Mendel Lebowitz
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Mendel Lebowitz has a total of 50 co-authors. Here are the top co-authors by number of shared publications:
- Anant Madabhushi
Department of Biomedical Engineering, Emory University, Atlanta, GA
4 shared publications - Gourav Modanwal
Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology and Emory University , Health Sciences Research Building II, 1750 Haygood Drive, Suite N647, Atlanta, GA 30322
4 shared publications Anant Madabhushi
Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology and Emory University , Health Sciences Research Building II, 1750 Haygood Drive, Suite N647, Atlanta, GA 30322
3 shared publications- Rohan Dhamdhere
Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology and Emory University , Health Sciences Research Building II, 1750 Haygood Drive, Suite N647, Atlanta, GA 30322
3 shared publications - Santosh Kumar Sirasapalli
Harrington Heart and Vascular Institute, University Hospitals/Case Western Reserve University , Cleveland, OH
3 shared publications - Amritpal Singh
Emory University Computer Science and Informatics Atlanta, GA, USA
2 shared publications
Latest Publications
- Open Access
- 13 November 2026
- Published by Oxford University Press (OUP) in European Heart Journal - Cardiovascular Imaging
Liver radiomics on routine cardiac CT identifies hepatic stress states that predict heart failure
Aims: Heart failure (HF) prevention remains limited because early multisystem cardiometabolic dysfunction is not captured by conventional risk tools. We evaluated whether liver radiomic phenotypes derived opportunistically from coronary artery calcium computed tomography (CAC-CT) identify imaging-defined hepatic stress states associated with incident HF. Methods and results: We analysed 59 237 individuals undergoing CAC-CT (CLARIFY: 46 028; Houston Methodist: 13 209). Liver regions were segmented and radiomic features extracted to derive a Liver Radiomics Score (LRS). The model was trained in a CLARIFY derivation cohort (Dtr, n = 9205), validated in a hold-out cohort (Dts, n = 36 823), and externally evaluated in Houston Methodist (Dext, n = 13 209). The mean follow-up was 1411 days (SD 760.1) in the pooled Dtr and Dts cohorts (1144 HF events) and 470 days (SD 596) in Dext (212 HF events). Continuous LRS was associated with incident HF [Dts: hazard ratio (HR) 1.63, 95% confidence interval (CI) 1.52–1.74; P < 0.001; Dext: HR 1.04, 95% CI 1.01–1.07; P = 0.004] and remained significant after multivariable adjustment in Dts. Increasing LRS quartiles showed progressively adverse hepatic, haemodynamic, and cardiometabolic profiles and higher HF risk. Among 17 806 Dts participants with PREVENT data, adding LRS modestly improved discrimination across CAC- and PREVENT-based models (P < 0.001) and improved reclassification metrics (7-year net reclassification improvement up to 0.25; including integrated discrimination improvement up to 0.015). High LRS identified elevated risk in low-risk groups: among participants with CAC=0, high LRS conferred ∼2.5-fold higher 5-year HF incidence, and ∼2.9-fold in PREVENT-low individuals. Conclusion: In this large multicentre cohort, liver radiomics from CAC-CT captured hepatic stress states linked to HF risk and modestly improved risk stratification beyond CAC and PREVENT scores.
