ldSNVs for each GWAS index SNV (r2 ≥ 0.8)

140 Background: Perioperative chemotherapy improves survival in resectable gastroesophageal adenocarcinoma (GEA), yet many patients do not complete the intended postoperative component. Multimodal prehabilitation targets modifiable physical, nutritional, and psychosocial factors before surgery. We assessed whether prehabilitation is associated with improved adherence to perioperative chemotherapy. Methods: We conducted a single-center retrospective cohort study of patients with localized gastroesophageal adenocarcinoma treated with perioperative chemotherapy and curative-intent surgery at the McGill University Health Centre from 2007 to 2025. Patients were identified from an institutional registry, the prehabilitation clinic dataset, and the electronic medical record (OACIS). Eligibility required complete treatment and outcome data and neoadjuvant chemotherapy given with intent to deliver a full perioperative course. Multivariable logistic regression evaluated predictors of receipt of adjuvant chemotherapy. We performed a subgroup analysis of patients treated with perioperative fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT). The institutional prehabilitation program began in 2022 and included supervised exercise, nutritional optimization, and psychosocial support. Results: We included 258 patients. Twenty-four patients underwent multimodal prehabilitation. In the adjusted analyses, younger age, neutrophil-to-lymphocyte ratio less than 3, and receipt of a regimen that included immunotherapy were independently associated with receipt of adjuvant chemotherapy. In the FLOT subgroup, prehabilitation was not associated with a higher likelihood of receiving adjuvant chemotherapy, but it was associated with higher completion of the preoperative FLOT course (24 of 24, 100%) versus patients without prehabilitation (54 of 64, 84.4%). Conclusions: In this real-world cohort, multimodal prehabilitation was associated with improved completion of preoperative FLOT, but not with receipt of adjuvant therapy. Prospective studies should test whether prehabilitation can increase completion of the full perioperative course and improve outcomes in GEA patients receiving FLOT or other perioperative regimens.

GEC risk-associated index SNVs and loci.

UXS1 is a selective dependency in gastroesophageal cancers.

Abstract Introduction: Overcoming chemoresistance remains a significant challenge in esophageal adenocarcinoma (EAC). Drug repurposing has gained interest for its potential to identify new uses for approved drugs, with advantages including lower costs, established safety, and accelerated timelines. Seeing as EAC tumors demonstrate increased reliance on lipids as an alternative energy source which is associated with therapeutic resistance, we sought to investigate, through an integrated multi-omic approach, the potential of repurposing lipid-lowering drugs as a therapeutic intervention. Materials and Methods: We integrated scRNAseq datasets, a gene network-dependent drug repurposing machine learning algorithm (NeMoCAD) and Hight Throughput Drug screening (HTS) on tumor organoids (PDOs) derived from EAC patients undergoing neoadjuvant chemotherapy to investigate therapy resistance. We applied scRNAseq data in NeMoCAD pipeline to identify agents within the LINCS database (approx. 30,000 drugs), capable of reversing the chemoresistance-specific tumor transcriptional states to states toward chemosensitive or normal phenotypes. Additionally, we leveraged a clinically characterized dataset on 450 EAC patients who underwent perioperative chemotherapy followed by surgical resection. Approximately one-third of which were prescribed statins for cardiovascular reasons. PDOs from non-statin users were selected to evaluate the synergy between hydrophobic statins (atorvastatin, simvastatin) and triplet chemotherapy in vitro. Results: NeMoCAD analysis predicted that hydrophobic statins, which inhibit 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR), can shift EAC tumor gene expression patterns toward those observed in normal gastroesophageal tissues. From our retrospective cohort of patients with pre-operative chemotherapy, statin users had a significantly higher 5-yr overall survival rate (58% vs. 38%; HR 0.55, p=0.018), as well as a 26% increase in major pathological response and significant reductions in recurrence (20%) and distant metastasis (37%). This improvement in survival is despite, statin users being older, higher BMI, more likely to have severe systemic diseases compared to non-statin users. HTS experiments with PDOs demonstrated response with several clinically approved statins, and focused validation drug testing showed pronounced synergistic effect when low doses of hydrophobic statins were combined with standard of care chemotherapeutic regimes. Conclusions: Our findings combining clinical patient outcome, scRNAseq data, drug repurposing machine learning pipelines, and functional validation with PDOs based HTS suggest that statins may represent a valuable adjunct to standard-of-care treatment for EAC. Citation Format: Sanjima Pal, Luís Nuno Cruz Santos Castro, Megan Sperry, Qian Qiu, Shuyuan Wang, Mingyan (Iris) Kong, Nicholas Bertos, Betty Giannias, Cedric Julien, Wotan Zeng, Swneke Bailey, Donald E. Ingber, Lorenzo Ferri. Repurposing statins for locally advanced esophageal adenocarcinoma: Integrating scRNAseq, organoid based high throughput drug screening, and patient level clinical data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB057.

