Comparison of sedated respiratory-gated computed tomography to anesthetized inspiratory-expiratory breath-hold computed tomography in dogs with respiratory disease

Objective To determine the diagnostic accuracy of sedated respiratory-gated (RG) CT as a minimally invasive surrogate for anesthetized, ventilator-assisted inspiratory-expiratory breath-hold (I:E-BH) CT scans by being able to identify CT lung patterns and subpatterns, abnormal and normal lung attenuation scores, and bronchomalacia (BM) in dogs with respiratory disease. Methods Sedated RG CT and anesthetized I:E-BH CT images were sequentially acquired in 50 client-owned dogs with respiratory clinical signs. Computed tomography lung patterns and subpatterns, 2 CT severity scores, and the presence of BM were assessed. Agreement was estimated using unweighted and weighted Cohen κ coefficients and Bland-Altman plots. McNemar tests and mixed-effects logistic or ordinal logistic regressions were utilized to evaluate discordant pairs between CT techniques. Results Both scan types had minimal motion artifact. Nodular pattern and airspace nodules and reticular and subpleural interstitial thickening subpatterns were significantly underestimated in RG CT compared to I:E-BH CT. A fair agreement was estimated between the 2 methods for normal lung scores but not for abnormal lung scores, and for BM. There was a significant difference between the two techniques in diagnosing clinical BM. Inspiratory-expiratory breath-hold CT yielded more BM diagnoses, especially in younger dogs with milder BM scores. Conclusions In dogs with respiratory disease, RG CT had good concordance with I:E-BH CT on most major lung patterns but not some nodular and linear subpatterns, underdetecting subtle but clinically relevant lesions. The probability of diagnosing clinical BM is higher with I:E-BH CT, particularly in milder cases. Clinical Relevance Compared to RG CT, anesthetized I:E-BH CT provides superior technique and detail that allows for characterization of some subpatterns, which could have diagnostic implications.