About

James D Thomas is a Professor of Medicine-Cardiology at Northwestern University, affiliated with the Feinberg School of Medicine. He holds an MD from Harvard Medical School, obtained in 1981, and has completed residency at Massachusetts General Hospital in Cardiology, as well as fellowships at the University of Vermont College of Medicine and Massachusetts General Hospital. His research focuses on cardiovascular disease, with notable contributions in echocardiography, including the comparison of global longitudinal strain measurements among different vendors, and the development of automated software for echocardiographic analysis. His work also encompasses the assessment of paravalvular regurgitation after valve replacement and the application of artificial intelligence in cardiology. Dr. Thomas has authored multiple publications in leading journals, advancing the understanding and evaluation of cardiac function and imaging techniques.

Research topics

• Artificial Intelligence
• Medicine
• Computer Science
• Cardiology
• Machine Learning
• Internal medicine
• Radiology
• Algorithm
• Pathology
• Data science
• Surgery
• Nursing

Selected publications

• Impact of measurement location on direct mitral regurgitation quantification using four-dimensional flow cardiovascular magnetic resonance

  Journal of Cardiovascular Magnetic Resonance · 2025-01-01 · 2 citations

  articleOpen access

  Four-dimensional (4D) flow cardiovascular magnetic resonance (CMR) shows promise for quantifying mitral regurgitation (MR) by allowing for direct regurgitant volume (RVol) measurement using a plane precisely placed at the MR jet. However, the ideal location of a measurement plane remains unclear. This study aims to systematically examine how varying measurement locations affect RVol quantification and determine the optimal location using the momentum conservation principle of a free jet. Patients diagnosed with MR by transthoracic echocardiography (TTE) and scheduled for CMR were prospectively recruited. Regurgitant jet flow volume (RVol jet ) and regurgitant jet flow momentum (RMom jet ) were quantified using 4D flow CMR at seven locations along the jet axis, x. The reference plane (mid-plane, x = 0 mm) was positioned at the peak velocity of the jet at each cardiac phase, and three additional planes were positioned on either side of the jet, each 2.5 mm apart. RVol jet was compared to RVol TTE , measured by the proximal isovelocity surface area method, and RVol indirect , measured by subtracting aortic forward flow volume from the left ventricle stroke volume derived from two-dimensional phase contrast at the aortic valve and a stack of short-axis cine CMR techniques. RVol jet and RMom jet were quantified in 45 patients (age 63±13, male 26). In patients with RVol jet at x = 0 mm ≥ 10 mL (n = 25), RVol jet consistently increased as the plane moved downstream. RVol jet measured furthest upstream (x = −7.5 mm) was significantly lower (39±11%, p<0.001) and RVol jet measured furthest downstream (x = 7.5 mm) was significantly higher (16±19%, p<0.001) than RVol jet at x = 0 mm. RMom jet similarly increased from x = −7.5 to 0 mm (57±12%, p<0.001) but stabilized from x = 0–7.5 mm (−2±17%). From x = −7.5 to 7.5 mm, RVol jet was in consistent moderate agreement with RVol indirect (n = 41, bias = −2±24 to 8±32 mL, intraclass correlation coefficient = 0.55–0.63, p<0.001). The location of a measurement plane significantly influences RVol quantification using the direct 4D flow CMR approach. Based on the converging profile of RMom jet , we propose the peak velocity of the jet as the optimal position.

  PublisherDOI

• FibriCheck Detection Capabilities for Atrial Fibrillation (FDA – AF): A Multicenter Validation Study

  Research Square · 2025-06-25

  preprintOpen accessSenior author
  PublisherOA PDFDOI

• Echocardiographic Parameters of the Right Ventricle in Patients With Pulmonary Hypertension: A Review

  Korean Circulation Journal · 2025-01-01 · 8 citations

  reviewOpen accessSenior author

  To diagnose pulmonary hypertension (PH) and assess its severity, accurate measurement of pulmonary artery (PA) pressure is crucial. However, there can be significant discrepancies between echocardiography (Echo) and invasive catheterization. The right ventricle (RV) has a complex structure, and its remodeling in PH is diverse, making it challenging to evaluate RV physiology with a single imaging modality. While right heart catheterization is the gold standard, its practicality in clinical settings is limited. Cardiac magnetic resonance imaging (MRI) is valuable for RV evaluation, with 4-dimensional flow MRI showing promise, yet accessibility remains a concern. Thus, in PH patient management, Echo plays a central role as a practical decision-making tool. This review aims to elucidate Echo parameters in PH patients, highlighting differences in PA systolic pressure measurements, RV-PA coupling, RV remodeling patterns crucial for understanding PH progression, and clinical evidence regarding RV strain. Additionally, it aims to introduce new Echo parameters that help understand RV in PH.

