Fatemeh Siavoshi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Fatemeh Siavoshi has a total of 43 co-authors. Here are the top co-authors by number of shared publications:
- Pavan Bhargava
Department of Neurology Johns Hopkins University School of Medicine Baltimore Maryland USA
6 shared publications - Kathryn C. Fitzgerald
Department of Neurology Johns Hopkins University School of Medicine Baltimore Maryland USA
5 shared publications - Dimitrios Christos Ladakis
Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, USA
3 shared publications - Angeliki G. Filippatou
Department of Neurology Johns Hopkins University School of Medicine Baltimore Maryland USA
2 shared publications - Blake E. Dewey
Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD, USA
2 shared publications - Elias S. Sotirchos
From the Department of Neurology, Johns Hopkins University School of Medicine, Baltimore, MD
2 shared publications
Latest Publications
- 1 November 2026
- Published by Wolters Kluwer Health in Neurology Neuroimmunology & Neuroinflammation
- Vol. 13 (6) , e200664
Metabolomic Profiles Associated With Disease Severity in Multiple Sclerosis
Sensitive, minimally invasive biomarkers that reflect functional decline in multiple sclerosis (MS) are still lacking. Untargeted metabolomic profiling allows unbiased discovery of circulating biomarkers and dysregulated biological pathways underlying MS onset and progression. We aimed to identify candidate circulating metabolite signatures correlated with clinical, imaging, and patient-reported outcomes in MS and to characterize the underlying metabolic pathways. We conducted a cross-sectional analysis of untargeted circulating metabolomic profiles from 541 samples of 323 people with multiple sclerosis (PwMS), including both relapsing-remitting and progressive forms, detecting 332 metabolites. Associations between metabolite levels and outcomes, including clinical disability (Expanded Disability Status Scale), functional tests (walking speed, manual dexterity, and processing speed), Neuro-QoL patient-reported domains, optical coherence tomography-derived retinal layer thicknesses, and MRI-derived measures, were assessed using linear mixed-effects models adjusted for age, sex, race, treatment status, and MS subtype. To identify clusters of metabolites jointly linked to outcomes, we applied Hierarchical All-Against-All Association Analysis and weighted correlation network analysis. Pathway enrichment was also performed using Fast Gene Set Enrichment Analysis on ranked effect sizes based on the RaMP database, excluding pathways with fewer than 3 metabolites. We identified multiple metabolites consistently associated with MS severity across clinical, patient-reported, and imaging domains. Metabolites such as p-cresol sulfate, p-cresol glucuronide, phenylacetylglutamine, and carnitine-derived species were linked to greater disease severity, whereas tryptophan betaine, ergothioneine, N-acetylcarnosine, and indolepropionate were associated with more favorable outcomes. Pathway analyses highlighted amino acid transport and metabolism, aromatic and branched-chain amino acid pathways, and solute carrier transporter systems, as well as bile acid, sphingolipid, and carnitine-related pathways linked to MS severity. Comprehensive integration of metabolomic, clinical, and imaging data in PwMS highlights distinct metabolites and metabolic pathways associated with MS severity. These candidate metabolites warrant external validation in independent cohorts and provide a rationale for future biomarker discovery studies aimed at improving disease monitoring and management in MS.
