Exciting Milestones in Precision Radiotherapy (PRT)
Our precision radiotherapy (PRT) efforts have reached new heights with **16 accepted abstracts** at ESTRO this year, surpassing last year’s 13. Collaborating with esteemed clinical and academic partners, these insights can revolutionize radiotherapy (RT) workflows. Key focus areas include automated treatment planning, vendor-agnostic adaptive radiation therapy, magnetic resonance-only RT planning, and uncertainty-aware automatic delineation of tumors and organs at risk, guided by consensus protocols.
End-to-End Automation for Prostate Cancer Treatment Planning
We're excited about ART-Plan®, which is soon set to offer comprehensive end-to-end automated and robust radiotherapy planning. By incorporating dose prediction into treatment planning, we address individual patient anatomy and tumor characteristics with precision. Our system efficiently delivers complex treatment parameters, ensuring optimal dosing while accounting for uncertainties (Ungun et al., Delasalles et al.). Dubois et al. took treatment planning further by introducing robustness through a graph-based multi-plan optimization strategy. AI-enabled automation improves planning reliability, enhances robustness, and tailors treatment to individual patients.
Adaptive Radiotherapy, CBCT Imaging Quality & Magnetic Resonance Workflows
In a groundbreaking achievement, our work has improved cone beam CT (CBCT) imaging to diagnostic quality, transforming radiotherapy treatment. Partnerships like Chalkia et al. demonstrated synthetic CBCT's equivalence for aligning with planning CT in head and neck radiotherapy. ART-Plan®, now equipped with a full dose calculation framework (Colombo et al.), enables real-time dose monitoring and adaptive replanning. This ensures personalized care for critical areas like the breast (Colombob et al.), head and neck (Colomboc et al.), and thorax (Colombod et al.). Leclercq et al. validated the efficacy of synthetic CT in offline adaptive replanning, confirming its transformative potential. Pantelis et al. explored an MRI-only protocol for intracranial radiosurgery, confirming its reliability. Finally, Cafaro et al. employed generative AI for volumetric reconstruction using bi-planar imaging.
Elevating Organ and Tumor Delineation
Achieving precise organ and target volume delineation is challenging. Perennec et al. enhanced automatic contouring by integrating anatomical reasoning into the process. Costea et al. showed AI models' generalizability across patient populations, while Schmidt-Mengin et al. introduced a novel weakly supervised method for automatic segmentation. Leroy et al.’s diffusion model, trained with histological ground truth and transferred to CT, outperformed consensus among experts in head and neck target volume delineation.
These strides mark significant progress in advancing radiotherapy precision, reinforcing our commitment to pioneering innovation and improving patient care.