The thoracic radiographic unstructured interstitial pattern underestimates and may fail to identify canine respiratory disease compared to computed tomography.

Abstract

In dogs, lung disease presenting with a radiographic unstructured interstitial pattern (UnIP) poses a diagnostic challenge due to heterogenous clinical signs, non-specific differentials, and need for tissue sampling to confirm the pathologic process. The terminology describing patterns on thoracic radiography (TR) can be misleading in assuming an interstitial pattern implies disease of the pulmonary interstitium. Thoracic computed tomography (CT) is more likely to predict anatomic localization on a subgross level with robust evidence for CT patterns/subpatterns having corresponding histologic correlates in people. The study objective was to show that dogs with a UnIP on TR (1) have multiple CT patterns and subpatterns reflecting pathology beyond the interstitium and that (2) CT supports final definitive diagnoses encompassing more disorders than a UnIP on TR would imply. Thirty-six dogs with respiratory clinical signs, a sole UnIP on TR, thoracic CT, and additional tests to determine final diagnosis were retrospectively enrolled. Thoracic CT scans were assessed for presence or absence of four major CT patterns and 14 subpatterns. Final diagnoses were obtained by comprehensive evaluation of clinicopathologic abnormalities. Thoracic CT identified disease beyond the interstitium in all patients with a UnIP including large airway, small airway, and mixed airway/parenchyma disease. Mean (range) number of final diagnoses was 5 (1-13) with 33/36 (92 %) dogs having > 1 final diagnosis. Dynamic segmental/subsegmental airway collapse (i.e., bronchomalacia; 21/36, 58 %) was missed on TR. Despite the classic paradigm for radiographic UnIP corresponding to interstitial disease, CT provides more comprehensive anatomic correlates, expanding the differential list for respiratory disease.