125 Background: Perioperative chemotherapy improves survival in resectable gastric and gastroesophageal junction (GEJ) adenocarcinoma; however, many patients fail to initiate postoperative therapy in routine practice. Host-related determinants of adjuvant treatment feasibility and survival remain incompletely defined. Methods: We conducted a retrospective cohort study of patients with localized gastric or GEJ adenocarcinoma treated with perioperative chemotherapy and curative-intent resection at the McGill University Health Centre (2006-2025). The primary outcome was adjuvant chemotherapy initiation. Overall survival (OS) was analyzed using Kaplan-Meier and Cox proportional hazards models; prespecified 3-month landmark analyses were performed to mitigate immortal time bias. Sequential multivariable logistic models evaluated tumor variables, host factors (age, ECOG, serum albumin), and postoperative inflammatory recovery (ΔlogNLR), with incremental discrimination assessed by AUC. Results: Among 248 patients (median age 64; 85% male), 168 (67.7%) initiated adjuvant chemotherapy, which was associated with improved OS (HR 0.45; 95% CI, 0.33–0.62), including in landmark analyses. Pathology-only modeling showed modest discrimination (AUC 0.66). Addition of host factors improved discrimination (AUC 0.75); increasing age was independently associated with lower odds of initiating adjuvant therapy (OR 0.94 per year; 95% CI, 0.89–0.98), whereas higher serum albumin was independently associated with greater odds of initiation (OR 1.09 per g/L; 95% CI, 1.01–1.18). In a prespecified nested analysis among 69 patients with available longitudinal inflammatory data, model discrimination improved sequentially from AUC 0.65 to 0.77 to 0.80 with addition of host factors and ΔlogNLR. Tumor factors and chemotherapy regimen were not independently associated with adjuvant initiation after adjustment. Conclusions: In localized upper gastrointestinal adenocarcinoma, postoperative chemotherapy feasibility and survival are more strongly associated with host factors and recovery than with tumor pathology or regimen selection. These findings support perioperative optimization strategies, including nutritional support, prehabilitation, and interventions targeting systemic inflammation, to improve treatment delivery. Incorporating host vulnerability into risk stratification may refine treatment planning beyond pathological response alone. Prospective validation of dynamic inflammatory biomarkers and targeted host-directed interventions is warranted to determine whether improving treatment feasibility translates into survival benefit.

Verification of the inducible UXS1 expression and knockdown by RT-qPCR.

SNV filtering using HiChIP data referenced in manuscript

Abstract Introduction: Gastroesophageal Adenocarcinoma (GEA) is among the top five malignant tumors with a 5-year survival of < 20% and limited improvement in clinical management over recent decades. Currently, docetaxel-based triplet chemotherapy remains the standard-of-care treatment. However, approximately 40% of patients show resistance to chemotherapy up front, and half of initial responders develop resistance on treatment. Identifying predictors of response to chemotherapy remains a key unmet clinical need. Single-cell transcriptomic profiling enables high-resolution characterization of gene expression changes at the cellular level. It holds significant promise as a new approach to address the clinical challenge. Given the lack of reliable predictive or prognostic biomarkers, this study aims to identify biomarkers predictive of treatment response to support clinical decision-making. Materials and Methods: Tissue samples were obtained from GEA patients, then processed and prepared for single-cell RNA sequencing (scRNA-seq). Gene transcripts and cells identified from scRNA-seq were clustered and annotated. Overall survival (OS) and disease-free survival (DFS) analyses were conducted to assess whether specific biomarkers were associated with survival outcomes. Immunohistochemistry (IHC) and immunofluorescence (IF) staining were performed to validate identified gene markers. Gene Set Enrichment Analysis (GSEA) and Pseudotime analyses were carried out to elucidate the underlying mechanisms of chemoresistance. Results: Our scRNA-seq analysis was performed on samples from 43 GEA patients, annotated with their pathological treatment response to chemotherapy assessed using Tumor Regression Grade (TRG). Among these, 10 patients were good responders (TRG 0 or 1), 25 were moderate responders (TRG 2), and 8 were poor responders (TRG 3). FOXA2 was identified as one of the most highly expressed genes in tumor epithelial cells. High FOXA2 expression before treatment was significantly associated with longer OS and DFS, and its expression was validated on primary tissue samples and matched patient-derived organoids. FOXA2 expression was significantly higher in treatment-naïve tissues from pathological good responders than in poor responders. Additionally, differential FOXA2 expression was associated with distinct cellular states, with cells expressing low levels of FOXA2 showing enrichment of HALLMARK pathways including epithelial-mesenchymal transition and DNA repair. Conclusions: Our results identify FOXA2 as a predictive biomarker of response to neoadjuvant standard-of-care chemotherapy in patients with GEA. Our data further suggest that FOXA2 may function as a tumor suppressor through inhibition of epithelial-mesenchymal transition. Citation Format: Shuyuan Wang, Qian Qiu, Sanjima Pal, Ece Ozmen, Ruoyu Ma, Cédric Julien, Wotan Zeng, Betty Giannias, France Bourdeau, Nicholas Bertos, Swneke Bailey, Lorenzo Ferri. FOXA2-associated chemoresistance in gastroesophageal adenocarcinoma: A single-cell transcriptomic exploratory analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB316.

