Aaron J. Masino

Aaron J. Masino

Auto-generated profile, based on publicly available information

Aaron J. Masino

School of Computing, Clemson University

Research Focus & Metrics

Academic Career Length

13Years (2014-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
22 February 2017 in Journal of the American Medical Informatics Association
Verified
Journal Article
Closed access
PDF file
ACAnne Cocos
Anne Cocos
AFAlexander G Fiks
Alexander G Fiks
AMAaron J Masino
Aaron J Masino
Objective: Social media is an important pharmacovigilance data source for adverse drug reaction (ADR) identification. Human review of social media data is infeasible due to data quantity, thus natural language processing techniques are necessary. Social media includes informal vocabulary and irregular grammar, which challenge natural language processing methods. Our objective is to develop a scalable, deep-learning approach that exceeds state-of-the-art ADR detection performance in social media.Materials and Methods: We developed a recurrent neural network (RNN) model that labels words in an input sequence with ADR membership tags. The only input features are word-embedding vectors, which can be formed through task-independent pretraining or during ADR detection training.Results: Our best-performing RNN model used pretrained word embeddings created from a large, non–domain-specific Twitter dataset. It achieved an approximate match F-measure of 0.755 for ADR identification on the dataset, compared to 0.631 for a baseline lexicon system and 0.65 for the state-of-the-art conditional random field model. Feature analysis indicated that semantic information in pretrained word embeddings boosted sensitivity and, combined with contextual awareness captured in the RNN, precision.Discussion: Our model required no task-specific feature engineering, suggesting generalizability to additional sequence-labeling tasks. Learning curve analysis showed that our model reached optimal performance with fewer training examples than the other models.Conclusions: ADR detection performance in social media is significantly improved by using a contextually aware model and word embeddings formed from large, unlabeled datasets. The approach reduces manual data-labeling requirements and is scalable to large social media datasets.

22 February 2019 in PLOS ONE
Verified
Journal Article
Open access
PDF file
AMAaron J. Masino
Aaron J. Masino
MHMary Catherine Harris
Mary Catherine Harris
DFDaniel Forsyth
Daniel Forsyth
SOSvetlana Ostapenko
Svetlana Ostapenko
LSLakshmi Srinivasan
Lakshmi Srinivasan
RGRobert W Grundmeier
Robert W Grundmeier
Rapid antibiotic administration is known to improve sepsis outcomes, however early diagnosis remains challenging due to complex presentation. Our objective was to develop a model using readily available electronic health record (EHR) data capable of recognizing infant sepsis at least 4 hours prior to clinical recognition. We performed a retrospective case control study of infants hospitalized ≥48 hours in the Neonatal Intensive Care Unit (NICU) at the Children’s Hospital of Philadelphia between September 2014 and November 2017 who received at least one sepsis evaluation before 12 months of age. We considered two evaluation outcomes as cases: culture positive–positive blood culture for a known pathogen (110 evaluations); and clinically positive–negative cultures but antibiotics administered for ≥120 hours (265 evaluations). Case data was taken from the 44-hour window ending 4 hours prior to evaluation. We randomly sampled 1,100 44-hour windows of control data from all times ≥10 days removed from any evaluation. Model inputs consisted of up to 36 features derived from routine EHR data. Using 10-fold nested cross-validation, 8 machine learning models were trained to classify inputs as sepsis positive or negative. When tasked with discriminating culture positive cases from controls, 6 models achieved a mean area under the receiver operating characteristic (AUC) between 0.80–0.82 with no significant differences between them. Including both culture and clinically positive cases, the same 6 models achieved an AUC between 0.85–0.87, again with no significant differences. Machine learning models can identify infants with sepsis in the NICU hours prior to clinical recognition. Learning curves indicate model improvement may be achieved with additional training examples. Additional input features may also improve performance. Further research is warranted to assess potential performance improvements and clinical efficacy in a prospective trial.

1 March 2021 in PLOS ONE
Verified
Journal Article
Open access
PDF file
SBSaurav Bose
Saurav Bose
CKChén C. Kenyon
Chén C. Kenyon
AMAaron J. Masino
Aaron J. Masino
Early childhood asthma diagnosis is common; however, many children diagnosed before age 5 experience symptom resolution and it remains difficult to identify individuals whose symptoms will persist. Our objective was to develop machine learning models to identify which individuals diagnosed with asthma before age 5 continue to experience asthma-related visits. We curated a retrospective dataset for 9,934 children derived from electronic health record (EHR) data. We trained five machine learning models to differentiate individuals without subsequent asthma-related visits (transient diagnosis) from those with asthma-related visits between ages 5 and 10 (persistent diagnosis) given clinical information up to age 5 years. Based on average NPV-Specificity area (ANSA), all models performed significantly better than random chance, with XGBoost obtaining the best performance (0.43 mean ANSA). Feature importance analysis indicated age of last asthma diagnosis under 5 years, total number of asthma related visits, self-identified black race, allergic rhinitis, and eczema as important features. Although our models appear to perform well, a lack of prior models utilizing a large number of features to predict individual persistence makes direct comparison infeasible. However, feature importance analysis indicates our models are consistent with prior research indicating diagnosis age and prior health service utilization as important predictors of persistent asthma. We therefore find that machine learning models can predict which individuals will experience persistent asthma with good performance and may be useful to guide clinician and parental decisions regarding asthma counselling in early childhood.

