S2776
RTT - RTT contouring, target definition, and treatment planning
ESTRO 2026
appropriate training, can perform OAR contouring without compromising clinical quality, and how AI- assisted autocontouring can be integrated into a safe and efficient workflow. Material/Methods: We evaluated data from 18 patients (9 males and 9 females) treated with SBRT for lung lesions. For each case, thirteen thoracic OARs were delineated: aorta, heart, pulmonary artery, right and left lungs, trachea, bronchial tree, esophagus, spinal cord, chest wall, right and left breast (in female patients only).Three contour sets were generated per patient: physician (reference standard), APRT (manual), and AI (ART-Plan ™).A total of 216 structures were analyzed. Geometric agreement was quantified using the Dice Similarity Coefficient (DSC) and Hausdorff Distance (HD). Dosimetric comparisons were performed by recalculating treatment plans on alternative contour sets and recording differences in key dose-volume parameters (Dmax, Dmean). Operational efficiency was assessed by recording the contouring time for each operator (physician, APRT, and AI–APRT hybrid workflow). Results: The median DSC for physician–APRT comparison was 0.92 (range 0.87–0.96), and 0.82 (range 0.66–0.93) for physician–AI comparison. The median HD was 10.75 mm (range 5.02–28.13 mm) for physician–APRT and 29.97 mm (range 11.09–58.80 mm) for physician– AI. During final physician review, 61% of APRT contours were accepted without modification and 39% required minor adjustments; none were classified as clinically significant. For AI-generated contours, 11% were approved without changes, 61% required minor corrections, and 28% required major corrections. Median contouring times were 37 minutes (range 29–45) for physicians, 53 minutes (range 42–69) for APRT manual contouring, and 17 minutes (range 5– 23) for the AI–APRT workflow. Dosimetric analysis showed no significant differences between physician and APRT contours, while deviations up to 29.1% in Dmax were observed in AI-only contours. Conclusion: The AI–APRT hybrid model provides accurate, safe, and time-efficient OAR contouring in lung SBRT. APRTs can reproduce physician-level quality, and AI support significantly reduces workload and contouring time (54% less time than physicians), supporting sustainable and high-quality radiotherapy practice. The main limitations of our study include the small sample size, single-site setting, retrospective design, and use of a single AI platform. References: 1.Trimpl M., Campbell S., Panakis N. et al. Deep learning-assisted interactive contouring of lung cancer: Impact on contouring time and consistency. Radiotherapy and Oncology 200 (2024) 1105002.Meyer
References: 1. Oar A, Lee M, Le H, Wilson K, Aiken C, Chantrill L, et al. AGITG MASTERPLAN: a randomised phase II study of modified FOLFIRINOX alone or in combination with stereotactic body radiotherapy for patients with high- risk and locally advanced pancreatic cancer. BMC Cancer. 2021;21(1):936.2. Oar A, Lee M, Le H, Hruby G, Dalfsen R, Pryor D, et al. Australasian Gastrointestinal Trials Group (AGITG) and Trans-Tasman Radiation Oncology Group (TROG) Guidelines for Pancreatic Stereotactic Body Radiation Therapy (SBRT). Pract Radiat Oncol. 2020;10(3):e136-e46. Keywords: quality assurance, pancreas, SBRT Digital Poster Highlight 1696 AI–APRT Hybrid Model for OAR Contouring in Lung SBRT: Accuracy, Safety, and Efficiency Michele Sala 1 , Suela Vukcaj 2 , Giulia Cipriano 3 , Giuseppe Ronci 4 , Davide Cicolari 4 , Sara Molli 2 , Fabio Piccoli 2 , Francesco Filippone 2 , Roberta Muni 2 , Micaela Motta 2 , Laura Maffioletti 2 , Angelo Di Naro 2 , Gian Luca Poli 4 , Maurizio Portaluri 2 1 Radiology, ASST PAPA GIOVANNI XXIII, Bergamo, Italy. 2 Radiotherapy, ASST PAPA GIOVANNI XXIII, Bergamo, Italy. 3 Radiotherapy, Milano-Bicocca University, Milano, Italy. 4 Medical Physics, ASST PAPA GIOVANNI XXIII, Bergamo, Italy Purpose/Objective: Accurate contouring of organs at risk (OARs) is a critical determinant of radiotherapy quality and patient safety. Manual delineation by physicians remains the clinical reference standard but is time- consuming and subject to inter- and intra-observer variability. This study investigates whether an Advanced Practice Radiation Therapist (APRT), with
Made with FlippingBook - Share PDF online