Diverse Patient Populations Research Articles

Constructing machine learning-based risk prediction model for osteoarthritis in population aged 45 and above: NHANES 2011–2018

Osteoarthritis is a widespread chronic joint disease, becoming increasingly prevalent, particularly among individuals over the age of 45. This condition causes joint pain and dysfunction, significantly disrupting daily life. The objective of this study is to develop an optimal machine learning model for predicting the risk of osteoarthritis in individuals aged 45 and older. This study utilized data from the National Health and Nutrition Examination Survey (NHANES) from 2011 to 2018, which included a total of 2980 individuals. The dataset was randomly divided into a training set (n = 2235) and a validation set (n = 745). Five machine learning algorithms were employed to develop the predictive model for osteoarthritis. The SHapley Additive exPlanation (SHAP) method was used to interpret the machine learning algorithms and identify the most significant features for predicting outcomes. The study involved 2980 participants and focused on predicting the probability of osteoarthritis occurrence using machine learning algorithms. Five algorithms were employed, analyzing 24 features from an average 60-year-old cohort, with 605 osteoarthritis diagnoses. After performing Recursive Feature Elimination (RFE) to select 20 features, the CatBoost model achieved an AUC of 0.8109 and an accuracy rate of 0.7315, making it the most efficient model. The most influential factors in the predictions were Gender, Age, BMI, Waist Circumference, and Race. This study demonstrates that the CatBoost model with 20 features can effectively predict the occurrence of osteoarthritis. This accurate prediction model can help inform early interventions and patient management strategies, potentially improving patient prognosis. Further research will focus on enhancing the model performance, such as incorporating additional relevant features or refining existing ones. Additionally, validating the model in more diverse patient populations, and investigating its potential for real-time implementation in clinical settings would further increase the study’s impact and facilitate its translation into clinical practice.

SELF-REPORTED DEMOGRAPHICS OF 154 HYPERMOBILE EHLERS-DANLOS SYNDROME PATIENTS.

Demographic data for patients with Hypermobile Ehlers-Danlos Syndrome (hEDS) has not been well established. hEDS patients often present with a constellation of symptoms; it is important to update clinical criteria for diagnosis and provide a framework for common co-morbidities. The primary objective of this study was to identify the prevalence of co-morbidities in hEDS patients to allow clinicians to better identify patients and their most common symptoms. The goal is to use this information to augment diagnostic and clinical demographic data to more accurately represent hEDS patients. This retrospective chart review utilized patient intake forms from 154 patients diagnosed with Hypermobile Ehlers-Danlos Syndrome in a clinic. Patient intake forms included responses to questions about the presence of various current and previous conditions. Our study revealed numerous co-morbidities with increased prevalence that differed from the current diagnostic criteria including Postural Orthostatic Tachycardia Syndrome, Mast Cell Activation Syndrome, anxiety, depression, Temporomandibular Joint Syndrome, headaches, and Gastroesophageal Reflux Disease. These findings will help inform clinicians of prevalent co-morbidities among hEDS patients and encourage further evaluation for screening and diagnosis. Hypermobile Ehlers-Danlos Syndrome patients presented with a wide range of co-morbidities not listed on current clinical criteria. Future studies are warranted across diverse and larger patient populations and beyond self-reported data.

Heterozygous and Homozygous RFC1 AAGGG Repeat Expansions are Common in Idiopathic Peripheral Neuropathy.

Biallelic intronic AAGGG repeat expansions in RFC1 cause Cerebellar Ataxia with Neuropathy and Vestibular Areflexia Syndrome and may also contribute to isolated sensory neuropathy. The clinical significance of both heterozygous and biallelic RFC1 expansions in more diverse patient populations remains unclear-partly due to the absence of accurate, user-friendly computational tools specifically tailored for tandem repeat analysis. To discern the relationship between RFC1 expansions and idiopathic peripheral neuropathy (iPN), we performed whole-genome sequencing (WGS) followed by PCR-based confirmation in a large, well-characterized U.S. cohort consisting of 788 iPN patients (369 pure small fiber neuropathy (SFN), 266 sensorimotor, 144 pure sensory, and 9 pure motor). We developed an integrative pipeline combining ExpansionHunter Denovo and Expansion Hunter coupled with unsupervised clustering to reliably detect and genotype RFC1 expansions from short-read WGS data, achieving 98.2% concordance with repeat-primed PCR based validation. Biallelic RFC1 expansions were absent in 879 controls but present in 2.8% of iPN patients (Fisher's exact p = 5.9×10 -8 ), including 6.2% of pure sensory, 2.2% of SFN, and 1.5% of sensorimotor neuropathy, indicating that motor nerve involvement should not exclude patients from RFC1 repeat screening. We also observed a markedly increased frequency of monoallelic expansions in iPN compared to controls (13.2% versus 2.5%; Fisher's exact p = 3.4×10 -17 ), without evidence of secondary mutations or expansions on the other allele. Our approach provides a robust, cost-effective method for detecting RFC1 expansions from WGS data. Our findings indicate that both heterozygous and homozygous AAGGG repeat expansions in RFC1 can contribute to development of iPN.

Medicine, healthcare and the AI act: gaps, challenges and future implications

Abstract The European Union's Artificial Intelligence Act (AI Act), published in July 2024, is a pioneering horizontal regulatory framework aimed at ensuring the ethical and safe integration of AI technologies across sectors, including healthcare. While it offers the potential to improve patient care and drive innovation, it also presents challenges for healthcare providers, such as identifying high-risk applications, ensuring transparency in algorithms, and mitigating data bias. However, there are several challenges in its implementation. These include unclear guidance for certain technologies, the need to ensure fairness for diverse patient populations, effective monitoring of AI performance after deployment, and clarifying responsibility in cases of errors. Additionally, varying levels of resources among EU countries may lead to inconsistent implementation of the regulations. This article explores the core elements of the AI Act and its relevance to cardiology and identifies key gaps and unanswered questions that need to be addressed to effectively advance AI-driven medical practices.

