Pupillary Patterns as Predictive Keys to detect Inherited Retinal Diseases in Infants: where Machine Learning Meets Medicine
Keywords:
Inherited Retinal Diseases, Retinitis Pigmentosa, Pediatric Ophthalmology, Chromatic Pupillometry, Machine Learning, Support Vector Machine, Long Short-Term Memory, Ensemble Learning Methods, Clinical Decision Support System, Early Detection, Diagnostic Accuracy, Pediatric HealthcareAbstract
In the realm of pediatric ophthalmology, early diagnosis and intervention for inherited retinal diseases (IRDs) are
crucial for mitigating potential vision loss. Traditional diagnostic methods, however, often present challenges in
pediatric patients due to the complexity of genetic factors and clinical presentations. To address this critical need,
this study presents a pioneering Clinical Decision Support System (CDSS) specifically designed for the automatic
detection of IRDs in children, with a primary focus on Retinitis Pigmentosa (RP).
This innovative CDSS leverages Chromatic Pupillometry, a non-invasive technique analyzing pupillary light
reflexes to gain valuable insights into retinal function. The system's power lies in its integration of sophisticated
machine learning algorithms, including Support Vector Machines (SVMs), Long Short-Term Memory networks
(LSTMs), and Ensemble Learning Methods (ELMs), for comprehensive data processing and analysis.
Through rigorous testing, the CDSS has achieved an impressive 89.6% accuracy, 93.7% sensitivity, and 78.6%
specificity in diagnosing RP, establishing its potential as a highly reliable diagnostic tool. This study marks a
significant breakthrough, representing the first successful application of machine learning, specifically SVMs,
LSTMs, and ELMs, to pupillometric data for diagnosing genetic diseases in children. The implications of this
technology are profound, potentially revolutionizing early detection and intervention strategies for IRDs and
leading to significantly improved visual outcomes for young patients.
Future research will focus on validating the system's performance with a larger and more diverse patient cohort,
exploring its applicability in diagnosing a broader range of IRDs, and investigating alternative pupillometry devices
to minimize the impact of movement artifacts, a common challenge in pediatric populations. This groundbreaking
approach holds immense promise for advancing pediatric healthcare by providing a reliable, non-invasive, and
efficient tool for early diagnosis and management of IRDs.
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.













