P2.05-032 CT-Based Surrogates of Pulmonary Ventilation in Lung Cancer: A Voxel-Level Comparison with HP Gas MRI: Topic: Multimodality Treatment

Abstract

Image registration of paired inspiratory & expiratory CT is a potential method for generating surrogates of regional ventilation by assuming that local lung expansion & density changes of corresponding parenchymal voxels equate to ventilation. Potential lung cancer applications include functional lung avoidance radiotherapy planning and longitudinal assessment of treatment response. However, the physiological accuracy of the technique has yet to be validated. The aim of this study was to compare the spatial correlation of ventilation CT & 3He MRI in a cohort of lung cancer patients. 5 patients underwent expiration & inspiration breath-hold CT. 3He & 1H MRI were also acquired in the same breath and at the same inflation state as inspiratory CT. Expiration CT was deformably registered to inspiration CT for calculation of ventilation CT from voxel-wise differences in Hounsfield units. Registration accuracy was validated using a reference CT dataset for 6 patients with 100 expert anatomical landmarks defined on both images. Inspiration CT was registered to 3He MRI via the same-breath anatomical 1H MRI to enable direct comparison of 3He MRI & CT ventilation in corresponding regions of interest located within the lungs as defined by a 1H MRI lung mask (see figure). Spearman’s rank coefficients were used to assess voxel-level correlation. The mean registration error for the reference dataset was 1.1±0.2mm (mean±SD). Successful registration enabled computation of ventilation CT images at the inspiratory state and direct comparison of ventilation CT with MRI. The median (range) Spearman’s coefficient was 0.68 (0.45-0.76). This work demonstrates a method of acquiring CT & 3He MRI in similar breath-holds to enable direct spatial comparison of ventilation maps. Initial results show moderate correlation between ventilation CT & 3He MRI. Further large-scale clinical trials are required before clinical implementation of the technology.