Background:Voiding cystourethrogram (VCUG) has traditionally been used to evaluate pediatric urethral pathology. Contrast enhanced voiding ultrasonography (ceVUS) has been described as an alternative method to evaluating the urethra but the validity is yet to be determined. Our study aims to compare differences in interpretation between VCUG and ceVUS images.Methods:Children who had undergone both ceVUS and VCUG with urethral imaging were considered for the study. Time between studies was recorded. Imaging studies spaced more than 24 months apart and patients who underwent urethral surgeries were excluded. Video screen captures excluding identifying information were recorded for both ceVUS and VCUG imaging. The videos were unlinked between patients and randomized. The randomized videos were interpreted independently and compared by 2 pediatric urologists. Interrater reliability was assessed using Cohen’s kappa (k) as a measure of agreement.Results:Thirty-nine patients met study criteria. Fifteen patients (38%) underwent urodynamics in conjunction with both imaging modalities. See Table 1 for patient demographics. Twenty-one patients (54%) had abnormal urethral pathology on one of their studies as deemed by at least one provider. For each patient, Urologist #1 saw that 28 (72%) of the 2 paired studies were congruent based on blinded interpretation (k=0.35), while Urologist #2 saw that 23 (59%) of the 2 studies were congruent (k=0.003). Only 1 patient had the 2 studies performed on the same day for which radiology read a posterior urethral valve on the ceVUS study and normal on VCUG, while the 2 urologists read all studies as normal. Between the 2 urologists, 30 (77%) of ceVUS studies and 33 (85%) VCUG studies were read the same. The agreement for ceVUS was slight with k=0.093, and the agreement for VCUG was substantial with k=0.620. Conclusions:VCUG appears to have better interrater reliability compared to ceVUS. In addition, the interpretation between the two studies were incongruent in 30-40% of cases which could be due to changes in pathology due to time lapse vs truly different appearance based on study modality. Clearly, a larger patient cohort is needed to support this preliminary study.
Table 1: Patient Demographics
| Characteristics (n=39) | |
| Gender:MaleFemale | 16 (41.0%)23 (59.0%) |
| Mean (SD) Age at Study (years):ceVUSVCUG | 3.3 (2.8)2.0 (2.7) |
| Condition:Vesicoureteral reflux concernsNeurogenic conditions | 24 (61.5%) 15 (38.5%) |
