Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Project Grant

McGill University

$321,300.00

Bayesian Approach to Trustworthy AI for Reducing Delays in Oral Cancer Detection

EF 2025-26

Résumé
Oral cancer is a serious disease that causes high levels of illness, disfigurement, and premature death. In Canada and worldwide, many people are diagnosed at a late stage, when treatment is more invasive, outcomes are poorer, and costs are higher. One of the main reasons for this is a delay in diagnosis. Many oral cancers begin as small lesions in the mouth that can be detected earlier, but recognizing these lesions can be challenging for general dentists, especially when they appear similar to harmless conditions. Earlier and more accurate diagnosis could greatly improve survival and quality of life. This project will develop and test a new approach that uses advanced computer models, known as generative artificial intelligence (AI), to assist dentists in diagnosing oral lesions from information recorded during a dental visit. These AI models can read and interpret the detailed notes dentists write about a patient's symptoms, history, and examination. Our innovation is a Bayesian method that allows the AI to express how certain or uncertain it is about a diagnosis. This is important because in health care, knowing when a tool is unsure can help guide safer decisions. We will create a high-quality benchmark dataset from real patient charts, licensing exam questions, and specially designed test cases to evaluate the AI's accuracy, ability to reason like a clinician, and level of certainty. We will also compare the AI's performance with that of Canadian dentists. The findings will guide safe and trustworthy use of AI in dental practice, help reduce delays in oral cancer detection, and improve population health.
Programme
Project Grant
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
CIHR Funding Decisions Database

Tel que publié

Id
533489
Name
Madathil Sreenath A
Region
Québec
Country
Canada
Orgtype
Institution Paid; Research Institution
Instname2
Cancer Research
Theme2
Social / Cultural / Environmental / Population Health
Prcname2
Public, Community & Population Health 3
Projecttitle
Bayesian Approach to Trustworthy AI for Reducing Delays in Oral Cancer Detection
Programtype2
Operating Grants
Orgnameinp2
McGill University
Cihramount2
$321,300
Cihrequipment2
$0
Programname2
Project Grant
Keyworddelim
Oral Cancer Detection; Trustworthy Ai; Uncertainty Quantification
Abstract
Oral cancer is a serious disease that causes high levels of illness, disfigurement, and premature death. In Canada and worldwide, many people are diagnosed at a late stage, when treatment is more invasive, outcomes are poorer, and costs are higher. One of the main reasons for this is a delay in diagnosis. Many oral cancers begin as small lesions in the mouth that can be detected earlier, but recognizing these lesions can be challenging for general dentists, especially when they appear similar to harmless conditions. Earlier and more accurate diagnosis could greatly improve survival and quality of life. This project will develop and test a new approach that uses advanced computer models, known as generative artificial intelligence (AI), to assist dentists in diagnosing oral lesions from information recorded during a dental visit. These AI models can read and interpret the detailed notes dentists write about a patient's symptoms, history, and examination. Our innovation is a Bayesian method that allows the AI to express how certain or uncertain it is about a diagnosis. This is important because in health care, knowing when a tool is unsure can help guide safer decisions. We will create a high-quality benchmark dataset from real patient charts, licensing exam questions, and specially designed test cases to evaluate the AI's accuracy, ability to reason like a clinician, and level of certainty. We will also compare the AI's performance with that of Canadian dentists. The findings will guide safe and trustworthy use of AI in dental practice, help reduce delays in oral cancer detection, and improve population health.
Competitiondate
202509
Approvedterm2
3 yrs 0 mth
Pinamesdelim
Madathil, Sreenath A; Reddy, Siva
Approvedterm
3 ans 0 mois
Cihrcontribution2
~ $321,300~ $0~