PedRetina Explorer: A Cloud-Based AI Tool for Optimizing Diagnosis and Treatment in Pediatric Retinoblastoma
Mehrana Fallah1 , Masoud Nasseripour2 , Reza Mirshahi2 , Mahdi Khaligh Razavi1 , Leila Satarian1 *
- 1. Department of Stem Cells and Developmental Biology, Cell Science Research Center, Royan Institute for Stem Cell Biology and Technology, ACECR, Tehran, Iran.
- 2. Department of Ophthalmology, Faculty of Medical Sciences, Iran University, Tehran, Iran.
Abstract: Retinoblastoma, the most common intraocular tumor in children, requires timely diagnosis and treatment to prevent severe outcomes like vision loss or eye enucleation. The PedRetina Explorer is a cloud-based tool designed to improve understanding of retinal responses to chemotherapy. It provides a user-friendly interface that allows ophthalmologists to quickly search for and access relevant patient cases, streamlining the process for diagnosis and treatment.
Methods: Initially, we conducted a comprehensive literature review of pertinent articles in pediatric ophthalmology and artificial intelligence to establish the project’s foundation. Subsequently, we collaborated with experts to develop advanced algorithms designed to predict retinal responses in children based on historical cases. We then implemented the PedRetina Explorer search engine, which includes custom filters and data visualization tools. Extensive testing and validation were carried out to ensure the accuracy and reliability of the search engine, involving medical professionals and utilizing diverse datasets.
Results: Traditionally, physicians have depended on manual searches through Electronic Health Records and Picture Archiving and Communication Systems, which can be time-consuming and inefficient. While existing search tools, such as radiology-based Case-Based Reasoning systems and oncology platforms like OncoSearch, offer some functionalities, these tools primarily focus on adult cancers or molecular profiling, leaving a significant gap for solutions tailored to pediatric needs. PedRetina Explorer addresses these gaps by integrating machine learning-based search capabilities that combine retinal images with clinical reports using advanced algorithms. This system enables quick and relevant retrieval of historical cases, empowering doctors to effectively match similar patient profiles, treatments, and outcomes.The incorporation of vector-based databases and natural language processing models bridges the divide between imaging data and textual information, enabling precise queries for swift decision-making. As a cloud-based platform, PedRetina Explorer provides global access with secure authorization, positioning itself as a transformative tool for enhancing patient care in pediatric retinal oncology.
Conclusion: PedRetina Explorer is an innovative tool that analyzes case data to assist physicians in decision-making, research, and improving pediatric retinal chemotherapy. With continuous learning and regular updates, it enhances treatment outcomes while maintaining its relevance.