- 1 November 2026
- Published by Wolters Kluwer Health in Neurology Neuroimmunology & Neuroinflammation
- Vol. 13 (6) , e200664
Metabolomic Profiles Associated With Disease Severity in Multiple Sclerosis
Sensitive, minimally invasive biomarkers that reflect functional decline in multiple sclerosis (MS) are still lacking. Untargeted metabolomic profiling allows unbiased discovery of circulating biomarkers and dysregulated biological pathways underlying MS onset and progression. We aimed to identify candidate circulating metabolite signatures correlated with clinical, imaging, and patient-reported outcomes in MS and to characterize the underlying metabolic pathways. We conducted a cross-sectional analysis of untargeted circulating metabolomic profiles from 541 samples of 323 people with multiple sclerosis (PwMS), including both relapsing-remitting and progressive forms, detecting 332 metabolites. Associations between metabolite levels and outcomes, including clinical disability (Expanded Disability Status Scale), functional tests (walking speed, manual dexterity, and processing speed), Neuro-QoL patient-reported domains, optical coherence tomography-derived retinal layer thicknesses, and MRI-derived measures, were assessed using linear mixed-effects models adjusted for age, sex, race, treatment status, and MS subtype. To identify clusters of metabolites jointly linked to outcomes, we applied Hierarchical All-Against-All Association Analysis and weighted correlation network analysis. Pathway enrichment was also performed using Fast Gene Set Enrichment Analysis on ranked effect sizes based on the RaMP database, excluding pathways with fewer than 3 metabolites. We identified multiple metabolites consistently associated with MS severity across clinical, patient-reported, and imaging domains. Metabolites such as p-cresol sulfate, p-cresol glucuronide, phenylacetylglutamine, and carnitine-derived species were linked to greater disease severity, whereas tryptophan betaine, ergothioneine, N-acetylcarnosine, and indolepropionate were associated with more favorable outcomes. Pathway analyses highlighted amino acid transport and metabolism, aromatic and branched-chain amino acid pathways, and solute carrier transporter systems, as well as bile acid, sphingolipid, and carnitine-related pathways linked to MS severity. Comprehensive integration of metabolomic, clinical, and imaging data in PwMS highlights distinct metabolites and metabolic pathways associated with MS severity. These candidate metabolites warrant external validation in independent cohorts and provide a rationale for future biomarker discovery studies aimed at improving disease monitoring and management in MS.
- Open Access
Proteomic Age Acceleration in Multiple Sclerosis Precedes Symptom Onset and Associates with Severity
Biological aging is accelerated in people with multiple sclerosis, but whether such acceleration occurs during the pre-symptomatic phase or varies by organ system is understudied. We analyzed two independent proteomics datasets profiled using distinct platforms: the Johns Hopkins cohort profiled using the SomaScan platform (348 multiple sclerosis/49 age-matched controls) and the Department of Defense cohort profiled using the Olink platform (134 multiple sclerosis/79 age-matched controls), including 117 pre-symptomatic samples from people with multiple sclerosis (median lead time: 4.0 years), to estimate systemic and organ-specific proteomic age gaps using established clocks in pre-symptomatic and symptomatic phases, and assess their associations with severity. In the Johns Hopkins cohort, people with multiple sclerosis demonstrated acceleration of systemic (β=2.2, 95% CI 1.2–3.2, P P =0.003, FDR=0.01), muscle (β=2.5, 95% CI 1.3–3.7, P P =0.003, FDR=0.01), with findings reproduced in the Department of Defense cohort for systemic (β=0.7, 95% CI 0.0–1.4, P =0.04, FDR=0.34) and brain age (3.2 years, 95% CI 2.1–4.3, P P =0.002, FDR=0.02); brain: (β=2.4, 95% CI 1.2–3.7, P P =0.005, FDR=0.03) and slower walking speed (β=0.02, 95% CI 0.01–0.03, P =0.006, FDR=0.04), while higher muscle age gap was associated with greater global Age-Related Multiple Sclerosis Severity Score (β=0.17, 95% CI 0.10–0.24, P P =0.03, FDR=0.30), slower walking speed (β=0.02, 95% CI 0.01–0.03, P =0.002, FDR=0.02), lower peripapillary retinal nerve fiber layer (β= −0.26, 95% CI −0.41 to −0.10, P =0.001, FDR=0.02) and ganglion cell-inner plexiform layer thicknesses (β= −0.35; 95% CI −0.65 to −0.05; P =0.02, FDR=0.30). Higher brain age gap was associated with several imaging measures, including lower whole-brain (β= −0.002, 95% CI −0.003 to −0.001, P =0.002, FDR=0.02), and lower peripapillary retinal nerve fiber layer thickness (β= −0.21, 95% CI −0.39 to −0.03, P =0.02, FDR=0.10). Proteomic age acceleration in multiple sclerosis is detectable years before symptom onset and distinct organ-specific aging signatures are associated with disease severity. Proteomic aging may provide a biologically informative marker of early disease processes and a clinically relevant readout of disease heterogeneity.