Integrating a novel tablet-based digital neurocognitive assessment tool in brain metastases patients.

Patients with brain metastasis frequently exhibit baseline neurocognitive decline (NCD), which can be further exacerbated by CNS-directed interventions. Despite efforts to measure and preserve neurocognitive function (NCF), the practicality of traditional methods remains a concern. As a contemporary alternative, this study evaluates the feasibility and patient feedback of a novel digital neurocognitive assessment (NCA) tool. Adult patients undergoing stereotactic radiosurgery (SRS), fractionated SRS (FSRS), or whole-brain radiotherapy (WBRT) from December 2021 to March 2024 were enrolled onto a prospective registry (NCT05504681), which formed the basis for this work. The multi-dimensional, multi-language digital NCA battery was administered at baseline and follow-up, and a feedback survey was collected to assess the tool's usability and patient satisfaction. A "high risk" cohort of 74 patients with brain metastases and a "low risk" cohort of 14 patients with small benign extra-axial tumors were evaluated. The high risk cohort had higher rates of NCD compared to the low risk cohort (36.5% vs. 7.1%, p = 0.032); driven by patients treated with WBRT vs. SRS/FSRS (73.3% vs. 27.1%, p = 0.001). Notably, 50.0% completed the NCAs in their native language (non-English). Patient feedback indicated that 94% found the NCAs easy to understand, 96% found it easy to use, and 74% found it relevant to their care. This multi-dimensional, multi-language tablet-based NCA tool demonstrated feasibility and patient acceptance in a diverse patient population as well as the ability to discriminate NCD among patients treated with SRS/FSRS vs. WBRT. This approach shows promise for broader implementation, warranting further validation to confirm its efficacy.

Current Evidence Supporting the Role of miRNA as a Biomarker for Lung Cancer Diagnosis Through Exhaled Breath Condensate Collection: A Narrative Review.

Lung cancer, the leading cause of cancer-related mortality, has brought exhaled breath condensate (EBC) into focus as a promising non-invasive sample for detecting molecular biomarkers, particularly microRNAs, which regulate gene expression and contribute to tumorigenesis. Ten key studies encompassing approximately 866 subjects consistently demonstrated distinct patterns of miRNA dysregulation in lung cancer. Notably, several reported panels achieved diagnostic sensitivity and specificity exceeding 75% through the identification of distinct miRNA signatures in EBC, with oncogenic miRNAs (e.g., miR-21) upregulated and tumor-suppressor miRNAs (e.g., miR-486) downregulated in lung cancer patients. Analytical advancements, including next-generation sequencing (NGS), have improved miRNA detection sensitivity and specificity, addressing prior limitations of low yield and variability. NGS enabled the identification of novel miRNAs and proved especially effective in overcoming the low RNA yield associated with EBC samples. However, challenges persist regarding standardization of collection, sample dilution, and potential contamination. Moreover, the reproducibility of miRNA signatures across diverse patient populations remains a critical issue. Large-scale, multicenter validation studies are needed to establish robust diagnostic algorithms integrating EBC-derived miRNAs with existing clinical tools. The potential of EBC miRNA profiling to support current screening strategies could significantly improve early lung cancer detection and patient outcomes. Nevertheless, its clinical transition requires further methodological optimization and biomarker validation. This review critically evaluates current evidence on miRNA detection in EBC for lung cancer diagnosis.

Prognostic value of the triglyceride-glucose index in advanced gastric cancer: A call for further exploration.

Gastric cancer (GC) remains a leading cause of cancer-related mortality worldwide, necessitating the identification of reliable prognostic indicators to enhance treatment outcomes. Recent research has highlighted the triglyceride-glucose (TyG) index as a potential surrogate marker for insulin resistance, which may significantly influence the prognosis of patients undergoing immunotherapy combined with chemotherapy. In this context, the study by Yao et al demonstrates that a high TyG index correlates with improved overall survival and progression-free survival in advanced GC patients receiving sintilimab and chemotherapy. Specifically, patients in the high TyG group had a significantly longer median progression-free survival of 9.8 months [95% confidence interval (CI): 9.2-10.9] compared to 8.0 months (95%CI: 7.5-8.5) in the low TyG group (hazard ratio = 0.58, 95%CI: 0.43-0.79, P < 0.001). Similarly, the median overall survival was significantly longer in the high TyG group at 23.1 months (95%CI: 21.2-NA) vs 16.5 months (95%CI: 13.9-18.3) in the low TyG group (hazard ratio = 0.30, 95%CI: 0.21-0.42, P < 0.001). These findings underscore the strong prognostic potential of the TyG index in guiding treatment strategies for advanced GC. These findings underscore the need for further investigation into the TyG index's role as a prognostic tool and its underlying mechanisms in influencing treatment efficacy. We advocate for additional multicenter studies to validate these results and explore the TyG index's applicability across diverse patient populations, ultimately aiming to refine treatment strategies and improve patient outcomes in advanced GC.