- 13 November 2026
- Published by Oxford University Press (OUP) in European Heart Journal - Cardiovascular Imaging
Open AccessLiver radiomics on routine cardiac CT identifies hepatic stress states that predict heart failure
Aims: Heart failure (HF) prevention remains limited because early multisystem cardiometabolic dysfunction is not captured by conventional risk tools. We evaluated whether liver radiomic phenotypes derived opportunistically from coronary artery calcium computed tomography (CAC-CT) identify imaging-defined hepatic stress states associated with incident HF. Methods and results: We analysed 59 237 individuals undergoing CAC-CT (CLARIFY: 46 028; Houston Methodist: 13 209). Liver regions were segmented and radiomic features extracted to derive a Liver Radiomics Score (LRS). The model was trained in a CLARIFY derivation cohort (Dtr, n = 9205), validated in a hold-out cohort (Dts, n = 36 823), and externally evaluated in Houston Methodist (Dext, n = 13 209). The mean follow-up was 1411 days (SD 760.1) in the pooled Dtr and Dts cohorts (1144 HF events) and 470 days (SD 596) in Dext (212 HF events). Continuous LRS was associated with incident HF [Dts: hazard ratio (HR) 1.63, 95% confidence interval (CI) 1.52–1.74; P < 0.001; Dext: HR 1.04, 95% CI 1.01–1.07; P = 0.004] and remained significant after multivariable adjustment in Dts. Increasing LRS quartiles showed progressively adverse hepatic, haemodynamic, and cardiometabolic profiles and higher HF risk. Among 17 806 Dts participants with PREVENT data, adding LRS modestly improved discrimination across CAC- and PREVENT-based models (P < 0.001) and improved reclassification metrics (7-year net reclassification improvement up to 0.25; including integrated discrimination improvement up to 0.015). High LRS identified elevated risk in low-risk groups: among participants with CAC=0, high LRS conferred ∼2.5-fold higher 5-year HF incidence, and ∼2.9-fold in PREVENT-low individuals. Conclusion: In this large multicentre cohort, liver radiomics from CAC-CT captured hepatic stress states linked to HF risk and modestly improved risk stratification beyond CAC and PREVENT scores.
- Open Access
- 10 November 2026
- Published by Elsevier in Journal of Cardiac Failure
- Vol. 32 (11) , 2034-2039
AI-Enabled Opportunistic Splenic Phenotyping on Calcium-Scoring CT Identifies Residual Heart Failure Risk in Patients Without Coronary Calcium
- 10 November 2026
- Published by Elsevier in Journal of Cardiac Failure
- Vol. 32 (11) , 2034-2039
Open AccessAI-Enabled Opportunistic Splenic Phenotyping on Calcium-Scoring CT Identifies Residual Heart Failure Risk in Patients Without Coronary Calcium
- Open Access
- 14 July 2026
- Published by Elsevier in EBioMedicine
- Vol. 130, 106385
CardioMorph Atlas: a statistical approach to evaluate association between pulmonary tuberculosis and cardiac morphology from chest X-rays
- 14 July 2026
- Published by Elsevier in EBioMedicine
- Vol. 130, 106385
Open AccessCardioMorph Atlas: a statistical approach to evaluate association between pulmonary tuberculosis and cardiac morphology from chest X-rays
- Open Access
- 9 July 2026
- Published by Oxford University Press (OUP) in European Heart Journal
Peri-aortic fat to assess cardiovascular aging using an AI-driven radiomic biomarker
Background and Aims: Chronological age is central to cardiovascular disease (CVD) risk estimation but poorly reflects interindividual heterogeneity in cardiovascular aging. This study developed a peri-aortic adipose tissue (PAAT) radiomic cardiovascular age (CV-Age) biomarker from routine chest CT and tested whether the PAAT age gap (ΔAge = CV-Age − chronological age) stratifies major adverse cardiovascular events (MACE) and improves risk classification. Methods: Four chest CT cohorts with different protocols were utilized to train (n = 4451) and externally validate (n = 44 214) a CV-Age model. Associations between ΔAge decile groups (resilient ≤10th; accelerated ≥90th) and incident 5-point MACE (myocardial infarction, stroke, revascularization, heart failure, and death) were assessed using survival models. Clinical utility was evaluated by substituting CV-Age for chronological age in PREVENT to quantify risk reclassification. Results: A 31-feature radiomic signature capturing PAAT volume, attenuation, and texture heterogeneity predicted age with good accuracy (training MAE 2.2 years; primary external validation MAE 2.7 years). In a separate higher-risk cohort, predictions showed a positive ΔAge (median +5.4 years), reflecting elevated baseline cardiovascular burden. ΔAge-stratified event-free survival differed across groups (log-rank P < .001); compared with normative agers, accelerated agers had higher MACE risk (HR 1.51, 95% CI 1.37–1.66), and resilient agers had lower risk (HR 0.70, 0.63–0.77). Substituting CV-Age into PREVENT improved risk reclassification (categorical NRI +0.03, P < .01, continuous NRI +0.04, P < .05). Conclusions: Radiomic profiling of PAAT on routine chest CT yields a scalable imaging biomarker of cardiovascular aging. CV-Age can help improve identification of higher-risk individuals missed by age-driven risk estimation, supporting its role as a risk-enrichment tool.
