ESTRO 2024 - Abstract Book

S4431

Physics - Machine learning models and clinical applications

ESTRO 2024

distinguishable. These results align with current clinical practices, underscoring the significance of considering early gamma changes to the skin in conjunction with factors like clinically evident dermatitis or weight loss, which should stimulate consideration of replanning. Our future direction involves extending this research into a prospective study on a larger patient cohort to develop a predictive model. This innovative approach offers seamless integration into clinical practice, requiring minimal time and financial resources, ultimately enhancing patient care and laying the foundation for patient-specific adaptation in head and neck radiotherapy.

Keywords: ART, toxicity predictors, head-and-neck

References:

1. Trotti, A., Toxicity in head and neck cancer: a review of trends and issues. International Journal of Radiation Oncology Biology Physics, 2000. 47(1): p. 1-12.

2. Weistrand, O. and S. Svensson, The ANACONDA algorithm for deformable image registration in radiotherapy. Medical physics, 2015. 42(1): p. 40-53.

709

Digital Poster

Exploring Varian RapidPlan models for prostate, cranial radiosurgery (SRS), and lung SBRT

Karen Pieri, Albino D. Lucas, Matheus F. Santos, José G. Calado, Ernesto H. Roesler, Roberta R.G.C. Kramer, Sonmi Lee, Susane P. Leite, Jonathan A. Melo

Real Hospital Português de Beneficência em Pernambuco, Radioerapia, Recife, Brazil

Purpose/Objective:

The integration of artificial intelligence (AI) into radiation therapy has transformed cancer treatment by significantly improving precision in radiation dose planning. AI's capability to analyze patient data and account for anatomical variations has revolutionized conventional methods that rely heavily on healthcare professionals' expertise. Varian's RapidPlan software has emerged as a leading solution, setting high standards for treatment quality. It provides customized treatment plans, optimizes radiation doses, streamlines planning processes, ensures consistency, and continually evolves through the incorporation of evolving patient data. [1] [2] This study's main objective is to create and rigorously validate three models, each tailored to different anatomical sites with varying complexities, all implemented using RapidPlan. The research aims to explore the dosimetric excellence of treatment plans generated by the software. By utilizing AI to personalize treatment plans based on individual patient anatomy and refine radiation dosages, RapidPlan significantly enhances the precision of cancer treatment while minimizing adverse side effects. Furthermore, the software's consistent and standardized approach