Abstract 5497: Sacituzumab govitecan screening in a large organoid panel reveals the importance of functional assays for predicting ADC tumor response

Introduction: Antibody-drug conjugates (ADCs) represent an innovative approach in targeted cancer therapy, combining antibodies' specificity with therapeutic modalities such as cytotoxic or targeted chemical compounds. This study presents an ADC testing platform, leveraging patient-derived organoids (PDO) and advanced 3D high-content imaging. Methods: The efficacy of FDA-approved ADC Sacituzumab Govitecan/SN38 (SG), was evaluated in the CellTiter-Glo (CTG)-based OrganoidXplore panel, a 3D in vitro screening panel of 50 organoid models across 7 indications. The panel was characterized by NGS (WES, RNAseq) but selected without bias for TROP2/TACSTD2 expression. The OrganoidXplore screen was followed up with validating SG and payload responses using CTG and high-content imaging (HCI). ADC target engagement was assessed via immunofluorescence (IF) and ADC responses were validated in a selection of in vivo PDX models. Results: SG demonstrated a broad differential response (&amp;lt;0.01nM&amp;lt;IC50&amp;lt;50nM) in the OrganoidXplore panel, with 18 models showing strong responses (IC50&amp;lt;0.01nM). Response rates did not correlate with mRNA expression TROP2/TACSTD2. IF staining of the original PDX tumor tissues and organoids confirmed the correspondence between TROP2/TACSTD2 mRNA and protein expression levels, reaffirming that SG sensitivity did not directly correlate with gene and protein expression. Additional studies were performed with SG and SN38 in selected resistant and sensitive organoid models by HCI analysis. Similarly to the CTG assay, TROP2/TACSTD2 protein expression did not correlate with increased SG efficacy, showing that PDO sensitivity was not primarily driven by absolute target levels but rather through the mode of action of ADC uptake and payload sensitivity. Preliminary in vivo ADC efficacy results supported the in vitro findings. Conclusion: The OrganoidXplore screening platform provides insights into ADC efficacy, and responses and target engagement were validated with HCI. Combined, this demonstrates a lack of correlation between target expression (mRNA and protein) and drug efficacy, and the importance of functional testing for efficacy across a diverse patient population. The unique application of the 3D HCI platform for additional characterization, including ADC binding and uptake assays, will be an essential step to gain insight into the mode of action requirements for ADCs to be effective. These findings highlight the application of functional screening in organoid to better understand the complexity of ADC efficacy and identify responder/non-responder patient profiles to guide on resistance mechanism, in vivo modeling, and biomarker identification for patient stratification. Citation Format: Marten Hornsveld, Jinxi Wang, Jun Zhou, Xiaolong Tu, Peter van Schaik, Mariusz Madej, Hester Bange, Jessie J.J. Wang, Ludovic Bourre, Marrit Putker. Sacituzumab govitecan screening in a large organoid panel reveals the importance of functional assays for predicting ADC tumor response [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5497.

Abstract 2042: ROS1 single nucleotide variants predict favorable survival outcomes on immunotherapy regimens in non-small cell lung cancer

Abstract Background: Only a fraction of non-small cell lung cancer patients who receive immune checkpoint inhibitor (ICI) treatment will respond, largely because current FDA-approved biomarkers still leave room for improvement to identify more responders. We hypothesized that modeling the time-dependent interaction between treatment and tumor genetics may reveal novel associations. We further hypothesized that real-world data (RWD), as opposed to clinical trial data, may contain these associations, as RWD possesses a potentially more diverse patient population and time-to-event treatment history across the diverse regimens used in routine clinical practice. Methods: We examined the relation between tumor genetics, treatment, and real-world progression-free survival (PFS) in two RWD cohorts (Tempus discovery cohort, n = 1458; Dana Farber Cancer Institute Profile independent validation cohort, n = 466) by utilizing a Cox proportional hazards model with a time-dependent covariate for whether a patient was receiving an ICI-containing treatment regimen. Specifically, we included an indicator variable that denoted whether an individual was receiving an ICI regimen (=1) or not (=0), and enumerated the time intervals associated with each treatment status. We performed a univariate analysis for each mutated gene from tumor sequencing, and included a gene-treatment interaction term to permit identification of genes that were predictive (i.e. specific to a treatment) rather than prognostic. Our model included additional covariates to control for known predictive (tumor mutation burden [TMB]) and prognostic factors (age, sex). Results: In the discovery cohort, we identified a significant association between ROS1 single nucleotide variants (SNV; prevalence = 5%) and favorable PFS on ICI therapy (interaction HR = 0.34, 95% CI = 0.18-0.64, p = 7.8e-4, q = 0.08) versus non-ICI treatment. This association was not explained by ROS1 rearrangement or ROS1 TKI therapy. ROS1 SNV associated with worse PFS when not on ICI treatment (marginal HR = 2.3, 95% CI = 1.4-3.6, p = 4e-4, q = 0.044). This ROS1 SNV association was replicated in the validation cohort, which exhibited a favorable interaction with ICI treatment (interaction HR = 0.11, 95% CI = 0.04-0.33, p = 7.46e-5), and poor PFS on non-ICI treatment (marginal HR = 5.39, 95% CI = 2.5-11.6, p = 1.96e-5). ROS1 SNV treatment interaction effect was also not explained by TMB or PD-L1 status. Finally, we observed the ROS1 SNV ICI-specific protective effect was specific to patients ICI-treated in the first line setting. Conclusions: We identified an association between ROS1 SNVs and favorable PFS specific to ICI treatment, independent of common pre-existing predictive biomarkers for ICIs. Because this association was specific to the first line ICI setting, our finding may clarify ICI usage where clinical practice for ICI treatment is still evolving. Citation Format: Yue Wu, Jaclyn LoPiccolo, Sebastian Carrasco Pro, Krishna C. Bulusu, Etai Jacob, Alexander Gusev, Gerald J. Sun. ROS1 single nucleotide variants predict favorable survival outcomes on immunotherapy regimens in non-small cell lung cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2042.

