Three radiographic methods for assessing left atrial enlargement in dogs: diagnostic performance and correlation with qualitative features.

INTRODUCTION/

OBJECTIVES

Left atrial (LA) enlargement is a critical indicator of cardiac disease progression in dogs. This study compared the diagnostic accuracy of vertebral heart size (VHS), vertebral left atrial size (VLAS), and radiographic left atrial dimension (RLAD) in detecting ≥moderate LA enlargement using an ordinal echocardiographic reference standard and investigated their correlations with qualitative radiographic features.

ANIMALS, MATERIALS AND

METHODS

This retrospective study included 177 dogs. Quantitative radiographic measurements (VHS, VLAS, and RLAD) and qualitative radiographic features (caudodorsal cardiac margin, left auricular bulge, and double wall sign) were evaluated. Left atrial size was classified using established left atrial-to-aortic root ratio (LA:Ao) thresholds (normal <1.6; mild 1.6-1.9; moderate 1.9-2.3; severe ≥2.3). Analyses focused on distinguishing ≥moderate LA enlargement using receiver operating characteristic analysis. Correlations with qualitative features were examined using Spearman correlation.

RESULTS

All radiographic measurements showed fair positive correlations with LA:Ao. For detection of ≥moderate LA enlargement, LA-specific measurements demonstrated good diagnostic performance than the fair diagnostic performance of VHS. Optimal cutoffs were VHS ≥10.9, VLAS ≥2.5, and RLAD ≥2.1. Both VLAS and RLAD correlated more strongly with caudodorsal cardiac margin morphology than did VHS.

STUDY LIMITATIONS: The RLAD was unmeasurable in 39.8% of cases, and demonstrated lower interobserver agreement than VHS and VLAS.

CONCLUSIONS

Measurements specific to the LA (VLAS and RLAD) demonstrated improved diagnostic performance for detecting ≥moderate LA enlargement and correlated more strongly with wedge-shaped caudodorsal margin than VHS. Vertebral left atrial size provided the optimal balance of performance, feasibility, and reliability for clinical use.