Accuracy of contrast-enhanced voiding urosonography using Optison™ for diagnosis of vesicoureteral reflux in children.

ABSTRACT

BACKGROUND: There is limited quality of evidence regarding the accuracy of contrast-enhanced voiding urosonography (ceVUS) for diagnosis of vesicoureteral reflux (VUR) compared to fluoroscopic voiding cystourethrography (VCUG), and minimal data on the use of the ultrasound contrast agent Optison™ for this purpose. OBJECTIVE: To compare the accuracy of ceVUS using Optison™ to VCUG, and to assess inter-rater agreement regarding presence and grading of VUR. STUDY DESIGN: In this retrospective investigation, all sequential ceVUS with Optison™ and VCUG studies performed in children between 2014 and 2017 were reviewed. Two raters independently graded all ceVUS studies using a 5-point scale. CeVUS sensitivity and specificity were estimated separately for each rater using the VCUG report as the ground truth for presence and degree of VUR. Logistic and ordinary linear regression models assessed rater-report agreement and inter-rater agreement for each kidney, Optison™ dose, and referral diagnosis. RESULTS: 97 children (51 females) with 101 paired studies were included. Sensitivity and specificity of ceVUS for VUR detection were identical for both raters right kidney 75%/90.9%; left kidney 85.7%/78.9% (Figure). There was no statistically significant difference in disagreement between raters and the VCUG report for the right or left kidney. Inter-rater agreement on ceVUS grading was 90% and 88% for right and left kidneys, respectively. There was a significant negative association between fetal hydronephrosis vs urinary tract infection and disagreement between Rater 2 and the VCUG report for the left kidney. There were no other significant associations with respect to either kidney, Optison™ dose, or referral diagnosis. DISCUSSION: Our study showed that detection of VUR with ceVUS and Optison™ is comparable to fluoroscopic VCUG. Based on the VCUG reports, the incidence of VUR in our patient population was substantially lower than in the meta-analysis of Chua et al. and in the study of Kim et al. The explanation for the large discrepancy in VUR incidence may reflect differences in the patient populations, and in our reporting of VUR with respect to kidney number rather than to pelviureteral units. Study limitations include its retrospective nature and potential bias in terms of patient selection. Since VUR is an intermittent phenomenon, sequential rather than simultaneous performance of the ceVUS and fluoroscopic studies might have influenced VUR detection. CONCLUSION: A blinded comparison of ceVUS performed with Optison™ to fluoroscopic VCUG showed moderate-good sensitivity and specificity for diagnosis of VUR.