eScholarship@UMassChan
eScholarship@UMassChan is a digital archive for UMass Chan Medical School's research and scholarship, including journal articles, theses, datasets and more. We welcome submissions from our faculty, staff, and students. eScholarship@UMassChan is a service of the Lamar Soutter Library, Worcester, MA, USA. See also our open access journal publishing services.
Questions? See the Help menu or contact escholarship@umassmed.edu.
Communities in eScholarship@UMassChan
Select a community to browse its collections.
Recent Publications
Publication Integrating Symbolic Regression for Generalizable and Interpretable Machine Learning in Cardiovascular Risk Prediction(2025-11-18)Introduction: Cardiovascular disease (CVD) risk prediction is crucial for timely, targeted risk factor modification. Machine learning (ML) is often applied to CVD risk prediction to improve diagnosis and treatment but with modest performance [1] and limited interpretability - often requiring large datasets with a risk of overfitting. Enhancing the generalizability and interpretability of CVD risk prediction ML models is critical for their broad adoption and clinical deployment. We explored the integration of symbolic regression (SR) with random forests (RF) to address this challenge. Methods: We analyzed 518,389 patients in the UMass Memorial Clarity data lake research database. These patients had at least one clinical encounter between Oct 1, 2017 and Nov 1, 2018, without any indication of death or CVD during that time based on ICD-10 codes. [2] Demographic features (e.g., age, race/ethnicity) and known risk factors (e.g., diagnoses, lab results) were extracted from the electronic health records. Pre-processing included examination of missing/spurious data, imputation using MissForest, one-hot encoding, and aggregation of laboratory and diagnostic features by condition and measured quantity. We developed and validated a gpu-accelerated SR-enhanced RF (SReRF), motivated by prior work hybridizing SR with decision trees, [3] to predict 5-year CVD risk. Performance metrics included precision, recall, F1, area under the ROC curve (AUROC), and area under the precision-recall curve (AUPRC). We did a 75-25 split for training/testing data and tuned an RF using cross validation and grid search. Results: The patient cohort had a mean (SD) age of 48.6 (18.7), with 5.8% CVD events observed in the 5-year window. SReRF yielded superior performance compared with classical RF. Specifically, classical RF had an AUROC (SD) of 0.79 (0.0027), AUPRC 0.18 (0.0027), and F1 0.27 (0.0032), compared with SReRF’s AUROC of 0.82 (0.0023), AUPRC 0.20 (0.0034), and F1 0.29 (0.0034). Here, SDs were computed using 500 bootstrapped samples of the test set. Visual examination of PR curves indicated that the SReRF’s performance improvement was substantial for thresholds associated with lower recall values, but modest for higher recall thresholds. Also, whereas the classical RF overfit the training set, the SReRF had nearly identical performance on both training and test sets. The main predictors in SReRF (higher in the decision trees) were age, HDL and LDL cholesterol, hypertension, atrial fibrillation, chronic obstructive pulmonary disease, obesity, albumin, creatinine, and self-reported Hispanic/Latinx ethnicity. Discussion and Conclusion: This study demonstrates promising results of integrating symbolic regression with random forest for CVD risk prediction. We found that SReRF outperformed classic RF. The top SReRF predictors align with known risk factors for CVD, adding face validity and reinforcing its potential clinical utility. Further, the symbolic expressions that determined splits produced compact and meaningful relationships between predictors, improving model interpretability. Expressions revealed the simultaneous presence of related comorbidities (e.g., hyperlipidemia and hypertension) and non-linear relationships between numerical predictors (e.g. age and LDL). Of note, ethnicity (Hispanic/Latinx) among identified main predictors may reflect underlying disparities in social determinants of health or biological risk factors. Despite recognized advantages of SR, limitations exist, including long training time, difficulty handling missing data, and limited scalability with high-dimension features. Future models will address these issues, include enhanced pre-processing, computational optimization, additional clinical features (e.g. medications), and be evaluated against other contemporary ML techniques. In conclusion, SReRF represents a promising, interpretable, and generalizable approach for CVD risk prediction, with strong potential for implementation in clinical settings, especially those with a high demand for transparency in predictive modeling. References 1. Soares C, Kwok M, Boucher K, et al. Performance of Cardiovascular Risk Prediction Models Among People Living With HIV: A Systematic Review and Meta-analysis. JAMA Cardiol. 2023;8(2):139–149. 2. https://vsac.nlm.nih.gov/valueset/2.16.840.1.113762.1.4.1078.90/expansion/Latest 3. Fong, Kei Sen, and Mehul Motani. 2024. “Symbolic Regression Enhanced Decision Trees for Classification Tasks.” Proceedings of the AAAI Conference on Artificial Intelligence 38 (11): 12033–42.Publication MYC and p53 Alterations Cooperate through VEGF Signaling to Repress Cytotoxic T-cell and Immunotherapy Responses in Prostate Cancer(2025-11-14)Patients with castration-resistant prostate cancer (CRPC) are generally unresponsive to tumor-targeted treatments and immunotherapies. Genetic alterations acquired during the evolution of CRPC may affect antitumor immunity and immunotherapy responses, which could inform personalized therapeutic strategies. Using our innovative electroporation-based mouse models, we generated distinct genetic subtypes of CRPC found in patients and uncovered unique immune microenvironments. Specifically, mouse and human prostate tumors with MYC amplification and p53 disruption had weak cytotoxic lymphocyte infiltration and an overall dismal prognosis. MYC and p53 cooperated to induce