Voxel-level Ventilation-Perfusion Functional Imaging for Lung RT: Proof-of-Concept Technical Feasibility and Preliminary Dosimetric Evaluation

Abstract

INTRODUCTION: Existing lung imaging lacks the dual assessment and spatial resolution needed for functional-avoidance radiotherapy. This proof-of-concept (PoC) feasibility study aimed to validate the technical feasibility of a novel voxel-level lung functional imaging technique based on ventilation-perfusion (VP) mapping and explore its preliminary utility in radiotherapy treatment planning. MATERIALS AND METHODS: We retrospectively analyzed 20 lung cancer patients with 4D-CT and PET/CT scans. Ventilation was calculated from 4D-CT deformable image registration, and perfusion from pre-processed PET images. High-function lung regions were defined as the top 40%. Three functional imaging modalities were generated: ventilation (V), perfusion (P), and VP imaging. The Dice Similarity Coefficient (DSC) and Bland-Altman plots evaluated agreement, and intensity-modulated radiation therapy (IMRT) plans were optimized for dosimetric comparison. RESULTS: The automated workflow took 45±15 minutes per patient, meeting clinical efficiency requirements. V-VP and P-VP showed moderate-to-good agreement in whole and low-function lungs (DSC up to 0.71; r up to 0.943) but poor agreement in high-function regions. VP-IMRT significantly reduced ipsilateral high-function lung V20Gy (81.03±56.03 to 61.14±64.44, p<0.001), with favourable dosimetric differences versus single-modality plans and no increased dose to organs at risk. CONCLUSION: VP-Imaging effectively integrates ventilation and perfusion data, establishes a reproducible workflow, and shows preliminary dosimetric advantages. This study prioritizes technical viability over diagnostic accuracy relative to SPECT V/Q, supporting the feasibility of VP-Imaging for personalized RT.