ESTRO 2025 - Abstract Book

S747

Clinical - CNS

ESTRO 2025

acceptable toxicity with the use of Cyberknife technology.

Keywords: Radiosurgery, cyberknife, metastases

4248

Digital Poster Advancing Pediatric Radiotherapy: Integrating AI-Driven Radiomics and Global Collaboration to Enhance Precision and Humanize Care Aida Angela Tummolo 1,2 , Silvia Chiesa 1 , Sabina Vennarini 2 , Elisa Meldolesi 1 , Silvia Mariani 1 , Elisa Marconi 3,1 , Mariangela Massaccesi 4 , Gerardina Stimato 1 , Luca Boldrini 1 , Antonio Ruggiero 5 , Angela Mastrouzzi 6 , Vincenzo Frascino 1 , Nicola Dinapoli 1 , Luca Tagliaferri 1 , Maria Antonietta Gambacorta 1 1 Gemelli ART, Scienze Radiologiche, Radioterapiche ed Ematologiche, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Roma, Italy. 2 Radioterapia Oncologica, Istituto Nazionale dei Tumori IRCCS, Milano, Italy. 3 UOS Psicologia Clinica, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Roma, Italy. 4 Gemelli ART, Scienze Radiologiche, Radioterapiche ed Ematologichemagini, ematologia e radioterapia, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Roma, Italy. 5 Scienze della salute della donna, del bambino e di sanità pubblica, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Roma, Italy. 6 Ematologia, Oncologia, Terapia Cellulare, Terapia Genetica e Trapianto Emopoietico, Ospedale Pediatrico Bambino Gesù, Roma, Italy Purpose/Objective: Radiation therapy represents a cornerstone in treating specific pediatric malignancies, necessitating advanced expertise in treatment planning and execution. With the advent of cutting-edge technologies, the development of precise and personalized therapies has become feasible. Notably, machine learning approaches facilitate the identification of predictive and prognostic radiomic parameters. For example, studies using MRI from patients with central nervous system (CNS) neoplasms have revealed radiomic indicators that, when integrated with conventional MRI data, provide more nuanced prognostic insights. Alongside these technological advances, the humanization of care remains critical. Many pediatric patients must travel, often internationally, to specialized centers due to limited local expertise or resources, disrupting their daily lives and detaching them from familiar surroundings. Material/Methods: This research initiative involves a comprehensive review of pediatric CNS neoplasm cases treated with radiotherapy over the past five years. The project will integrate a radiomic approach, applying quantitative MRI image analysis to identify radiomic parameters with predictive and prognostic significance, particularly for high-risk neoplasms. Additionally, the project envisions creating an extensive database involving collaboration with other pediatric radiotherapy centers. This initiative aims to establish a national and international network to standardize and enhance clinical practices, with particular attention to centers in developing countries. An educational program will underpin this endeavor, offering both theoretical and hands-on training led by field experts. Results: We hypothesize that machine learning-based identification of radiomic parameters will support the development of highly personalized treatments, potentially improving therapeutic outcomes. Establishing a comprehensive database will elucidate variations in treatment protocols across centers, enhancing our understanding of indications, prescription methodologies, and radiotherapy planning and delivery processes. Conclusion: Analyzing inter-center procedural differences, alongside clinical follow-up data, will aid in devising a standardized and widely accepted treatment framework. The educational initiatives will foster a collaborative network for clinical case sharing, enabling incremental expertise acquisition by participating centers. This approach aspires to ensure that pediatric patients can access radiotherapy within their local context, preserving geographic and familial