Abstract 4650: The impact of GLI3 mutations and expression on prognosis in gastric cancer

Abstract Background: GLI3, a member of the GLI zinc finger family, has been extensively investigated in various cancers, including gastric cancer. It has been observed that GLI3 is significantly upregulated in solid cancer types. These findings suggest the involvement of GLI3 in the malignant progression of different cancers. However, the precise impact of GLI3 mutations and expression on the prognosis of gastric cancer patients remains unclear. Further investigation is warranted to elucidate the prognostic significance of GLI3 alterations in gastric cancer. Methods: This study included 269 gastric cancer patients from TCGA and 2903 patients from China. The mutation rate of GLI3 was analyzed using Excel. Additionally, the survival rates of gastric cancer patients with GLI3 mutations were compared to those with wild-type using KM survival curves in the TCGA cohort. In addition, the impact of high and low expression of GLI3 on survival was analyzed using a 50% threshold. The expression of GLI3 in tumor tissues were evaluated using Mann-Whitney analysis to compare with those in normal tissues. Results: The mutation rate of GLI3 was 13.55% in the TCGA cohort, whereas it was significantly lower at 7.79% in the Chinese cohort. KM curve analysis demonstrated that patients with GLI3 mutations in the TCGA cohort had a significantly higher survival rate than those with the wild group (p &amp;lt; 0.05). Additionally, GLI3 expression was significantly higher in tumor tissues compared to normal tissues (p &amp;lt; 0.01). Furthermore, based on GLI3 expression analysis, patients with a 30% higher expression of GLI3 had a significantly worse prognosis compared to the remaining 70% of patients (p &amp;lt; 0.05). The analysis of GLI3 mutations and expression revealed that patients with GLI3 mutations had significantly lower RNA expression compared to those with the wild-type. Conclusion: The mutation rate of GLI3 was found to be higher in the TCGA cohort compared to the Chinese cohort. Both GLI3 mutations and expression were closely associated with the prognosis of gastric cancer patients. These findings highlight the significance of GLI3 as a potential prognostic indicator in gastric cancer. Further research is needed to elucidate the underlying mechanisms and validate these observations in larger and more diverse patient populations. Citation Format: Wanfu Zhang, Wendi Jin, Hongwei Wan, Chao Niu, Chu Yan, Chunxiao Wang, Juanjuan Liu, Xiaogang Li, Qinglan Zong, Ning Gao, Ding Zhang. The impact of GLI3 mutations and expression on prognosis in gastric cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4650.

Abstract 5545: Dynamic Precision™: a scalable tech-bio platform to identify immune-mediated resistance through dynamic monitoring

Abstract Background: Understanding immune-mediated resistance requires recognizing the immune system as a complex adaptive system and studying its dynamic responses to drugs in diverse patient populations. Current approaches are hindered by (1) reliance on static, pre-treatment tumor analyses or preclinical models, and (2) sampling bias, as 80% of cancer patients are treated in community settings without enrollment in translational research. We hypothesized that combining dynamic analysis of immune response on treatment with broader patient representation would reveal previously unrecognized resistance mechanisms. Methods: We developed a tech-bio platform to overcome these limitations. First, applying complex system insights, we pioneered Dynamic Precision™, a AI/ML-based approach to identify emerging resistance targets and blood-based biomarkers from longitudinal patient data. Second, we developed ApricityCare, a HIPAA-compliant virtual care service, to engage diverse populations outside academic institutions in research under the SMART (Serial Molecular and Clinical Annotation of Responses to Therapies) protocol. SMART is IRB-approved for serial liquid biopsy collection, incorporating ApricityCare to enable research-grade data collection. ApricityCare offers 24/7 symptom monitoring and virtual management by nurses augmented with a clinical intelligence co-pilot, ensuring standardized side effect management, documentation, and automated data capture. Coupled with mobile phlebotomy, SMART reduces participation burdens for providers and patients. Results: We analyzed longitudinal clinical and serial plasma proteomic data from melanoma patients on standard-of-care ICI therapy. Dynamic Precision™ identified early on-treatment signatures predictive of treatment failure and uncovered dynamically regulated immune proteins as potential therapeutic targets. Triangulation with clinical insights revealed druggable immune regulators with validated resistance mechanisms to T-cell checkpoint inhibitors. In parallel, ApricityCare demonstrated its ability to provide proactive symptom monitoring and side effect management while capturing high-quality clinical data without increasing workload for in-clinic care team. Leveraging mobile phlebotomy, SMART has been piloted successfully in multiple states. Conclusions: We demonstrate here a dynamic approach to identifying immune-mediated resistance mechanisms, with a blood-based biomarker strategy for longitudinal monitoring to detect the molecular signals of resistance months before clinical progression of disease. With ApricityCare-enabled SMART protocol, we have also a scalable platform to bridge the representation gaps, generating more inclusive datasets to elucidate a more diverse range of drug resistance mechanisms. Proof-of-concept data will be presented. Citation Format: Ronan O'Hagan, Lynda Chin. Dynamic Precision™: a scalable tech-bio platform to identify immune-mediated resistance through dynamic monitoring [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5545.