- 9 July 2026
- Published by Oxford University Press (OUP) in European Heart Journal
Open AccessPeri-aortic fat to assess cardiovascular aging using an AI-driven radiomic biomarker
Background and Aims: Chronological age is central to cardiovascular disease (CVD) risk estimation but poorly reflects interindividual heterogeneity in cardiovascular aging. This study developed a peri-aortic adipose tissue (PAAT) radiomic cardiovascular age (CV-Age) biomarker from routine chest CT and tested whether the PAAT age gap (ΔAge = CV-Age − chronological age) stratifies major adverse cardiovascular events (MACE) and improves risk classification. Methods: Four chest CT cohorts with different protocols were utilized to train (n = 4451) and externally validate (n = 44 214) a CV-Age model. Associations between ΔAge decile groups (resilient ≤10th; accelerated ≥90th) and incident 5-point MACE (myocardial infarction, stroke, revascularization, heart failure, and death) were assessed using survival models. Clinical utility was evaluated by substituting CV-Age for chronological age in PREVENT to quantify risk reclassification. Results: A 31-feature radiomic signature capturing PAAT volume, attenuation, and texture heterogeneity predicted age with good accuracy (training MAE 2.2 years; primary external validation MAE 2.7 years). In a separate higher-risk cohort, predictions showed a positive ΔAge (median +5.4 years), reflecting elevated baseline cardiovascular burden. ΔAge-stratified event-free survival differed across groups (log-rank P < .001); compared with normative agers, accelerated agers had higher MACE risk (HR 1.51, 95% CI 1.37–1.66), and resilient agers had lower risk (HR 0.70, 0.63–0.77). Substituting CV-Age into PREVENT improved risk reclassification (categorical NRI +0.03, P < .01, continuous NRI +0.04, P < .05). Conclusions: Radiomic profiling of PAAT on routine chest CT yields a scalable imaging biomarker of cardiovascular aging. CV-Age can help improve identification of higher-risk individuals missed by age-driven risk estimation, supporting its role as a risk-enrichment tool.
- 23 January 2026
- Published by SAGE Publications in Diabetes Technology & Therapeutics
- Vol. 28 (8) , 782-791
Diagnostic Accuracy of Artificial Intelligence-Enabled Retinal Biomarkers for Detecting Type 2 Diabetes and Prediabetes Among Asian Indians [DART Study]
Purpose: To identify retinal microvascular features that distinguish normoglycemia from diabetes mellitus (DM)[Study-1] and prediabetes (PreDM)[Study-2] among Asian Indians using artificial intelligence (AI); evaluate their diagnostic accuracy; and examine independent associations of oculomics scores [Oculomic Diabetes Score (ODiS) and Prediabetes Score (OPreS)] with glycemic status. Methods: We analyzed 273 retinal images from 139 participants (19 = normoglycemia, 100 = DM, 20 = PreDM; mean age: 49.4 ± 12.9 yrs; male: 71.2%) with dataset randomly split (50:50) into training and test sets. We extracted 226 quantitative vessel tortuosity features separately for arteries and veins using machine vision-based approaches. Top six discriminating features were separately selected (using Wilcoxon Rank-Sum test and Linear Discriminant Analysis) on training sets and validated on independent blinded test sets. Model performances were evaluated using area under precision-recall curve (AUPRC) for DM and area under the receiver operating characteristic curve (AUC) for PreDM. Independent association of ODiS and OPreS (adjusting for age, sex, height, weight, blood pressure, serum cholesterol, serum creatinine) was assessed by multivariable logistic regression. Results: Specific oculomics-based retinal vascular features distinguished DM/PreDM from normoglycemia. Vein-only model achieved AUPRC = 0.96 (95% confidence interval [95% CI]: 0.9–1.00) (sensitivity = 95%, specificity = 72.2%, precision = 95%) on S v1 for DM and an AUC = 0.80 (95% CI: 0.63–0.94) (sensitivity = 70%, specificity = 80%, precision = 82.35%) on S v2 for PreDM. Oculomics scores were independently associated with DM(ODiS) [Adjusted Odds ratio (AdjOR): 2.00 (95% CI: 1.56–2.57, P < 0.001)] but not with PreDM(OPreS) [AdjOR: 1.32 (95% CI: 0.65–2.71, P = 0.48)]. Conclusions: In this proof-of-concept study, AI-informed retinal venous features on routine fundus images, with further prospective and multisite validation, could potentially serve as noninvasive DM detection using AI models among Asian Indians.
