Purpose: Ventilation and perfusion (V/Q) imaging has been used extensively to determine initial lung function and change in lung function in response to an intervention such as radiotherapy or surgery. Typically nuclear medicine techniques have been used; however there has been a strong interest in deriving these images from computed tomography (CT) imaging. In this work, we present a protocol for acquiring V/Q surrogates from dual energy CT (DECT) with initial results from 10 baseline scans in a prospective clinical trial.
Methods: Patients received a 4D V/Q positron emission tomography (PET)/CT radiotherapy scan and a DECT scan on the same day. In the DECT session an inhale/exhale breath hold single energy CT was acquired, followed by a DECT image in inhale breath hold with arterial phase contrast. CT ventilation (CT-V) images were derived using multiple methods based on deformation and/or density changes between the breath-hold inhale/exhale images. A perfusion surrogate iodine map (DECT-Q) was generated based on the arterial phase DECT inhale breath hold images. The CT derived V/Q maps were compared to PET V/Q with a voxel-wise Spearman correlation.
Results: Spearman correlation for the ventilation images were similar to previous breath hold CT studies with coefficients, r_s=median (range), for density based r_s=0.55 (0.13-0.77), deformation based r_s=0.56 (-0.09-0.84) and combined density and deformation models r_s=0.55 (0.04-0.83). The Spearman correlation for the perfusion images resulted in r_s=0.44 (0.20-0.61), which is lower than what has previously been reported using region-based correlation with SPECT/CT.
Conclusion: Initial results of a dual-energy CT ventilation and perfusion imaging protocol show variable voxel-wise correlation when compared to 4D PET V/Q images. Further work is required to determine if the observed variation in voxel-wise correlation translates to variation in the clinical application of these images, in lobar or other regional functional evaluation.