tumor-intrinsic secretion of VEGF, which signaled through VEGFR2 expressed on CD8+ T cells to directly inhibit T-cell migration and effector functions. Targeting VEGF-VEGFR2 signaling in vivo remodeled the immunosuppressive prostate tumor microenvironment, leading to CD8+ T-cell-mediated primary tumor and metastasis growth suppression and significantly increased overall survival in MYC- and p53-altered CRPC. VEGFR2 blockade also led to the induction of PD-L1 in tumors and produced antitumor efficacy in combination with PD-L1 immune checkpoint blockade in multiple preclinical CRPC mouse models. Thus, these results identify a genetic mechanism of immunosuppression through VEGF signaling in prostate cancer that can be targeted to reactivate immune and immunotherapy responses in an aggressive subtype of CRPC. Significance: VEGFR2 blockade inhibits VEGF-mediated T-cell suppression and potentiates the effects of PD-L1 immune checkpoint blockade to treat castration-resistant prostate cancer driven by MYC and p53 alterations.Publication Travel nurses' experience with ethical challenges in practice: A qualitative descriptive study(2025-11-13)BackgroundNursing turnover rates are among the highest measured in recent years, contributing to financial and staffing challenges in the healthcare industry. Citing ethical challenges and subsequent moral distress, nurses have increasingly turned to travel nurse positions. Current literature regarding moral distress and ethical challenges has largely focused on staff nurses, and it is unknown how travel nurses experience ethical challenges and any associated moral distress.AimTo explore travel nurses' experiences with ethical challenges, inclusive of how they respond and feel or mitigate moral distress related to these challenges. Sources of support will be identified, with a focus on personal and organizational resources.Research designA qualitative descriptive study conducted through individual interviews. Participants were recruited with a purposive sampling strategy through flyers and social media. The data was analyzed using inductive and deductive content analysis. Participants and research context: Nurses ( = 15) working as travel nurses in the United States of America. Data was collected between August 2024 and February 2025.Ethical considerationsThis study received approval by the Institutional Review Board at the UMass Chan Medical School and Baystate Medical Center.FindingsThree themes described the participants' experience of ethical challenges. Strategies that travel nurses use to address and cope with ethical challenges include reflective thought, formation of a support network, and contemplation of action strategies.DiscussionTravel nurses' experience with ethical challenges has some unique qualities, and coping strategies have some distinctions from recommended strategies to avoid and mitigate moral distress.ConclusionThis study will inform practice models of travel nursing and strategies to support all nurses encountering ethical challenges in their practice. Nursing leaders should foster strategies for feedback from travel nurses, including ways to improve the ethical environment.Publication Myocardial ultrastructure in dogs and cats: review of normal structure, abnormal findings, and rationale for use in veterinary medicine(2025-11-12)Electron microscopy is an important imaging tool to identify ultrastructural abnormalities of the myocardium, diagnose certain diseases, and expand the understanding of mechanisms of cardiovascular disease processes. Transmission electron microscopy (TEM) is still an important part of human cardiovascular pathology, but its use in veterinary medicine has become uncommon. Even in patients with minimal histopathological changes on light microscopy, there can be important ultrastructural findings that help to achieve a diagnosis. In addition, TEM can be valuable for research by shedding light on underlying mechanisms or investigating new forms of cardiomyopathies. This review highlights the importance of TEM, details normal and pathological ultrastructural findings, and provides insight into clinical and research uses. We also provide tips on tissue preparation, processing, and analysis that are key to successful use of this valuable tool.Publication The Impact of Ethanol and Adolescent Social Isolation Stress on Nucleus Accumbens Medium Spiny Neurons Integration of Prefrontal Cortex and Basolateral Amygdala Inputs(UMass Chan Medical School, 2025-11-10)Drug addiction, including alcohol use disorder (AUD), hijacks the brain’s reward system and promotes drug-seeking behavior. Medium spiny neurons (MSNs) in the nucleus accumbens (NAc), a key reward center, receive glutamatergic input from the medial prefrontal cortex (mPFC) and basolateral amygdala (BLA), both of which are affected by alcohol. However, the cellular mechanisms by which the NAc integrates these inputs, and their relevance to AUD, remain poorly understood. Additionally, the effects of stress on glutamatergic signaling in the NAc are not well characterized. Using ex vivo electrophysiology and optogenetics, I found that under alcohol-naïve conditions, the NAc prioritizes mPFC and BLA inputs in an age- and sex-dependent manner. Adolescent males favor mPFC input more than females, but this difference disappears in adulthood, when females instead show greater BLA input than males. After two weeks of binge drinking, adolescent mice exhibit no changes in input prioritization. In contrast, adult males show a shift toward BLA input relative to controls. Social stress in alcohol-naïve adolescent mice enhances BLA input in both sexes compared to non-isolated females. Acute alcohol (50 mM) strengthens BLA input in non-isolated males and isolated females. These findings suggest that both stress and alcohol dynamically modulate BLA and mPFC input to the NAc in a sex-specific manner. This work enhances our understanding of how stress and alcohol interact to influence reward circuitry and highlights the importance of considering sex differences in addiction vulnerability and relapse risk.