Research
Elucidating Genetic and Environmental Second Hits in Racial and Ethnic Minorities with APOL1 High-Risk Genotypes
Mutations in a specific gene (Apolipoprotein L1 or APOL1) seen only in people of African ancestry (African Americans or Hispanic Americans) is linked to high rates of kidney disease, with only some of the individuals with these mutations developing kidney disease, indicating that other genetic or environmental factors are involved. We will explore these factors including other genes and common environmental exposures like air pollution/neighborhood environment, and heavy metal exposure. We will understand what additional risk factors contribute to kidney disease in individuals with this mutation, which will lead to preventative efforts to address the persistent racial and ethnic disparities in kidney disease.
​
Relevant Publications:
Nadkarni GN, Gignoux CR, Sorokin EP, Daya M, Rahman R, Barnes KC, Wassel CL, Kenny EE. Worldwide Frequencies of APOL1 Renal Risk Variants. N Engl J Med. 2018 Dec 27;379(26):2571-2572. PMID: 30586505; PMCID: PMC6482949.
​
Paranjpe I, Chaudhary K, Paranjpe M, O'Hagan R, Manna S, Jaladanki S, Kapoor A, Horowitz C, DeFelice N, Cooper R, Glicksberg B, Bottinger EP, Just AC, Nadkarni GN. Association of APOL1 Risk Genotype and Air Pollution for Kidney Disease. Clin J Am Soc Nephrol. 2020 Mar 6;15(3):401-403. PMID: 32079610; PMCID: PMC7057301.
​
Elucidating hereditary transthyretin-mediated heart failure risk using machine learning, polygenic risk and recall by genotype approaches in African ancestry individuals
Transthyretin V122I, is a highly penetrant mutation for hereditary transthyretin amyloid cardiomyopathy (hATTR-CM) that is common in African Americans and Hispanic Americans. Recent therapies have been approved to treat hATTR-CM, which have led to decreased mortality; thus, understanding who will develop hATTR-CM, accurate identification of V122I carriers and understanding hidden disease burden is important. In this proposal, we examine the population health impact of V122I using polygenic risk scores, machine learning approaches on electronic health records, and patient recall by genotype, which can lead to new population health strategies for heart failure in racial and ethnic minorities.
​
Relevant Publications:
Damrauer SM, Chaudhary K, Cho JH, Liang LW, Argulian E, Chan L, Dobbyn A, Guerraty MA, Judy R, Kay J, Kember RL, Levin MG, Saha A, Van Vleck T, Verma SS, Weaver J, Abul-Husn NS, Baras A, Chirinos JA, Drachman B, Kenny EE, Loos RJF, Narula J, Overton J, Reid J, Ritchie M, Sirugo G, Nadkarni G*, Rader DJ*, Do R*. Association of the V122I Hereditary Transthyretin Amyloidosis Genetic Variant With Heart Failure Among Individuals of African or Hispanic/Latino Ancestry. JAMA. 2019 Dec 10;322(22):2191-2202. . PMID: 31821430; PMCID: PMC7081752.
Artificial Intelligence to Predict Outcomes in Patients with Acute Kidney Injury on Continuous Renal Replacement Therapy
Continuous renal replacement therapy (CRRT) is the preferred dialysis treatment for critically ill patients with acute kidney injury and hemodynamic instability, yet mortality is high (~75%). Presently, there are no universally accepted approaches for predicting kidney recovery, survival or individual response to fluid removal during CRRT. We propose to develop and validate innovative deep learning approaches to dynamically predict these outcomes, which could guide CRRT decision-making including intensification, de-escalation, and enrich clinical trials focusing on fluid removal during CRRT.
​
Relevant Publications:
1) Chaudhary K, Vaid A, Duffy Á, Paranjpe I, Jaladanki S, Paranjpe M, Johnson K, Gokhale A, Pattharanitima P, Chauhan K, O'Hagan R, Van Vleck T, Coca SG, Cooper R, Glicksberg B, Bottinger EP, Chan L, Nadkarni GN. Utilization of Deep Learning for Subphenotype Identification in Sepsis-Associated Acute Kidney Injury. Clin J Am Soc Nephrol. 2020 Nov 6;15(11):1557-1565. PMID: 33033164.
Long term kidney outcomes of COVID-19
Over a third of patients hospitalized with COVID-19 suffer from severe acute kidney injury and likely will experience a large burden of chronic kidney disease. We will leverage clinical data and blood, DNA, and urine samples from the Mount Sinai COVID-19 Center of Excellence Longitudinal Registry to study predictors and mechanisms of COVID-19 associated kidney disease. The project has potential to provide important insights for therapeutic and clinical management of COVID-19 related kidney disease.
​
Relevant Publications:
Chan L, Chaudhary K, Saha A, Chauhan K, Vaid A, Zhao S, Paranjpe I, Somani S, Richter F, Miotto R, Lala A, Kia A, Timsina P, Li L, Freeman R, Chen R, Narula J, Just AC, Horowitz C, Fayad Z, Cordon-Cardo C, Schadt E, Levin MA, Reich DL, Fuster V, Murphy B, He JC, Charney AW, Böttinger EP, Glicksberg BS, Coca SG, Nadkarni GN; Mount Sinai COVID Informatics Center (MSCIC). AKI in Hospitalized Patients with COVID-19. J Am Soc Nephrol. 2020 Sep 3. Epub ahead of print. PMID: 32883700.
