Abstract: Background: Paired inspiratory and expiratory breath-hold computed tomographic scans (I:E-BH CT) capture images with control of respiratory phase, important with dynamic airway collapse. CT attenuation histograms allow objective quantification of distribution of Hounsfield Units (HU).
Hypothesis/Objectives: We hypothesize that dogs with bronchomalacia (BM), will demonstrate significant differences in objective metrics of attenuation on I:E-BH CT reflecting changes in airway caliber and impaired downstream parenchymal aeration compared to dogs without BM (NoBM). We aimed to document objective metrics of percentage low and high attenuating areas (LAA% and HAA%) and mean lung attenuation (MLA) on I:E-BH CT in BM/NoBM using automated software.
Animals: Client-owned dogs with respiratory signs and BM (n=118) and NoBM (n=25).
Methods: This retrospective study utilized 3D Slicer software to assess CT attenuation. Analysis used a Spearman correlation and a 2-way ANOVA utilizing beta regression.
Results: Strong correlations were identified within 4 LAA and 4 HAA metrics; thus the representative LAA%-856 and HAA%-700 were selected. BM had a significantly lower LAA%-856 (19%) than NoBM (31%) (p=0.013), independent of phase (I/E). Groups differed for HAA%-700 with accentuation in expiration versus inspiration (p< 0.001). On expiration, BM versus NoBM had 61% and 35% of lung <-700HU. MLA between groups was phase dependent with BM versus NoBM on inspiration being -708HU and -776HU and on expiration, -567HU and -686HU, respectively (p< 0.0001).
Conclusion and clinical importance: Higher lung attenuation in BM dogs supports downstream parenchymal aeration to segmental/subsegmental airway collapse. Quantitative imaging holds promise for objectively evaluating changes with BM.
