Advancing artificial intelligence-assisted pre-screening for fragile X syndrome

Area of Science:

• Medical Informatics
• Genetics
• Artificial Intelligence

Background:

• Fragile X syndrome (FXS) is a leading inherited cause of intellectual disability and autism, yet it is significantly underdiagnosed.
• Diagnostic challenges stem from heterogeneous phenotypes, subtle early signs, and symptom overlap with other conditions.
• There is a critical need for improved screening methods to facilitate earlier FXS identification.

Purpose of the Study:

• To evaluate the effectiveness of an artificial intelligence (AI) and electronic health records (EHRs) system for accelerating the diagnosis of Fragile X syndrome.
• To develop and test an AI-assisted pre-screening tool for identifying individuals with FXS.

Main Methods:

• A retrospective study utilized EHR data from 2.1 million patients at University of Wisconsin Health System (1988-2021).
• A supervised random forest classifier was employed to build an AI pre-screening tool.
• Individuals diagnosed with FXS after age 10 were included, with the AI tool aiming to identify cases 5 years prior to clinical diagnosis.

Main Results:

• The AI-assisted pre-screening tool identified FXS cases 5 years earlier than clinical diagnosis, achieving an AUROC of 0.717.
• A model trained and tested on UW Health cases demonstrated a higher AUROC of 0.798.
• The study identified 52 FXS cases matched with 5200 controls.

Conclusions:

• The developed AI tool shows significant potential for accelerating FXS diagnosis in clinical practice.
• Earlier diagnosis of FXS can facilitate prompt intervention and access to necessary services.
• Timely diagnosis is crucial for improving health outcomes for individuals with Fragile X syndrome.