Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: DRA : Canada Graduate Research Scholarship - Doctoral Program (CGRS D)

Direct Payment

$120,000.00

Enabling Multimodal Clinical Decision Making with ECG, Imaging, EHR, and Genomic Data

EF 2025-26

Résumé
Modern healthcare systems generate massive amounts of data. From the notes that family doctors type out, to MRI scans that can view our internal organs, we have the potential to create incredibly detailed views of a patient's health status. However, our modern medical tests need to be interpreted by highly specialized doctors, which can make it challenging for a single doctor to fully utilize the vast amount of medical data at their disposal. This presents an opportunity for multimodal machine learning models, a form of artificial intelligence that can learn from different types of data, like images and text, all at once. Imagine a model that can combine MRI images, chest X-rays, blood report results, and text based health records all together before making a clinical decision. These models could detect the risk of diseases like cancer well before their symptoms show, allowing for early detection and better healthcare outcomes. They could also learn patterns in the medical data that we as humans have not been able to find, allowing for more accurate clinical decision making. One area of modern science that is largely underutilized is the personalized genomic data, or DNA, of patients. One barrier is that most physicians cannot readily use this data in their daily practice, even if in theory it might be useful. For instance, a patient's genomic information might help to determine which antidepressant works best for them, but without the data and models to make these connections, physicians often rely on trial and error. Machine learning models could make this data easier for physicians to work with. My work aims to take existing models that work on specific data types, like images or text, and combine them such that a single model has a comprehensive view of a patient's health. I plan to use large datasets that have been created for medical research, including the UK Biobank, the All of Us project in the USA, and potentially Canada's own Gemini project.
Programme
DRA : Canada Graduate Research Scholarship - Doctoral Program (CGRS D)
Bénéficiaire
Emplacement
Massachusetts, United States of America
Type
Grant
Source
CIHR Funding Decisions Database

Tel que publié

Id
556163
Name
Kahidi Ramin
Region
Massachusetts
Country
United States of America
Orgtype
Research Institution
Instname2
Population and Public Health
Theme2
Biomedical
Prcname2
DT1 (Biomedical and/or Clinical Research) CGRS Doctoral
Projecttitle
Enabling Multimodal Clinical Decision Making with ECG, Imaging, EHR, and Genomic Data
Programtype2
Studentship Programs
Orgnameinp2
Direct Payment
Cihramount2
$120,000
Cihrequipment2
$0
Programname2
DRA : Canada Graduate Research Scholarship - Doctoral Program (CGRS D)
Keyworddelim
Ai Agents; Artificial Intelligence; Electronic Health Records; Genomics; Healthcare Datasets; Imaging; Large Language Models; Machine Learning; Multimodal; Predictive Models
Abstract
Modern healthcare systems generate massive amounts of data. From the notes that family doctors type out, to MRI scans that can view our internal organs, we have the potential to create incredibly detailed views of a patient's health status. However, our modern medical tests need to be interpreted by highly specialized doctors, which can make it challenging for a single doctor to fully utilize the vast amount of medical data at their disposal. This presents an opportunity for multimodal machine learning models, a form of artificial intelligence that can learn from different types of data, like images and text, all at once. Imagine a model that can combine MRI images, chest X-rays, blood report results, and text based health records all together before making a clinical decision. These models could detect the risk of diseases like cancer well before their symptoms show, allowing for early detection and better healthcare outcomes. They could also learn patterns in the medical data that we as humans have not been able to find, allowing for more accurate clinical decision making. One area of modern science that is largely underutilized is the personalized genomic data, or DNA, of patients. One barrier is that most physicians cannot readily use this data in their daily practice, even if in theory it might be useful. For instance, a patient's genomic information might help to determine which antidepressant works best for them, but without the data and models to make these connections, physicians often rely on trial and error. Machine learning models could make this data easier for physicians to work with. My work aims to take existing models that work on specific data types, like images or text, and combine them such that a single model has a comprehensive view of a patient's health. I plan to use large datasets that have been created for medical research, including the UK Biobank, the All of Us project in the USA, and potentially Canada's own Gemini project.
Competitiondate
202511
Approvedterm2
3 yrs 0 mth
Pinamesdelim
Kahidi, Ramin
Approvedterm
3 ans 0 mois
Cihrcontribution2
~ $120,000~ $0~