Abstract 3604: Diabetes and prostate cancer-specific mortality in the Multiethnic Cohort Study: Looking deeper into diagnostic timing and disease duration

Abstract Background: Diabetes is found to be associated with higher cancer-specific mortality for several cancers, but the findings have been conflicting for prostate cancer (PCa). The goal of this study was to evaluate the association between diabetes and PCa-specific mortality (PCSM) in a racial/ethnically diverse population of PCa patients. Methods: The association between diabetes and PCSM was assessed in 10, 464 African American (20.9%), Latino (23.6%), Native Hawaiian (5.4%), Japanese American (28.2%), and non-Hispanic White men (21.9%) diagnosed with incident PCa in the Multiethnic Cohort (MEC). Based on questionnaires and administrative data, diabetes status was defined as non-diabetic, pre-diagnostic diabetic, and post-diagnostic diabetic (relative to PCa diagnosis). In a subsequent analysis, diabetes status was further categorized given the timing of diabetes diagnosis (0-5 or &amp;gt; 5 years pre- or post-diagnosis of PCa). Sub-hazard ratios (SHRs) and 95% confidence intervals (CIs) were estimated using the Fine and Grey subdistribution hazards models, with time since PCa diagnosis as the time metric and adjustments on potential confounders. All analyses were performed in all PCa patients and stratified by tumor stage (localized, advanced) and Gleason score (high-grade, low-grade). Results: During an average follow-up of 8.5 years, 1, 311 (12.5%) men with PCa died of the disease. Compared to men without diabetes, men with pre-diagnostic diabetes had a significantly lower PCSM (SHR=0.73, 95% CI: 0.62-0.86) and the reduction in PCSM appeared to be greater for men with post-diagnostic diabetes (SHR=0.58, 95% CI: 0.48-0.70). Similar trends of the associations for pre- and post-diagnostic diabetes were observed among patients with high-grade (SHR 0.72 vs. 0.51) or advanced (SHR 0.72 vs. 0.56) disease. Among patients with low-grade or localized disease, the inverse associations with PCSM were similar for men with pre- (SHR 0.70-0.76) or post-diagnostic diabetes (SHR 0.65-0.75). Further categorization given the timing of diabetes diagnosis did not lead to differences in trends of the associations for pre- and post-diagnostic diabetes on PCSM among all, high-grade and advanced PCa cases. Among low-grade or localized cases, the inverse association was more evident among those with 5+ years of diabetes before cancer diagnosis and those who became diabetic within 5 years of cancer diagnosis (SHR 0.48-0.69 vs. 0.81-0.91). Statistical significance was defined as P &amp;lt; 0.05. Conclusions: We found that both pre- and post-diagnostic diabetes were significantly associated with a lower PCSM among PCa patients, and the association appeared to be stronger for post-diagnostic diabetes. Further studies are needed to confirm these findings in other multiethnic populations and to understand the mechanisms driving these associations. Citation Format: Wei (Melody) Xiong, David Bogumil, Peggy Wan, Gertraud Maskarinec, David V. Conti, Lynne R. Wilkens, Loïc Le Marchand, Christopher A. Haiman, Fei Chen. Diabetes and prostate cancer-specific mortality in the Multiethnic Cohort Study: Looking deeper into diagnostic timing and disease duration [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3604.

Abstract 6943: RV-1521: A broad-spectrum mRNA vaccine targeting KRAS mutations in cancer

Abstract KRAS mutations, particularly G12 and G13 variants, are common in cancers but remain difficult to target therapeutically. While KRAS G12C inhibitors, such as sotorasib and adagrasib, have made significant progress, their mutation-specificity excludes up to 85% of KRAS-mutant cancers, including those with G12D, G12V, and G13D mutations. To address this gap, we developed RV-1521, a broad-spectrum mRNA vaccine targeting multiple KRAS mutations, including G12D and G13D, and evaluated its efficacy in a CT26 colon cancer model in Balb/c mice. RV-1521 induced robust immune responses, with serum IgG titers increasing 3.5-fold (p &amp;lt; 0.01). B cell epitope mapping identified four immunogenic KRAS regions. T cell profiling revealed activation hotspots in peptides P1, P35, and P36, with significant IFN-γ secretion in response to both wild-type and G12D-mutant KRAS epitopes. Flow cytometry analysis showed a 2.7-fold increase in IFN-γ+ CD4+ T cells in splenocytes from vaccinated mice. These results highlight RV-1521's capacity to activate diverse immune mechanisms targeting KRAS-mutant tumor elimination, particularly through CD4+ T cell-mediated responses. As a monotherapy, RV-1521 demonstrated dose-dependent antitumor efficacy, reducing tumor volume by 50% (p &amp;lt; 0.001) and tumor mass by 45%. This was accompanied by a 3-fold increase in CD8+ T cell infiltration and a 2.8-fold increase in granzyme B expression. These findings were consistent with reduced tumor proliferation and enhanced T cell-mediated cytotoxicity. When combined with anti-PD-1 therapy, RV-1521 further amplified its efficacy, achieving a 70% reduction in tumor volume (p &amp;lt; 0.001) and a 2.5-fold increase in tumor IFN-γ levels, while preserving cytotoxic T cell activity and reversing T cell exhaustion. Transcriptomic analysis revealed upregulation of immune pathways, including T cell receptor and Th1/Th2 signaling, alongside downregulation of key oncogenic drivers such as EGR1, FOS, and JUN. Strain-specific analysis revealed stronger immune responses in Balb/c mice compared to C57BL/6 mice, suggesting that genetic background may influence immune activation. Notably, RV-1521 elicited immune responses against both wild-type and mutant KRAS, addressing tumor heterogeneity and minimizing immune escape. The unexpected cytotoxic role of CD4+ T cells further underscores the therapeutic potential of this vaccine. By targeting multiple KRAS mutations, RV-1521 overcomes the limitations of mutation-specific therapies. Its ability to synergize with anti-PD-1 therapy and elicit a robust, multi-faceted antitumor immune response positions it as a promising strategy for treating KRAS-driven cancers. Ongoing studies will evaluate its safety and adaptability across diverse microenvironments and patient populations. Citation Format: Renxiang Chen, Yongsik Bong, Wei Liu, Ezra Chung, David Brown, Evan Lewoczko, Zachary Dorsey, Steven Long, Cun Yu Wang, Dong Shen. RV-1521: A broad-spectrum mRNA vaccine targeting KRAS mutations in cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 6943.

