S793
Clinical - Lung
ESTRO 2026
Ane Mugica 1 , Maider Campo 1 , Amaya Sanchez 1 , Leyre Gonzalez 1 , Elena Guimón 1 , Intza Uranga 1 , Sara Palacios 1 , Beraldo Martinez 1 , Xabier Gurutzeaga 1 , Ane Otaegi 1 , Ane Dehesa 1 , Sebastian Luevano 1 , Ainhoa Diez 1 , Jose María Urraca 1 , Arrate Querejeta 1 1 Radiation Oncology, Onkologikoa – UGC Oncología, Gipuzkoa, Spain. 2 Faculty of Engineering, Mondragon University, Gipuzkoa, Spain. 3 Radiology, Hospital Universitario Donostia, Gipuzkoa, Spain Purpose/Objective: To assess whether radiological patterns on 3–6 month post-SBRT computed tomography (CT) scans, combined with dosimetric and clinical variables, can identify patients at risk of local or locoregional Post-treatment images from 92 patients with a solitary pulmonary nodule, either histologically confirmed or highly suspicious for carcinoma, treated with SBRT using TomoTherapy between 2013 and 2019 were retrospectively analyzed. A radiologist and a radiation oncologist classified the radiological evolution of the treated area according to patterns described in the literature: diffuse ground-glass (DGG), patchy ground- glass (PGG), patchy consolidation and ground-glass (PCGG), masslike (ML), scarlike (SL), modified recurrence (LR/LRR). Material/Methods: conventional (MC), diffuse consolidation pattern (DC) and no change (NC). The radiological pattern at 3–6 month (or the latest available if recurrence occurred before 3 months) was selected and coded. Clinical variables (age at treatment, sex, previous lung disease, tumor location, diagnostic PET SUVmax) and dosimetric variables (PTV, total dose, fraction dose, mean lung dose (MLD), V5 and V20 total-lung and V20 ipsilateral-lung) were included. After correlation analysis of dosimetric variables, fraction dose and MLD were selected. A predictive model was developed using a classification tree (CART) with internal 5-fold cross-validation. This method involves creating a hierarchical decision structure that iteratively splits data by the most informative variables to predict the target outcome. The dependent variable was LR/LRR. Model stability and variable importance were evaluated, and an AUC/ROC analysis performed. Results: Among 92 patients, 16 (17.4 %) developed LR/LRR, while 76 (82.6 %) remained progression-free. The final model incorporated four variables: PTV, 3–6 month post-SBRT radiological pattern, age at treatment, and mean lung dose. PTV and 3–6 month post-SBRT radiological pattern, were the most influential variables in the classification tree (100 % and 93.9 %, respectively). The radiological patterns DGG, PCGG and ML, as grouped by the tree structure, were associated with higher recurrence risk, whereas NC, PGG, DC, MC, and SL were associated with favorable
Conclusion: Single-fraction SBRT provided excellent local control and durable survival with minimal toxicity in the curative treatment of NSCLC. These real-world results, from one of the largest European cohorts treated with this schedule, demonstrate that this regimen offers both strong oncological outcomes and major logistical advantages by reducing treatment time and optimising resource use. (2) Given its safety and efficacy, this approach warrants prospective evaluation in phase III trials. References: 1. Videtic GM, Paulus R, Singh AK, Chang JY, Parker W, Olivier KR, et al. Long-term Follow-up on NRG Oncology RTOG 0915 (NCCTG N0927): A Randomized Phase 2 Study Comparing 2 Stereotactic Body Radiation Therapy Schedules for Medically Inoperable Patients With Stage I Peripheral Non-Small Cell Lung Cancer. Int J Radiat Oncol Biol Phys. 2019;103(5):1077-84.2. Tas KT, Lishewski P, Sheikhzadeh FF, Smalec E, Recknagel N, Wündisch T, et al. From protocol to practice: long-Term outcomes of single-Fraction stereotactic body radiotherapy for primary non-Small cell lung cancer. Strahlenther Onkol. 2025. Keywords: Single-fraction SBRT, Real-world outcomes Digital Poster 2528 Early post-SBRT CT patterns and dosimetric factors predict recurrence in solitary lung nodules María Pagola 1 , Josune Urien 2 , Ainhoa Galardi 3 , Julian Mínguez 1 , Alai Goñi 1 , Eva Saenz de Urturi 1 , Usoa Iceta 1 , Nuria Bultó 1 , Mikel Eguiguren 1 , Daniel Alberto Roura 1 ,
Made with FlippingBook - Share PDF online