Abstract Topic Esophageal Cancer: Adjuvant and Neo-Adjuvant Therapies Background Esophageal cancer is the seventh cause of cancer mortality worldwide. Despite recent advances in the perioperative management of esophageal adenocarcinoma with systemic and immunotherapy, significant therapy resistance persists. New approaches are in demand, and repurposing approved drugs for other indications is an attractive and low-cost approach to address this issue. Interfering with the adaptive metabolism of cancer cells holds promise, but complexity and redundancy makes it challenging to identify specific targets. Lipid and cholesterol metabolism play a key role in cancer cell homeostasis. Statins are ubiquitous cholesterol-lowering oral agents that could have a role in esophageal malignancy. Methods With a focus on esophageal adenocarcinoma (EAC), inputs from our single center highly annotated clinical database and biobank comprising more than 400 esophageal cancer patients were integrated. We used single-cell RNA sequencing data from a select cohort and employed an artificial intelligence-driven platform called NeMoCAD (Network Model for Causality-Aware Discovery) to predict compounds capable of reversing cancer cells transcriptomic state. High-throughput drug screening and synergy assays were employed to assess the in vitro effect of statins alone or in combination with chemotherapy on a patient-derived organoid cancer model. Parallel to this, we collected retrospective data from our prospectively maintained clinical database for analysis of demographics, tumor characteristics, response to systemic therapy and survival, grouping patients by statin use status. Results: With the NeMoCAD algorithm several statins were predicted to potentially alter the transcriptomic profile of chemoresistant tumours to resemble that of normal surrounding tissues, notably simvastatin and atorvastatin. Drg screening and synergy assays showed statins alone displayed variable cytotoxicity and that overall interaction patterns with chemotherapy were additive, but yielding peak synergy scores at low concentrations. From our clinical database, 449 locally advanced EAC patients receiving curative-intent neoadjuvant chemotherapy followed by surgery were included. A quarter of patients were on cholesterol-lowering statins for cardiovascular indications (n=128). They were significantly older and had higher BMI. A 30% increase in major pathological response was seen in statin users (NS) as well as significantly improved overall survival (5-year OS 54.1% vs 40.7%, HR 0.69, p=0.019). This impact was more pronounced for those receiving current standard-of-care FLOT. Of note, lipophilic statins (e.g., Atorvastatin, Simvastatin) showed better survival outcomes than hydrophilic statins. Conclusion In EAC, statins may play a role in improving response to neo-adjuvant systemic chemotherapy and may lead to prolonged survival in this population. Our clinical observations are in accordance, with increased overall survival (HR 0.69) for statin users. Moreover, these observations validate the predictions of in silico models employed, further corroborated by in vitro synergy essays. Confirmation of these findings is planned within a phase 2 clinical trial.

Expression of UXS1 in normal and tumor tissues.

List of genes and their LRT scores surrounding each GWAS index SNV.

