Sci 2022;18:242-60. [Crossref] [PubMed]
26. Ashad-Bishop K, Garikapati K, Lindley LE, et al. Loss of Limb-Bud-and-Heart (LBH) attenuates mammary hyperplasia and tumor development in MMTV-Wnt1 transgenic mice. Biochem Biophys Res Commun 2019;508:536-42. [Crossref] [PubMed]
27. Quan Y, Zhang H, Wang M, et al. UQCRB and LBH are correlated with Gleason score progression in prostate cancer: Spatial transcriptomics and experimental validation. Comput Struct Biotechnol J 2024;23:3315-26. [Crossref] [PubMed]
28. Jiang Y, Zhou J, Zou D, et al. Overexpression of Limb-Bud and Heart (LBH) promotes angiogenesis in human glioma via VEGFA-mediated ERK signalling under hypoxia. EBioMedicine 2019;48:36-48. [Crossref] [PubMed]
29. Wu SS, Chen J, Yan Y, et al. Limb-bud and heart as a novel biomarker for gastric intestinal type adenocarcinoma. Oncol Lett 2020;20:2209-16. [Crossref] [PubMed]
30. Eastham JA, Kattan MW, Scardino PT. Nomograms as predictive models. Semin Urol Oncol 2002;20:108-15. [Crossref] [PubMed]
31. Kattan MW. Nomograms are superior to staging and risk grouping systems for identifying high-risk patients: preoperative application in prostate cancer. Curr Opin Urol 2003;13:111-6. [Crossref] [PubMed]
32. Alsina Maqueda M, Teijo Quintáns A, Cuatrecasas M, et al. Biomarkers in gastroesophageal cancer 2025: an updated consensus statement by the Spanish Society of Medical Oncology (SEOM) and the Spanish Society of Pathology (SEAP). Clin Transl Oncol 2025;27:3580-94. [Crossref] [PubMed]
33. Li L, Zhang YH, Wang HJ, et al. ASPN was higher expression in gastric cancer and associated with poor prognosis through promoting invasion and migration and inducing macrophage M2 polarization. BMC Cancer 2025;25:1851. [Crossref] [PubMed]
34. Pinto R, Sjurgard IV, Pharo H, et al. Molecular stratification of esophageal adenocarcinoma: implications for prognosis and treatment strategy. Oncogene 2026;45:353-67. [Crossref] [PubMed]