  PublisherOA PDFDOI

• Development and evaluation of prompts for a large language model to screen titles and abstracts in a living systematic review

  BMJ Mental Health · 2025-07-01 · 6 citations

  reviewOpen access

  BACKGROUND: Living systematic reviews (LSRs) maintain an updated summary of evidence by incorporating newly published research. While they improve review currency, repeated screening and selection of new references make them labourious and difficult to maintain. Large language models (LLMs) show promise in assisting with screening and data extraction, but more work is needed to achieve the high accuracy required for evidence that informs clinical and policy decisions. OBJECTIVE: The study evaluated the effectiveness of an LLM (GPT-4o) in title and abstract screening compared with human reviewers. METHODS: Human decisions from an LSR on prodopaminergic interventions for anhedonia served as the reference standard. The baseline search results were divided into a development and a test set. Prompts guiding the LLM's eligibility assessments were refined using the development set and evaluated on the test set and two subsequent LSR updates. Consistency of the LLM outputs was also assessed. RESULTS: Prompt development required 1045 records. When applied to the remaining baseline 11 939 records and two updates, the refined prompts achieved 100% sensitivity for studies ultimately included in the review after full-text screening, though sensitivity for records included by humans at the title and abstract stage varied (58-100%) across updates. Simulated workload reductions of 65-85% were observed. Prompt decisions showed high consistency, with minimal false exclusions, satisfying established screening performance benchmarks for systematic reviews. CONCLUSIONS: Refined GPT-4o prompts demonstrated high sensitivity and moderate specificity while reducing human workload. This approach shows potential for integrating LLMs into systematic review workflows to enhance efficiency.

  PublisherOA PDFDOI

• Age and gender influence left ventricular diastolic filling flow assessed by 4D flow MRI

  Journal of Cardiovascular Magnetic Resonance · 2025-01-01

  articleOpen access
  PublisherDOI

• Correction: p38γ MAPK Is Essential for Aerobic Glycolysis and Pancreatic Tumorigenesis

  Cancer Research · 2025-02-17 · 1 citations

  erratumOpen access
  PublisherOA PDFDOI

• Opportunities, challenges and risks of using artificial intelligence for evidence synthesis

  BMJ evidence-based medicine · 2025-01-09 · 23 citations

  article
  PublisherDOI

• Vendor differences in 2D-speckle tracking global longitudinal strain: an update on a 10-year standardization effort

  European Heart Journal - Cardiovascular Imaging · 2025-05-26 · 17 citations

  article

  AIMS: To assess the inter-vendor differences in global longitudinal strain measurements and determine a potential improvement compared with the situation 10 years ago. METHODS AND RESULTS: 372 echocardiographic exams were performed in 62 subjects (50 male, age 56 ± 17) with LV ejection fraction ranging from 30% to 68%, using ultrasound systems from six manufacturers: GE, Philips, Canon, Siemens, Fujifilm and Esaote. Each subject was scanned consecutively on all machines by the same assigned sonographer, with two image sets per subject to assess test-retest setting reproducibility. Average peak systolic global strain from the three apical views (GLSAV) was measured on three vendor-specific (Canon, Siemens, and Fujifilm) and six vendor-agnostic (GE, Philips, US2.AI, Caas Qardia, Medis, and Epsilon) software solutions (SWS). Endocardial and mid-/full-wall GLS were measured and compared with the mean GLS of contemporary semi-automated clinical software: GE, Philips, Canon, Fujifilm, and Caas Qardia. Endocardial and mid-/full-wall GLS measurements from contemporary semi-automated clinical software showed minimal inter-vendor differences, with an average maximum bias of 0.6% strain units. There was a remaining inter-vendor bias with and among some other vendors. The average minimal detectable change with contemporary semi-automated clinical software was 2.5 and 2.4 strain% for endocardial and mid-/full-wall GLS, resp. These values were higher for and among some other vendors. Test-retest variability of GLS measurements was good and similar to that of LV ejection fraction (6.6% vs. 6.5%, P > 0.05), indicating consistent results across repeated scans. CONCLUSION: In this controlled study setting, GLS measurements from companies that provide contemporary semi-automated clinical software have become more consistent, compared with 10 years ago. Mid-/full-wall strain was now available in all but one software.

  PublisherDOI

• Impact of Preoperative Left Ventricular Dysfunction on Residual and Recurrent Mitral Regurgitation After Degenerative Mitral Valve Repair

  The Annals of Thoracic Surgery · 2025-11-22

  article
  PublisherDOI

• The Edwardians and the Making of a Modern Spanish Obsession

  Hispanic Research Journal · 2025-07-04

  article1st authorCorresponding
  PublisherDOI

Recent grants

• NIH Grant M01RR018390

  NIH · $519k · 2007

• NIH Grant R01HL056688

  NIH · $546k · 2001

Frequent coauthors

• Zoran Popović

  Cleveland Clinic

  285 shared

• Neil Greenberg

  King's College London

  275 shared

• Mario J. García

  Albert Einstein College of Medicine

  227 shared

• Arthur E. Weyman

  Massachusetts General Hospital

  216 shared

• Benjamin D. Levine

  The University of Texas Southwestern Medical Center

  198 shared

• Patrick M. McCarthy

  196 shared

• Anand Prasad

  The University of Texas Health Science Center at San Antonio

  177 shared

• Takahiro Shiota

  Cedars-Sinai Medical Center

  161 shared