AUTHOR=Chen Zhi , Huang Yu-Hua , Kong Feng-Ming , Ho Wai Yin , Ren Ge , Cai Jing 

TITLE=A super-voxel-based method for generating surrogate lung ventilation images from CT

JOURNAL=Frontiers in Physiology

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2023.1085158

DOI=10.3389/fphys.2023.1085158

ISSN=1664-042X

ABSTRACT=Purpose: This study aims to develop and evaluate a novel super-voxel based lung ventilation imaging method CTVI_SVD for robust lung ventilation imaging from CT images. Methods and Materials: 21 lung cancer patients with 4DCT, single-photon emission computed tomography (SPECT), and corresponding lung masks from the VAMPIRE challenge were included in this study. For each patient, the lung volume of the exhale CT was segmented into hundreds of super-voxels by using the Simple Linear Iterative Clustering (SLIC) method. These super-voxel segments were applied on the CT and SPECT image correspondingly to calculate the mean density values (Dmean) of CT and mean ventilation values (Ventmean) of SPECT. The final CTVI_SVD was generated by interpolation from Dmean. For the performance evaluation, the Spearman correlation and Dice Similarity Coefficient index were used to compare the voxel-wise and region-wise differences between CTVI_SVD with SPECT. As a comparison, two DIR-based methods CTVI_HU and CTVI_Jac were also generated and compared to the SPECT. Results: The correlation between the Dmean and the Ventmean of the super-voxel was 0.59 ± 0.09, showing a moderate-to-high correlation between the mean CT density values and mean SPECT ventilation values in the super-voxel level. For the voxel-wise evaluation, the proposed CTVI_SVD method achieved a strong average correlation of 0.62 ± 0.10 with SPECT, which is significantly better than the CTVI_HU (0.33 ± 0.14, p < 0.05) and CTVI_Jac (0.23 ± 0.11, p < 0.05) methods. For the region-wise evaluation, the Dice Similarity Coefficient of high functional region (〖DSC〗_h) for CTVI_SVD was 0.63 ± 0.07, which is significantly higher than CTVI_HU (0.43 ± 0.08, p < 0.05) and CTVI_Jac (0.42 ± 0.05, p < 0.05) methods. Conclusion: The strong correlations between CTVI_SVD with SPECT demonstrated that this new ventilation estimation method holds the potential for ventilation imaging.