​
Chan L, Jaladanki SK, Somani S, Paranjpe I, Kumar A, Zhao S, Kaufman L, Leisman S, Sharma S, He JC, Murphy B, Fayad ZA, Levin MA, Bottinger EP, Charney AW, Glicksberg BS, Coca SG, Nadkarni GN; Mount Sinai COVID Informatics Center (MSCIC). Outcomes of Patients on Maintenance Dialysis Hospitalized with COVID-19. Clin J Am Soc Nephrol. 2020 Oct 30. PMID: 33127607.
Integration of Multi-Modal Data to Improve Prediction and Clinical Implementation
Prediction of key clinical outcomes in hospitalized patients with COVID-19 is critical for appropriate clinical management, workflow optimization, and intelligent resource allocation. We will integrate multi-modal patient
data (EKG, physiological, narrative data) not routinely used for prediction, develop machine learning models, and prospectively validate and deploy them in clinical care, along with dissemination of our data, models and code for broad adoption. This project has the potential to reshape how hospitalized patients are managed.
​
Relevant Publications:
Vaid A, Somani S, Russak AJ, De Freitas JK, Chaudhry FF, Paranjpe I, Johnson KW, Lee SJ, Miotto R, Richter F, Zhao S, Beckmann ND, Naik N, Kia A, Timsina P, Lala A, Paranjpe M, Golden E, Danieletto M, Singh M, Meyer D, O'Reilly PF, Huckins L, Kovatch P, Finkelstein J, Freeman RM, Argulian E, Kasarskis A, Percha B, Aberg JA, Bagiella E, Horowitz CR, Murphy B, Nestler EJ, Schadt EE, Cho JH, Cordon-Cardo C, Fuster V, Charney DS, Reich DL, Bottinger EP, Levin MA, Narula J, Fayad ZA, Just AC, Charney AW, Nadkarni GN*, Glicksberg BS*. Machine Learning to Predict Mortality and Critical Events in a Cohort of Patients With COVID-19 in New York City: Model Development and Validation. J Med Internet Res. 2020 Nov 6;22(11) PMID: 33027032; PMCID: PMC7652593.
Natural Language Processing to Improve Phenotyping
Electronic and administrative health records have multidimensional and longitudinal data in structured, unstructured and mixed formats, including diagnostic codes, laboratory data, test reports, demographics, and clinician notes. Health systems and researchers have largely relied on structured data elements in the EHR for tasks like risk modeling. In contrast, few have tapped unstructured data in the EHR. NLP and deep learning open the door to using this rich resource by enabling the extraction of specific features and mapping them to biomedical terms, transforming unstructured data into structured and encoded data for analysis. We are using this for the use cases of end stage kidney disease, psychiatric disease and diagnosing cognitive impairment in individuals at high risk.
​
Relevant Publications:
Van Vleck TT, Chan L, Coca SG, Craven CK, Do R, Ellis SB, Kannry JL, Loos RJF, Bonis PA, Cho J, Nadkarni GN. Augmented intelligence with natural language processing applied to electronic health records for identifying patients with non-alcoholic fatty liver disease at risk for disease progression. Int J Med Inform. 2019 Sep;129:334-341. PMID: 31445275; PMCID: PMC6717556. 2) Chan L, Beers K, Yau AA, Chauhan K, Duffy Á, Chaudhary K, Debnath N, Saha A, Pattharanitima P, Cho J, Kotanko P, Federman A, Coca SG, Van Vleck T, Nadkarni GN. Natural language processing of electronic health records is superior to billing codes to identify symptom burden in hemodialysis patients. Kidney Int. 2020 Feb;97(2):383-392.PMID: 31883805; PMCID: PMC7001114.
Federated Learning for Prediction of Outcomes
Larger representation from additional populations is needed for generalizability particularly for machine learning . Large-scale initiatives are meta-analyzing data from several hospitals, but this does not allow joint machine learning. In light of patient privacy, federated learning is emerging as a promising strategy where patient data are fragmented across health systems . Federated learning allows for decentralized refinement of independently built ML models via iterative exchange of model parameters to a central aggregator without sharing raw data.
​
Relevant Publications:
Vaid A, Jaladanki SK, Xu J, Teng S, Kumar A, Lee S, Somani S, Paranjpe I, De Freitas JK, Wanyan T, Johnson KW, Bicak M, Klang E, Kwon YJ, Costa A, Zhao S, Miotto R, Charney AW, Böttinger E, Fayad ZA, Nadkarni GN*, Wang F*, Glicksberg BS*. Federated Learning of Electronic Health Records Improves Mortality Prediction in Patients Hospitalized with COVID-19. medRxiv [Preprint]. 2020 Aug 14:2020.08. PMID: 32817979; PMCID: PMC7430624.