Abstract 7288: Targeted degradation of mobile immune checkpoints to advance cancer therapy

Abstract Immune checkpoint inhibitors (ICIs) targeting T cell-intrinsic checkpoints, such as CTLA-4 and PD-1, have shown promising outcomes across diverse cancer patient populations. However, resistance to these ICIs remains prevalent, especially in patients with an immunosuppressive tumor microenvironment (TME), suggesting the existence of additional immune checkpoints that act through orthogonal mechanisms. To overcome therapeutic resistance, there is an unmet need to identify novel immune checkpoints involved in cancer progression and to develop innovative therapeutic strategies capable of reprogramming the TME by neutralizing tumor-promoting factors. Here, we identify sialic acid-binding immunoglobulin-like lectin (Siglec)-7 and -9 as mobile immune checkpoints acquired by T cells from myeloid cells via trogocytosis in the TME. These trogocytosed Siglec checkpoints suppress T cell activity by dephosphorylating T cell receptor (TCR)-related signaling pathways through engagement with tumor-associated sialoglycans. To target these mobile checkpoints, we leveraged sulfur(VI) fluoride exchange (SuFEx) click chemistry to develop a high-affinity Siglec-7/-9 ligand and converted it into a degrader for lysosomal degradation of both Siglecs. This degrader efficiently reduced Siglec levels in T cells and myeloid cells, significantly restoring T cell function and enhancing anti-tumor immunity. The degrader, particularly when combined with anti-CTLA-4, remodeled the TME and generated durable T cell memory, resulting in excellent tumor control in multiple murine models. These findings underscore the significance of degrading exogenously acquired Siglec checkpoints on T cells as a novel strategy to overcome resistance to established ICIs in cancer therapy. Citation Format: Chao Wang, Yingqin Hou, Jaroslav Zak, Kelli A. McCord, Qinheng Zheng, Shereen Chung, Ruben D. Peraza, James C. Paulson, John R. Teijaro, Xu Zhou, K Barry Sharpless, Je!rey V. Ravetch, Matthew S. Macauley, Peng Wu. Targeted degradation of mobile immune checkpoints to advance cancer therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 7288.

Abstract 2461: Federated learning enables multi-institution collaboration without sharing Hand E slides

Abstract Hematoxylin and eosin slides provide vital visual information on a collected tissue sample and are a key component of cancer diagnostics. Previous work in computational pathology has utilized deep learning architectures to accurately and efficiently tackle complex tasks such as tumor classification. However, the generalizability of these algorithms is currently limited due to the inherent difficulty of cross institution collaboration. Here, we propose leveraging federated learning (FL) which aggregates model level information between institutions to enable collaboration without exposing sensitive data. This allows the deep learning model to incorporate diverse patient populations and heterogeneous data sources which are needed for generalizability while preserving patient privacy. As a proof-of-concept, we used attention based multiple instance learning within the FL framework to perform cancer subtype classification between lung adenocarcinoma and lung squamous cell carcinoma with three distinct cohorts: Dana-Farber Cancer Institute (DFCI), Massachusetts General Hospital (MGH), and The Cancer Genome Atlas (TCGA). We present comparisons using three approaches: FL, institution specific analyses, and a centralized model where data is combined at one site (gold standard). We find all three approaches have comparable overall AUC. Additionally, we see that there is significant heterogeneity in AUC between cohorts which may indicate FL is more robust to overfitting and more generalizable. Comparing features generated using the foundation pathology model (UNI) to those generated from a ResNet50 model pre-trained on ImageNet (ResNet), we observed more institutional level heterogeneity in the UNI results than in ResNet which may indicate UNI is learning institution specific features that limit its generalizability. Overall our work shows FL enables generalizable deep learning across institutions without risking patient privacy. Citation Format: Kodi Taraszka, Oded Rosolio, Intae Moon, Yevgeniy Semenov, Alexander Gusev. Federated learning enables multi-institution collaboration without sharing Hand E slides [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2461.

Abstract 4632: Molecular determinants to neoadjuvant chemotherapy resistance in breast cancer patients: Insights from a Latin-American cohort

Abstract Neoadjuvant chemotherapy (NAC) is vital in treating primary breast cancer, yet the mechanisms behind its success are not fully understood. Understanding transcriptomic changes associated with NAC resistance is crucial for improving treatment. This study sought to identify transcriptomic alterations linked to NAC resistance, enhance comprehension of treatment resistance, and lay the groundwork for developing targeted strategies for patients unresponsive to standard NAC protocols.We analyzed the transcriptomes of 44 matched breast cancer specimens from nonresponder patients to NAC at the Colombian National Cancer Institute. Samples were collected before and after NAC, representing luminalA, luminalB/HER2-positive, luminalB/HER2-negative, and TNBC subtypes. We performed differential gene expression and enrichment analyses and compared hallmark gene sets. Furthermore, we used deconvolution algorithms to estimate tumor microenvironment (TME) infiltration and assess immune population activity within the TME. Our findings were compared with publicly available GEO database data.Gene expression patterns before and after treatment were distinct across subtypes, with FOS and NR4A1 common to all, and ZBTB16 unique to luminalB. FOSB was found in all luminal samples. ATF3, DUSP1, EGR1, and JUN were shared among HER2-negative subtypes. GEM was common to luminalA and LuminalB/HER2-negative subtypes. Pathway variations included complement and coagulation cascades, and IL-17, TNF, MAPK, and NF-κB signaling pathways among subtypes and post-NAC specimens. Post-treatment LuminalA samples had increased CD4 memory cells, Tregs, and neutrophils. The LuminalB/HER2+ group showed higher CD8+ and CD4+ memory-activated cells post-treatment. In LuminalB/HER2- post-treatment samples, B cell CD4 memory resting, Macrophages M2, and neutrophils were elevated. TNBC post-treatment samples had higher resting mast cells and CD4+ memory cells. MSIGBD comparison of gene sets linked to cisplatin, docetaxel, and doxorubicin resistance identified several shared genes. GeoDatabases confirmed FOS and NR4A1 across all groups, with other DEGs shared among subtypes.Our research indicates that NAC affects gene expression, cell function, and interactions within the TME, potentially contributing to treatment resistance. Assessing TME components is crucial for predicting NAC effectiveness and reducing the risk of recurrence. These results highlight the importance of studying diverse patient populations to gain fresh perspectives and address disparities in treatment outcomes. This investigation emphasizes the need to understand how TME influences chemotherapy resistance, paving the way for the development of more precise and efficient therapeutic strategies. Citation Format: H Michelle Guevara-Nieto, Carlos A. Orozco, Rafael S. Parra-Medina, Jovanny Zabaleta, Liliana Lopez-Kleine, Alba L. Combita. Molecular determinants to neoadjuvant chemotherapy resistance in breast cancer patients: Insights from a Latin-American cohort [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 4632.