Open AccessProteomic Age Acceleration in Multiple Sclerosis Precedes Symptom Onset and Associates with Severity
Biological aging is accelerated in people with multiple sclerosis, but whether such acceleration occurs during the pre-symptomatic phase or varies by organ system is understudied. We analyzed two independent proteomics datasets profiled using distinct platforms: the Johns Hopkins cohort profiled using the SomaScan platform (348 multiple sclerosis/49 age-matched controls) and the Department of Defense cohort profiled using the Olink platform (134 multiple sclerosis/79 age-matched controls), including 117 pre-symptomatic samples from people with multiple sclerosis (median lead time: 4.0 years), to estimate systemic and organ-specific proteomic age gaps using established clocks in pre-symptomatic and symptomatic phases, and assess their associations with severity. In the Johns Hopkins cohort, people with multiple sclerosis demonstrated acceleration of systemic (β=2.2, 95% CI 1.2–3.2, P P =0.003, FDR=0.01), muscle (β=2.5, 95% CI 1.3–3.7, P P =0.003, FDR=0.01), with findings reproduced in the Department of Defense cohort for systemic (β=0.7, 95% CI 0.0–1.4, P =0.04, FDR=0.34) and brain age (3.2 years, 95% CI 2.1–4.3, P P =0.002, FDR=0.02); brain: (β=2.4, 95% CI 1.2–3.7, P P =0.005, FDR=0.03) and slower walking speed (β=0.02, 95% CI 0.01–0.03, P =0.006, FDR=0.04), while higher muscle age gap was associated with greater global Age-Related Multiple Sclerosis Severity Score (β=0.17, 95% CI 0.10–0.24, P P =0.03, FDR=0.30), slower walking speed (β=0.02, 95% CI 0.01–0.03, P =0.002, FDR=0.02), lower peripapillary retinal nerve fiber layer (β= −0.26, 95% CI −0.41 to −0.10, P =0.001, FDR=0.02) and ganglion cell-inner plexiform layer thicknesses (β= −0.35; 95% CI −0.65 to −0.05; P =0.02, FDR=0.30). Higher brain age gap was associated with several imaging measures, including lower whole-brain (β= −0.002, 95% CI −0.003 to −0.001, P =0.002, FDR=0.02), and lower peripapillary retinal nerve fiber layer thickness (β= −0.21, 95% CI −0.39 to −0.03, P =0.02, FDR=0.10). Proteomic age acceleration in multiple sclerosis is detectable years before symptom onset and distinct organ-specific aging signatures are associated with disease severity. Proteomic aging may provide a biologically informative marker of early disease processes and a clinically relevant readout of disease heterogeneity.
- 20 February 2026
- Published by SAGE Publications in Multiple Sclerosis Journal
- Vol. 32 (4) , 421-428
Accelerated metabolomic aging and its association with social determinants of health in multiple sclerosis
Objectives: Biological age may better capture differences in disease course among people with multiple sclerosis (PwMS) of identical chronological age. We investigated biological age acceleration through metabolomic age (mAge) in PwMS and its association with social determinants of health (SDoH) measured by area deprivation index (ADI). Methods: mAge was calculated for two cohorts: 323 PwMS and 66 healthy controls (HCs); 101 HCs and 71 DMT-naïve PwMS, using an aging clock derived from 11,977 healthy adults. mAge acceleration, the difference between mAge and chronological age, was compared between groups using generalized linear and mixed-effects models, and its association with ADI was assessed via linear regression. Results: Estimated mAge showed a modest correlation with chronological age (R 2 ≈0.20–0.25). Cross-sectionally, PwMS had higher age acceleration than HCs: 9.77 years in adult PwMS (95% CI: 6.57–12.97, p = 5.3e-09), and 4.90 years in adult DMT-naïve PwMS (95% CI: 0.85–9.01, p = 0.02). Longitudinally, PwMS aged 1.19 mAge years per chronological year (95% CI: 0.18–2.20; p = 0.02), faster than HCs. In PwMS, a 10-percentile increase in ADI was associated with a 0.63-year (95% CI: 0.10–1.18; p = 0.02) increase in age acceleration. Discussion: We demonstrated accelerated mAge in PwMS and its association with social disadvantage.
