RESEARCH COMMUNICATIONS OF THE 28th ECVIM-CA CONGRESS

Performance of the summary score was assessed using an anonymised validation dataset of 119 suspect HAC cases from UCD Veterinary Hospital classified as HAC (n=58) or NAI (n=61). A diagnosis of HAC was supported by a documented improvement with treatment (n=42), post‐mortem examination (n=3) and independent expert review of the medical records (n=13). A diagnosis of NAI was supported by an alternative diagnosis (n=55), spontaneous resolution of clinical signs (n=3) and absence of progression or onset of additional clinical signs over a follow‐up period > 1 year (n=3). The sensitivity of the score ranged from 53 to 66%, depending on whether equivocal results were treated as positive or negative for HAC. Specificity was 93%. Area under ROCC was 0.926 indicating that cut‐off adjustment could improve diagnostic sensitivity. Four false negative scores were complex cases where HAC was diagnosed in the face of a negative LDDST adversely affecting sensitivity.