A Novel Method for Generating a Fused Ventilation–Perfusion Map Using CT‐Derived Ventilation and FDG‐PET for Lung Functional Avoidance in Radiotherapy: A Feasibility Study

Abstract Background Pulmonary ventilation–perfusion function plays a crucial role in both radiotherapy planning and prognosis assessment in patients with lung cancer. However, there remains a lack of rapid and cost‐effective imaging modalities capable of accurately capturing this functional parameter. Purpose This study aimed to develop a lung functional image that reflects ventilation–perfusion characteristics by integrating computed tomography (CT) and (positron emission tomography) PET imaging techniques, with the objective of enhancing lung dose evaluation in radiotherapy planning. Approach A retrospective analysis was performed on twenty lung cancer patients using CT images acquired at two respiratory phases and FDG‐PET/CT images. The Elastic Distortion algorithm was applied for deformable image registration, with the end‐expiration phase CT serving as the baseline. Values derived from the determinant Jacobian matrices of ventilation CT (V‐CT) images and the gray‐value matrices of PET images were normalized to a range of 0 to 1. These normalized values were multiplied to generate a ventilation–perfusion matrix. Three types of lung functional images were produced from these matrices: ventilation‐imaging (V‐imaging), perfusion‐imaging (P‐imaging), and ventilation–perfusion‐ imaging (VP‐imaging). The Dice Similarity Coefficient (DSC) and Bland–Altman plots were used to assess the correlations and discrepancies among the imaging modalities. Results The DSC values for the entire lung, regions with low 30% functionality, and regions with high 40% functionality were 0.39 ± 0.05, 0.50 ± 0.03, and 0.20 ± 0.05 for V–P; 0.58 ± 0.03, 0.73 ± 0.03, and 0.32 ± 0.02 for V–VP; and 0.68 ± 0.04, 0.78 ± 0.04, and 0.34 ± 0.04 for P–VP, respectively. Notably, significant concordance was observed between V–VP and P–VP images within the delineated functional lung regions.#x02013;Altman analysis supported the DSC results, revealing high correlation coefficients in the low 30% functional lung region: 0.628 for V–P, 0.857 for V–VP, and 0.779 for P–VP. In contrast, similarity within the high 40% functional regions was markedly lower. Conclusion This study developed a novel method for generating a fused VP map by integrating CT‐derived ventilation and FDG‐PET data. The method demonstrated feasibility, and the resulting VP map provided a balanced representation of both ventilation and perfusion signals, particularly in regions with reduced lung function.