Abstract 1056: Exploring the impact of ancestry on metabolic dysregulation in lung adenocarcinoma: Implications for racial health inequities and targeted therapeutics

Abstract Introduction. The occurrence of lung adenocarcinoma (LUAD) is approximately 68.3 per capita for Black men and 61.5 per capita for White men, reflecting an 11% higher rate in the former. Black patients experience a poorer prognosis and significantly higher mortality rates compared to their White counterparts. Purpose. Understanding the genetic and molecular foundations that underly racial health inequities is essential for developing comprehensive interventions and advancing more precise therapeutic modalities for diverse patient populations. Single Nucleotide Polymorphisms (SNPs) can influence oncogenic transformation and may be associated with distinct molecular landscapes in tumors and adjacent TME. Furthermore, ancestry may play a pivotal role in shaping individual susceptibility and risk for mitochondrial dysfunction and ensuing oncogenic transformation. Methods. We employed germline VCFs for ancestry analysis of 36 LUAD patients (18 Black and 18 White) from the ORIEN data set, while genotype data from the 1000 Genomes Project phase 3, grouped by super population served as reference. Oncoplots were generated using a combination of patients’ somatic VCF and metadata from ORIEN. Regression analysis adjusting for age, staging, and gender was performed to determine association between ancestry and gene mutations. Global admixture was estimated using STRUCTURE v2.3.4 with k=5, NUMREPS=2000, and BURNIN=50000. VCFs were subset and merged by 1126 ancestry-informative markers (AIMs) and converted to STRUCTURE format with PLINK. PCA was conducted using R. Local ancestry analysis was performed with LAMP-LD v1.3 using a window length of 50 SNPs and 20 states, based on 20, 803 SNPs. Results were visualized with ggplot2 in R. Results. In Black patients, admixture analysis showed predominant AFR genetic similarity with varying proportions of EUR, AMR and SAS similarity. Local ancestry analysis using LAMPLD reveals admixture across different regions of the genome illustrating the variation in ancestral contributions within and between individuals of the same cohort and compared to counterparts in other populations. Mutations in ACADVL, HK2, SREBF1, COX6L, BID, POLG2, SLC2A2, SLC22A5, and ACKR1 are associated with AFR status &amp;gt;/= 0.5. Mutations in ACADSB, HK3, PPARGC1A are associated with AFR status &amp;lt; 0.5. Conclusions. Mutations in metabolic and mitochondrial genes emerged as being highly enriched in LUAD patients with high relative to low African genetic ancestry, suggesting opportunities for improved risk-prediction and therapeutic targeting in this subpopulation. Citation Format: Benecia A. Jackson, Yuxin Jin, Yonatan Amzaleg, Angel Perez Hunt, Nikita Jinna, Dan Raz, Loretta Erhunmwunsee, John Carpten, Aaron Neely. Exploring the impact of ancestry on metabolic dysregulation in lung adenocarcinoma: Implications for racial health inequities and targeted therapeutics [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1056.

Abstract 3586: Fusobacterium nucleatum presence and colorectal cancer-specific mortality in a racially and ethnically diverse patient population