- 20 February 2026
- Published by SAGE Publications in Multiple Sclerosis Journal
- Vol. 32 (4) , 421-428
Accelerated metabolomic aging and its association with social determinants of health in multiple sclerosis
Objectives: Biological age may better capture differences in disease course among people with multiple sclerosis (PwMS) of identical chronological age. We investigated biological age acceleration through metabolomic age (mAge) in PwMS and its association with social determinants of health (SDoH) measured by area deprivation index (ADI). Methods: mAge was calculated for two cohorts: 323 PwMS and 66 healthy controls (HCs); 101 HCs and 71 DMT-naïve PwMS, using an aging clock derived from 11,977 healthy adults. mAge acceleration, the difference between mAge and chronological age, was compared between groups using generalized linear and mixed-effects models, and its association with ADI was assessed via linear regression. Results: Estimated mAge showed a modest correlation with chronological age (R 2 ≈0.20–0.25). Cross-sectionally, PwMS had higher age acceleration than HCs: 9.77 years in adult PwMS (95% CI: 6.57–12.97, p = 5.3e-09), and 4.90 years in adult DMT-naïve PwMS (95% CI: 0.85–9.01, p = 0.02). Longitudinally, PwMS aged 1.19 mAge years per chronological year (95% CI: 0.18–2.20; p = 0.02), faster than HCs. In PwMS, a 10-percentile increase in ADI was associated with a 0.63-year (95% CI: 0.10–1.18; p = 0.02) increase in age acceleration. Discussion: We demonstrated accelerated mAge in PwMS and its association with social disadvantage.
- 19 September 2025
- Published by Wiley in Annals of Neurology
- Vol. 98 (6) , 1278-1282
Fasting Mimicking Diets Reverse Accelerated Biological Aging in Multiple Sclerosis
Reversing the aging process may yield significant benefits in people with multiple sclerosis (PwMS), as accelerated biological aging is observed in this population. Secondary analyses of 2 previously conducted dietary interventions including a 6‐month modified ketogenic diet in 39 participants and an 8‐week randomized comparison of intermittent and daily calorie restriction versus a weight‐stable regimen in 36 participants demonstrated significant decreases in metabolomic age (mAge) following the ketogenic diet ( p = 0.009) and intermittent calorie restriction ( p = 0.04), whereas daily calorie reduction had no effect. These findings indicate that fasting‐mimicking diets (FMDs) can reverse accelerated metabolomic aging in PwMS. ANN NEUROL 2025;98:1278–1282
- 19 September 2025
- Published by Wiley in Annals of Neurology
- Vol. 98 (6) , 1278-1282
Fasting Mimicking Diets Reverse Accelerated Biological Aging in Multiple Sclerosis
Reversing the aging process may yield significant benefits in people with multiple sclerosis (PwMS), as accelerated biological aging is observed in this population. Secondary analyses of 2 previously conducted dietary interventions including a 6‐month modified ketogenic diet in 39 participants and an 8‐week randomized comparison of intermittent and daily calorie restriction versus a weight‐stable regimen in 36 participants demonstrated significant decreases in metabolomic age (mAge) following the ketogenic diet ( p = 0.009) and intermittent calorie restriction ( p = 0.04), whereas daily calorie reduction had no effect. These findings indicate that fasting‐mimicking diets (FMDs) can reverse accelerated metabolomic aging in PwMS. ANN NEUROL 2025;98:1278–1282
- Open Access
Accelerated Metabolomic Aging and Its Association with Social Determinants of Health in Multiple Sclerosis
Objectives: Biological age may better capture differences in disease course among people with multiple sclerosis (PwMS) of identical chronological age. We investigated biological age acceleration through metabolomic age (mAge) in PwMS and its association with social determinants of health (SDoH) measured by area deprivation index (ADI). Methods: mAge was calculated for three cohorts: 323 PwMS and 66 healthy controls (HCs); 101 HCs and 71 DMT-naïve PwMS; and 64 HCs and 67 pediatric-onset MS/clinically isolated syndrome patients, using an aging clock derived from 11,977 healthy adults. mAge acceleration, the difference between mAge and chronological age, was compared between groups using generalized linear and mixed-effects models, and its association with ADI was assessed via linear regression. Results: Cross-sectionally, PwMS had higher age acceleration than HCs: 9.77 years in adult PwMS (95% CI:6.57–12.97, p=5.3e-09), 4.90 years in adult DMT-naïve PwMS (95% CI:0.85–9.01, p=0.02), and 6.98 years (95% CI:1.58–12.39, p=0.01) in pediatric-onset PwMS. Longitudinally, PwMS aged 1.19 mAge years per chronological year (95% CI:0.18, 2.20; p=0.02), faster than HCs. In PwMS, a 10-percentile increase in ADI was associated with a 0.63-year (95% CI:0.10-1.18; p=0.02) increase in age acceleration. Discussion: We demonstrated accelerated mAge in adult and pediatric-onset PwMS and its association with social disadvantage.