- 23 January 2026
- Published by SAGE Publications in Diabetes Technology & Therapeutics
- Vol. 28 (8) , 782-791
Diagnostic Accuracy of Artificial Intelligence-Enabled Retinal Biomarkers for Detecting Type 2 Diabetes and Prediabetes Among Asian Indians [DART Study]
Purpose: To identify retinal microvascular features that distinguish normoglycemia from diabetes mellitus (DM)[Study-1] and prediabetes (PreDM)[Study-2] among Asian Indians using artificial intelligence (AI); evaluate their diagnostic accuracy; and examine independent associations of oculomics scores [Oculomic Diabetes Score (ODiS) and Prediabetes Score (OPreS)] with glycemic status. Methods: We analyzed 273 retinal images from 139 participants (19 = normoglycemia, 100 = DM, 20 = PreDM; mean age: 49.4 ± 12.9 yrs; male: 71.2%) with dataset randomly split (50:50) into training and test sets. We extracted 226 quantitative vessel tortuosity features separately for arteries and veins using machine vision-based approaches. Top six discriminating features were separately selected (using Wilcoxon Rank-Sum test and Linear Discriminant Analysis) on training sets and validated on independent blinded test sets. Model performances were evaluated using area under precision-recall curve (AUPRC) for DM and area under the receiver operating characteristic curve (AUC) for PreDM. Independent association of ODiS and OPreS (adjusting for age, sex, height, weight, blood pressure, serum cholesterol, serum creatinine) was assessed by multivariable logistic regression. Results: Specific oculomics-based retinal vascular features distinguished DM/PreDM from normoglycemia. Vein-only model achieved AUPRC = 0.96 (95% confidence interval [95% CI]: 0.9–1.00) (sensitivity = 95%, specificity = 72.2%, precision = 95%) on S v1 for DM and an AUC = 0.80 (95% CI: 0.63–0.94) (sensitivity = 70%, specificity = 80%, precision = 82.35%) on S v2 for PreDM. Oculomics scores were independently associated with DM(ODiS) [Adjusted Odds ratio (AdjOR): 2.00 (95% CI: 1.56–2.57, P < 0.001)] but not with PreDM(OPreS) [AdjOR: 1.32 (95% CI: 0.65–2.71, P = 0.48)]. Conclusions: In this proof-of-concept study, AI-informed retinal venous features on routine fundus images, with further prospective and multisite validation, could potentially serve as noninvasive DM detection using AI models among Asian Indians.
- 23 October 2025
- Published by Oxford University Press (OUP) in European Heart Journal
When bias correction becomes bias: lessons for cardiovascular age models
- 23 October 2025
- Published by Oxford University Press (OUP) in European Heart Journal
When bias correction becomes bias: lessons for cardiovascular age models
- 20 June 2024
- Published by Elsevier in Hormones and Behavior
- Vol. 164, 105579
Infants' salivary oxytocin and positive affective reactions to people
- 20 June 2024
- Published by Elsevier in Hormones and Behavior
- Vol. 164, 105579
Infants' salivary oxytocin and positive affective reactions to people
- 2 October 2023
- Published by Taylor & Francis in Parenting
- Vol. 23 (3-4) , 171-196
Maternal Stress: The First 14 Months Postpartum
- 2 October 2023
- Published by Taylor & Francis in Parenting
- Vol. 23 (3-4) , 171-196
Maternal Stress: The First 14 Months Postpartum