Abstract Purpose: There are persistent racial and ethnic disparities in colorectal cancer (CRC) mortality. We examine the association between prevalence of Fusobacterium nucleatum (Fn) in tumor tissue and CRC-specific mortality in a diverse population. Methods: The present study includes 569 African American, Alaska Native, Latinx, and non-Hispanic White adults with stage I-III CRC who were part of the Translational Research Program in Colorectal Cancer Disparities (TRPCD). 188 patients with CRC who died of CRC were matched to 381 patients with CRC who did not die of CRC. Genomic DNA was extracted from colorectal tumor tissue samples. Droplet digital polymerase chain reaction (ddPCR) was used to analyze 1) the nusG gene from Fn (Fn(nusG)), and 2) the SLCO2A1 gene as a housekeeping (HK) gene reference from the human genome. Fn(nusG) positive droplet counts were normalized by HK gene positive droplet counts for each duplicated assay, and then averaged. Fn(nusG) prevalence was analyzed as a dichotomous variable in primary analyses (not present, present) and as a categorical variable (not present, low prevalence, high prevalence). The median normalized average Fn(nusG) count among those with Fn(nusG) present was the cut-point between low and high prevalence. We fit logistic regression models for the association between Fn(nusG) presence and CRC-specific mortality, adjusted for matching factors (age at diagnosis, year of diagnosis, sex, stage, tumor site, race and ethnicity). Effect modification by age, sex, stage, tumor site, and racial and ethnic group was assessed using a likelihood ratio test. Results: Fn(nusG) was present in tumors for 24% of patients with CRC who died of CRC (N=46) and 16% of patients with who did not die of CRC (N=61). Presence was lower in patients who were Alaska Native (14%), with stage I CRC (12%), and with rectal cancer (16%), and higher in those diagnosed in 2011-2017 (22%). Presence of Fn(nusG) was associated with a higher odds of CRC-specific mortality (OR: 1.76; 95% CI: 1.12-2.74), relative to no presence of Fn(nusG), after adjustment for matching factors. Estimates for the association between low prevalence (OR: 1.77; 95% CI: 0.97-3.21) and high prevalence (OR: 1.74; 95% CI: 0.95-3.14), relative to no presence of Fn(nusG), and CRC-specific mortality were similar. Further analysis is underway. Conclusions: These findings provide evidence of an association between presence of fusobacterium in CRC tumor tissue and CRC-specific mortality in a diverse population. As the association did not differ by low or high prevalence of fusobacterium, any amount of fusobacterium in CRC tumor tissue may confer poorer prognosis. Fusobacterium may contribute less to CRC-specific mortality in Alaska Native patients with CRC. Identifying microbiota associated with CRC mortality can inform targeted interventions to address racial and ethnic disparities in this disease. Citation Format: Nicole C. Loroña, Scott D. LaBrie, Claire E. Thomas, Hang Yin, Jeroen Huyghe, Conghui Qu, Sushma Thomas, Sosun Nayemi, Hamza Ammar, Orsalem Kahsai, Li Hsu, Diana Redwood, Li Li, Christopher I. Li, Ulrike Peters, Timothy K. Thomas, Amanda I. Phipps, Jane C. Figueiredo, Meredith Hullar. Fusobacterium nucleatum presence and colorectal cancer-specific mortality in a racially and ethnically diverse patient population [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3586.

Leveraging Electronic Health Records in International Humanitarian Clinics for Population Health Research: Cross-Sectional Study.

As more humanitarian relief organizations are beginning to use electronic medical records in their operations, data from clinical encounters can be leveraged for public health planning. Currently, medical data from humanitarian medical workers are infrequently available in a format that can be analyzed, interpreted, and used for public health. This study aims to develop and test a methodology by which diagnosis and procedure codes can be derived from free-text medical encounters by medical relief practitioners for the purposes of data analysis. We conducted a cross-sectional study of clinical encounters from humanitarian clinics for displaced persons in Mexico between August 3, 2021, and December 5, 2022. We developed and tested a method by which free-text encounters were reviewed by medical billing coders and assigned codes from the International Classification of Diseases, Tenth Revision (ICD-10) and the Current Procedural Terminology (CPT). Each encounter was independently reviewed in duplicate and assigned ICD-10 and CPT codes in a blinded manner. Encounters with discordant codes were reviewed and arbitrated by a more experienced medical coder, whose decision was used to determine the final ICD-10 and CPT codes. We used chi-square tests of independence to compare the ICD-10 codes for concordance across single-diagnosis and multidiagnosis encounters and across patient characteristics, such as age, sex, and country of origin. We analyzed 8460 encounters representing 5623 unique patients and 2774 unique diagnosis codes. These free-text encounters had a mean of 20.5 words per encounter in the clinical documentation. There were 58.78% (4973/8460) encounters where both coders assigned 1 diagnosis code, 18.56% (1570/8460) encounters where both coders assigned multiple diagnosis codes, and 22.66% (1917/8460) encounters with a mixed number of codes assigned. Of the 4973 encounters with a single code, only 11.82% (n=588) had a unique diagnosis assigned by the arbitrator that was not assigned by either of the initial 2 coders. Of the 1570 encounters with multiple diagnosis codes, only 3.38% (n=53) had unique diagnosis codes assigned by the arbitrator that were not initially assigned by either coder. The frequency of complete concordance across diagnosis codes was similar across sex categories and ranged from 30.43% to 46.05% across age groups and countries of origin. Free-text electronic medical records from humanitarian relief clinics can be used to develop a database of diagnosis and procedure codes. The method developed in this study, which used multiple independent reviews of clinical encounters, appears to reliably assign diagnosis codes across a diverse patient population in a resource-limited setting.

Nanoparticles as Novel Drug Delivery Systems for Cancer Treatment: Current Status and Future Perspectives.

Cancer continues to pose a significant global health challenge, demanding innovative therapeutic approaches to overcome the limitations of conventional treatments like chemotherapy and radiotherapy. Nanoparticles (NPs) have emerged as promising tools for cancer therapy due to their unique physicochemical properties that enable targeted drug delivery, reduced systemic toxicity, and enhanced therapeutic efficacy. This comprehensive review delves into the mechanisms of NP-based drug delivery, highlighting both passive and active targeting strategies. It categorizes and discusses diverse NP types, including polymeric, lipid-based, and metallic nanoparticles, emphasizing their applications in enhancing the bioavailability and specificity of anticancer agents. This review also explores the integration of advanced technologies, such as theranostics and artificial intelligence, to optimize NP design and functionality for personalized medicine. However, challenges remain, including issues related to toxicity, drug resistance, and manufacturing scalability. Addressing these barriers requires interdisciplinary research focused on developing stimuli-responsive NPs, improving biocompatibility, and incorporating multimodal therapeutic platforms. Although substantial progress has been made, this review is limited by the paucity of clinical trials validating NP efficacy and safety in diverse patient populations. Future endeavors should prioritize translational research to bridge the gap between preclinical innovations and clinical applications, ensuring that these transformative technologies benefit a broader spectrum of cancer patients. The review underscores the immense potential of NPs in redefining cancer treatment while advocating for sustained research to address existing limitations and unlock their full therapeutic promise.