Open AccessAccelerated Metabolomic Aging and Its Association with Social Determinants of Health in Multiple Sclerosis
Objectives: Biological age may better capture differences in disease course among people with multiple sclerosis (PwMS) of identical chronological age. We investigated biological age acceleration through metabolomic age (mAge) in PwMS and its association with social determinants of health (SDoH) measured by area deprivation index (ADI). Methods: mAge was calculated for three cohorts: 323 PwMS and 66 healthy controls (HCs); 101 HCs and 71 DMT-naïve PwMS; and 64 HCs and 67 pediatric-onset MS/clinically isolated syndrome patients, using an aging clock derived from 11,977 healthy adults. mAge acceleration, the difference between mAge and chronological age, was compared between groups using generalized linear and mixed-effects models, and its association with ADI was assessed via linear regression. Results: Cross-sectionally, PwMS had higher age acceleration than HCs: 9.77 years in adult PwMS (95% CI:6.57–12.97, p=5.3e-09), 4.90 years in adult DMT-naïve PwMS (95% CI:0.85–9.01, p=0.02), and 6.98 years (95% CI:1.58–12.39, p=0.01) in pediatric-onset PwMS. Longitudinally, PwMS aged 1.19 mAge years per chronological year (95% CI:0.18, 2.20; p=0.02), faster than HCs. In PwMS, a 10-percentile increase in ADI was associated with a 0.63-year (95% CI:0.10-1.18; p=0.02) increase in age acceleration. Discussion: We demonstrated accelerated mAge in adult and pediatric-onset PwMS and its association with social disadvantage.
- Open Access
- 26 August 2024
- Published by Wiley in Annals of Clinical and Translational Neurology
- Vol. 11 (9) , 2485-2498
Ocrelizumab alters the circulating metabolome in people with relapsing–remitting multiple sclerosis
Background: Circulating metabolite levels are altered in multiple sclerosis (MS) and are associated with MS severity. However, how metabolic profiles shift following highly efficacious therapies, like ocrelizumab remains unclear.Objective: Circulating metabolite levels are altered in multiple sclerosis (MS) and are associated with MS severity. However, how metabolic profiles shift following highly efficacious therapies, like ocrelizumab remains unclear. To assess changes in the circulating metabolome produced by ocrelizumab treatment in people with relapsing–remitting MS (RRMS).Methods: Thirty‐one individuals with RRMS eligible for beginning treatment with ocrelizumab were recruited and followed with demographic, clinical, quality‐of‐life, and global metabolomics data collected at each visit. Modules of highly correlated metabolites were identified using the weighted correlation network analysis approach. Changes in each module's eigenmetabolite values and individual metabolites during the study were evaluated using linear mixed‐effects models.Results: Patients with a mean age of 40.8 (SD = 10.30) years, and median disease duration of 4.0 (IQR = 8.5) years, were monitored for a median of 3.36 (IQR = 1.43) years. Two out of twelve identified sets of metabolites were altered significantly. The first module mainly contained androgenic and pregnenolone steroids (p‐value p‐value = 0.016, coefficient: −0.12), which its reduction was significantly associated with improvement based on overall disability response score (OR 3.09e‐01, 95% CI: 6.83e‐02, 9.09e‐01, p‐value = 3.15E‐02).Interpretation: In this longitudinal observational study, using a global untargeted metabolomics approach, we showed significant alteration in circulating metabolome in RRMS patients undergoing ocrelizumab treatment. In particular, we observed a significant reduction in metabolites involved in the lysophospholipid pathway, which was associated with patients' improvement.