21 July 2014 in BMC Bioinformatics
Verified
Journal Article
Open access
PDF file
AMAaron J Masino
Aaron J Masino
EDElizabeth T Dechene
Elizabeth T Dechene
MDMatthew C Dulik
Matthew C Dulik
AWAlisha Wilkens
Alisha Wilkens
NSNancy B Spinner
Nancy B Spinner
PWPeter S White
Peter S White
Exome sequencing is a promising method for diagnosing patients with a complex phenotype. However, variant interpretation relative to patient phenotype can be challenging in some scenarios, particularly clinical assessment of rare complex phenotypes. Each patient’s sequence reveals many possibly damaging variants that must be individually assessed to establish clear association with patient phenotype. To assist interpretation, we implemented an algorithm that ranks a given set of genes relative to patient phenotype. The algorithm orders genes by the semantic similarity computed between phenotypic descriptors associated with each gene and those describing the patient. Phenotypic descriptor terms are taken from the Human Phenotype Ontology (HPO) and semantic similarity is derived from each term’s information content. Model validation was performed via simulation and with clinical data. We simulated 33 Mendelian diseases with 100 patients per disease. We modeled clinical conditions by adding noise and imprecision, i.e. phenotypic terms unrelated to the disease and terms less specific than the actual disease terms. We ranked the causative gene against all 2488 HPO annotated genes. The median causative gene rank was 1 for the optimal and noise cases, 12 for the imprecision case, and 60 for the imprecision with noise case. Additionally, we examined a clinical cohort of subjects with hearing impairment. The disease gene median rank was 22. However, when also considering the patient’s exome data and filtering non-exomic and common variants, the median rank improved to 3. Semantic similarity can rank a causative gene highly within a gene list relative to patient phenotype characteristics, provided that imprecision is mitigated. The clinical case results suggest that phenotype rank combined with variant analysis provides significant improvement over the individual approaches. We expect that this combined prioritization approach may increase accuracy and decrease effort for clinical genetic diagnosis.

1 May 2017 in Journal of Biomedical Informatics
Verified
Journal Article
Closed access
ACAnne Cocos
Anne Cocos
TQTing Qian
Ting Qian
CCChris Callison-Burch
Chris Callison-Burch
AMAaron J. Masino
Aaron J. Masino

10 June 2025 in JMIR Cancer
Verified
Journal Article
Open access
ECEmma Coen
Emma Coen
GFGuilherme Del Fiol
Guilherme Del Fiol
KKKimberly A. Kaphingst
Kimberly A. Kaphingst
EBEmerson Borsato
Emerson Borsato
JSJackilen Shannon
Jackilen Shannon
CACaitlin Allen
Caitlin Allen
The increasing demand for population-wide genomic screening and the limited availability of genetic counseling resources have created a pressing need for innovative service delivery models. Chatbots powered by large language models (LLMs) have shown potential in genomic services, particularly in pretest counseling, but their application in returning positive population-wide genomic screening results remains underexplored. Leveraging advanced LLMs like GPT-4 offers an opportunity to address this gap by delivering accurate, contextual, and user-centered communication to individuals receiving positive genetic test results. This project aimed to design, implement, and evaluate a chatbot integrated with GPT-4, tailored to support the return of positive genomic screening results in the context of South Carolina’s In Our DNA SC program. This initiative offers free genetic screening to 100,000 individuals, with over 33,000 results returned and numerous positive findings for conditions such as Lynch syndrome, hereditary breast and ovarian cancer syndrome, and familial hypercholesterolemia. A 3-step prompt engineering process using retrieval-augmented generation and few-shot techniques was used to create the chatbot. Training materials included patient frequently asked questions, genetic counseling scripts, and patient-derived queries. The chatbot underwent iterative refinement based on 13 training questions, while performance was evaluated through expert ratings on responses to 2 hypothetical patient scenarios. The 2 scenarios were intended to represent common but distinct patient profiles in terms of gender, race, ethnicity, age, and background knowledge. Domain experts rated the chatbot using a 5-point Likert scale across 8 predefined criteria: tone, clarity, program accuracy, domain accuracy, robustness, efficiency, boundaries, and usability. The chatbot achieved an average score of 3.86 (SD 0.89) across all evaluation metrics. The highest-rated criteria were tone (mean 4.25, SD 0.71) and usability (mean 4.25, SD 0.58), reflecting the chatbot’s ability to communicate effectively and provide a seamless user experience. Boundary management (mean 4.0, SD 0.76) and efficiency (mean 3.88, SD 1.08) also scored well, while clarity and robustness received ratings of 3.81 (SD 1.05) and 3.81 (SD 0.66), respectively. Domain accuracy was rated 3.63 (SD 0.96), indicating satisfactory performance in delivering genetic information, whereas program accuracy received the lowest score of 3.25 (SD 1.39), highlighting the need for improvements in delivering program-specific details. This project demonstrates the feasibility of using LLM-powered chatbots to support the return of positive genomic screening results. The chatbot effectively handled open-ended patient queries, maintained conversational boundaries, and delivered user-friendly responses. However, enhancements in program-specific accuracy are essential to maximize its utility. Future research will explore hybrid chatbot designs that combine the strengths of LLMs with rule-based components to improve scalability, accuracy, and accessibility in genomic service delivery. The findings underscore the potential of generative artificial intelligence tools to address resource limitations and improve the accessibility of genomic health care services.