A major limitation in studying gastroesophageal adenocarcinoma (GEA) has been the lack of reliable models that represent the disease’s complexity. We present lessons learned from a comprehensive large-scale biobanking effort combining traditional sample collection with several in vitro models, including 3-dimensional patient-derived organoids (PDOs), 2-dimensional cancer-associated fibroblasts (CAFs), tumor-infiltrating lymphocytes (TILs), and/or in vivo xenografts. This initiative started in 2018, integrating multiple advanced ex vivo models such as PDOs, patient-derived xenografts (PDXs), and organoids (PDXOs). This unique resource now includes tumor avatars from over 380 consented patients, making it the world’s largest living GEA biobank. We achieved a >90% success rate in creating per-patient models, including 227 tumor-derived and 203 neighboring normal PDOs. These organoids accurately mirror key features of the original tumors, such as their histology (e.g., microsatellite instability), mutations, and drug response across treatment points. Notably, PDOs can predict individual patient responses to chemotherapy within five weeks, underscoring their clinical relevance. Furthermore, high-throughput drug screening on PDO subsets with known genetic landscapes generates personalized chemosensitivity profiles for 22 drugs. Through a process of continued refinement of culture techniques and tumor sampling approach, our large-scale comprehensive collection of GEA avatars represents a unique and valuable preclinical experimental resource for precision oncology.

Verification of the reference and variant sequences of the luciferase constructs.

The rs147518036 SNV genotypes for the gastric cancer cell lines.

Chronic lung injury generates metaplasia which occasionally, but ominously, progresses to squamous dysplasia and squamous lung cancer. To identify mechanisms through which disrupted tissue homeostasis contributes to malignant initiation and progression, we used in vivo and in vitro heterotypic recombinant models of human b ronchial e pithelial c ells (hBECs) and fibroblasts. We demonstrate that injury-associated TGF-β signaling creates a fibroblast state dependent upon HSP47 upregulation. These fibroblasts accumulated collagen, thus elevating tissue stiffness and activating mechanosignaling that sustained YAP-dependent embryonic-like, pro-malignant activities in adjacent hBECs. This S tress/ T ension-Instructive F ibroblast (STIF) state, exhibited by stressed fibroblasts in premalignant and malignant lesions across multiple cancer types, was sufficient to reprogram disease-free hBECs to metaplasia and to drive hBECs with compromised tumor suppressor function to dysplasia, yet could be inhibited and reversed. STIFs suffice to activate epithelial phenotypes reminiscent of oncogene-mediated cell transformation and induce (pre)malignancy via increased force transmission, providing novel targets for prevention. Statement of significance:Tissue injury creates a regenerative pro-tumorigenic S tress/ T ension-Instructive F ibroblast (STIF) state which is sufficient to activate a YAP-dependent, pre-malignant program to induce or unmask pre-cancerous phenotypes in epithelial cells through mechanotransduction. Inhibition of STIF activity or mechanosignaling prevents metaplasia and progression to dysplasia. Highlights:Tissue injury creates a pro-tumorigenic Stress/Tension-Instructive Fibroblast (STIF) state in multiple organs that precedes and persists through cancerSTIF signaling alone, working through fibroblasts and not epithelial cells, is sufficient to activate embryonic-like plasticity and induce epithelial pre-cancerous metaplastic lesionsSTIFs program (pre)malignant phenotypes in adjacent epithelial cells through mechanosignaling by activating YAP prior to tumor formationInhibiting STIFs or mechanosignaling prevents/reverts metaplasia and prevents progression to dysplasia.

We present the clinical results of a phase 2 trial combining neoadjuvant docetaxel, cisplatin, 5 Flourouracil, and the PD-L1 inhibitor avelumab in locally advanced gastro-esophageal adenocarcinoma (GEA). Fifty-one patients receive neoadjuvant therapy with 50 proceeding to surgery. Grade 3-4 adverse events occur in 40%; complete/major pathological response is found in 7/50 (14%) and 9/50 (18%), with 2-year disease-free survival of 67.5%. There is no correlation between tumor regression and PD-L1 or mismatch repair (MMR) status. Multiplex immunohistochemistry and longitudinal single-cell transcriptomic profiling reveal alterations in certain innate immune cell populations, particularly noting an M2-tumor-associated macrophage (M2-TAM) proliferation in non-responding tumors. These findings describe the effective nature of this treatment regimen for GEA and reveal associated features of the inflammatory milieux associated with response to chemo-immunotherapy. The specific character of the inflammatory environment in non-responders may, in the future, help personalize treatment. This study was registered at ClinicalTrials.gov (NCT03288350).