- 26 August 2024
- Published by Wiley in Annals of Clinical and Translational Neurology
- Vol. 11 (9) , 2485-2498
Open AccessOcrelizumab alters the circulating metabolome in people with relapsing–remitting multiple sclerosis
Background: Circulating metabolite levels are altered in multiple sclerosis (MS) and are associated with MS severity. However, how metabolic profiles shift following highly efficacious therapies, like ocrelizumab remains unclear.Objective: Circulating metabolite levels are altered in multiple sclerosis (MS) and are associated with MS severity. However, how metabolic profiles shift following highly efficacious therapies, like ocrelizumab remains unclear. To assess changes in the circulating metabolome produced by ocrelizumab treatment in people with relapsing–remitting MS (RRMS).Methods: Thirty‐one individuals with RRMS eligible for beginning treatment with ocrelizumab were recruited and followed with demographic, clinical, quality‐of‐life, and global metabolomics data collected at each visit. Modules of highly correlated metabolites were identified using the weighted correlation network analysis approach. Changes in each module's eigenmetabolite values and individual metabolites during the study were evaluated using linear mixed‐effects models.Results: Patients with a mean age of 40.8 (SD = 10.30) years, and median disease duration of 4.0 (IQR = 8.5) years, were monitored for a median of 3.36 (IQR = 1.43) years. Two out of twelve identified sets of metabolites were altered significantly. The first module mainly contained androgenic and pregnenolone steroids (p‐value p‐value = 0.016, coefficient: −0.12), which its reduction was significantly associated with improvement based on overall disability response score (OR 3.09e‐01, 95% CI: 6.83e‐02, 9.09e‐01, p‐value = 3.15E‐02).Interpretation: In this longitudinal observational study, using a global untargeted metabolomics approach, we showed significant alteration in circulating metabolome in RRMS patients undergoing ocrelizumab treatment. In particular, we observed a significant reduction in metabolites involved in the lysophospholipid pathway, which was associated with patients' improvement.
- Open Access
- 10 February 2023
- Published by Springer Nature in Scientific Reports
- Vol. 13 (1) , 1-11
Generalizable machine learning approach for COVID-19 mortality risk prediction using on-admission clinical and laboratory features
We aimed to propose a mortality risk prediction model using on-admission clinical and laboratory predictors. We used a dataset of confirmed COVID-19 patients admitted to three general hospitals in Tehran. Clinical and laboratory values were gathered on admission. Six different machine learning models and two feature selection methods were used to assess the risk of in-hospital mortality. The proposed model was selected using the area under the receiver operator curve (AUC). Furthermore, a dataset from an additional hospital was used for external validation. 5320 hospitalized COVID-19 patients were enrolled in the study, with a mortality rate of 17.24% (N = 917). Among 82 features, ten laboratories and 27 clinical features were selected by LASSO. All methods showed acceptable performance (AUC > 80%), except for K-nearest neighbor. Our proposed deep neural network on features selected by LASSO showed AUC scores of 83.4% and 82.8% in internal and external validation, respectively. Furthermore, our imputer worked efficiently when two out of ten laboratory parameters were missing (AUC = 81.8%). We worked intimately with healthcare professionals to provide a tool that can solve real-world needs. Our model confirmed the potential of machine learning methods for use in clinical practice as a decision-support system.
- 10 February 2023
- Published by Springer Nature in Scientific Reports
- Vol. 13 (1) , 1-11
Open AccessGeneralizable machine learning approach for COVID-19 mortality risk prediction using on-admission clinical and laboratory features
We aimed to propose a mortality risk prediction model using on-admission clinical and laboratory predictors. We used a dataset of confirmed COVID-19 patients admitted to three general hospitals in Tehran. Clinical and laboratory values were gathered on admission. Six different machine learning models and two feature selection methods were used to assess the risk of in-hospital mortality. The proposed model was selected using the area under the receiver operator curve (AUC). Furthermore, a dataset from an additional hospital was used for external validation. 5320 hospitalized COVID-19 patients were enrolled in the study, with a mortality rate of 17.24% (N = 917). Among 82 features, ten laboratories and 27 clinical features were selected by LASSO. All methods showed acceptable performance (AUC > 80%), except for K-nearest neighbor. Our proposed deep neural network on features selected by LASSO showed AUC scores of 83.4% and 82.8% in internal and external validation, respectively. Furthermore, our imputer worked efficiently when two out of ten laboratory parameters were missing (AUC = 81.8%). We worked intimately with healthcare professionals to provide a tool that can solve real-world needs. Our model confirmed the potential of machine learning methods for use in clinical practice as a decision-support system.