Impact of PET/CT imaging on target volume delineation and radiotherapy planning in patients with locally advanced esophageal cancer

Abstract Background and purpose Combined chemoradiotherapy is a key treatment strategy for esophageal cancer in both neoadjuvant and definitive settings. Accurate and consistent target volume delineation is essential to optimize therapeutic efficacy and minimize toxicity. 18 F-fluorodeoxyglucose ( 18 F-FDG) positron emission tomography/computed tomography (PET/CT) enables visualization of metabolically active tumor regions and may improve the accuracy and reproducibility of target volume definition. This retrospective study aimed to evaluate the impact of PET/CT-guided contouring on target volume size, target volume geometry, organ-at-risk exposure, and interobserver variability compared with conventional CT-based delineation. Materials and methods Twenty patients treated between 2022 and 2024 at the Department of Radiation Oncology, University Hospital Magdeburg, were included. Gross tumor volumes (GTVs) were independently delineated on CT and PET/CT datasets. Clinical target volumes (CTVs) and planning target volumes (PTVs) were subsequently generated. Geometric agreement was assessed using the Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (HD95). Interobserver variability was evaluated using simultaneous truth and performance level estimation (STAPLE)-generated reference contours. Paired t‑tests (α = 0.05) were used to assess volumetric and dosimetric differences. Results PET/CT-based planning significantly reduced PTV volumes, with mean PTV decreasing from 602.3 ± 267.4 cm 3 to 453.2 ± 237.4 cm 3 ( p < 0.001). In four patients (20%), more than 100 cm 3 of the PET-defined target volume was located outside the original CT-based PTV, indicating clinically relevant geometric modification of irradiation fields. Eight patients (40%) showed HD95 values exceeding 25 mm between CT- and PET/CT-based target volumes. PET/CT-based planning also reduced pulmonary radiation exposure, with lung V20 decreasing from 10.39 ± 1.23% to 7.85 ± 1.18% ( p = 0.001) and lung V30 from 4.41 ± 0.76% to 2.46 ± 0.37% ( p = 0.003). No significant reduction in cardiac or left anterior descending coronary artery (LAD) exposure was observed. Interobserver analysis demonstrated comparable overall agreement between CT- and PET/CT-based contouring, with lower HD95 variability for PET/CT-based delineation. Conclusion PET/CT-guided radiotherapy planning substantially influences target volume geometry in esophageal cancer and enables identification of metabolically active disease not adequately covered by CT-based planning alone. In addition to significant target volume reduction and lower pulmonary dose exposure, PET/CT-based contouring may support more consistent target volume delineation. Prospective studies are warranted to determine whether these geometric and dosimetric advantages translate into improved oncologic outcomes and reduced treatment-related toxicity.