Receiver operating characteristics of computed tomography (CT) compared to cystoscopy in diagnosis of canine ectopic...

The aim of the study was to determine receiver operating characteristics (ROC) of computed tomographic excretory urography (CTEU) in predicting cystoscopic findings of ureteral anatomy.

Retrospective cohort study.

Thirty-five client-owned dogs.

The medical records of dogs referred for suspected ectopic ureters were reviewed. Inclusion criteria included CTEU findings reported by board-certified radiologists, followed by rigid cystoscopy with or without ureteral cystoscopic laser ablation (CLA). Data included signalment, urinary incontinence degree, body condition score, weight, degree of colon distension, CTEU and cystoscopy findings. ROC analysis was used to compare CT-predicted ureteral orthotopia/ectopia to cystoscopy findings. Additionally, ROC of CT predicted ureteral orifice locations was analyzed. Regression covariate analysis was performed to identify factors that may have influenced accuracy of diagnosis.

The ability of CT to identify a normal and intra-or extramural ectopic ureters conclusively and correctly was 13/26 (50%) and 32/41(78%), respectively. Sensitivity and specificity of identifying extramural versus intramural ureters was 2/7 versus 30/46 (29 vs. 65%) and 61/63 versus 17/24 (97 vs. 71%), respectively. Ectopic orifice determination sensitivity and specificity varied widely depending on location from 0% to 76% and 67% to 97%, respectively. Covariate analysis failed to identify interfering factors.

CT did not accurately predict anatomy of ureters; CT findings may need confirmation by cystoscopy and possibly intraoperative fluoroscopy prior to determining if CLA is indicated or not.

Our results may be of importance for surgeons interpreting the CTEU findings. CTEU prediction of the location of the ureteral orifice shows low sensitivity especially in or close to the urethral sphincter area.