RESUMO
OBJECTIVE: To compare the sensitivity and specificity of two- (2D) and three- (3D) dimensional transperineal ultrasound (TPUS) and 3D endovaginal ultrasound (EVUS) with the gold standard 3D endoanal ultrasound (EAUS) in detecting residual defects after primary repair of obstetric anal sphincter injuries (OASIS). METHODS: External (EAS) and internal (IAS) anal sphincters were evaluated by the four ultrasound modalities in women with repaired OASIS. 2D-TPUS was evaluated in real-time, whereas 3D-TPUS, 3D-EVUS and 3D-EAUS volumes were evaluated offline by six blinded readers. The presence/absence of any tear in EAS or IAS was recorded and defects were scored according to the Starck system. Sensitivity, specificity and predictive values were calculated, using 3D-EAUS as reference standard. Inter- and intraobserver analyses were performed for all 3D imaging modalities. Association between patients' symptoms (Wexner score) and ultrasound findings (Starck score) was calculated. RESULTS: Images from 55 patients were analyzed. Compared with findings on 3D-EAUS, the agreement for EAS evaluation was poor for 3D-EVUS (κ = 0.01), fair for 2D-TPUS (κ = 0.30) and good for 3D-TPUS (κ = 0.73). The agreement for IAS evaluation was moderate for both 3D-EVUS (κ = 0.41) and 2D-TPUS (κ = 0.52) and good for 3D-TPUS (κ = 0.66). Good intraobserver (3D-EAUS, κ = 0.73; 3D-TPUS, κ = 0.78) and interobserver (3D-EAUS, κ = 0.68; 3D-TPUS, κ = 0.60) agreement was reported. Significant association between Starck and Wexner scores was found only for 3D-EAUS (Spearman's rho = 0.277, P = 0.04). CONCLUSIONS: 2D-TPUS and 3D-EVUS are not accurate modalities for the assessment of anal sphincters after repair of OASIS. 3D-TPUS shows good agreement with the gold standard 3D-EAUS and a high sensitivity in detecting residual defects. It, thus, has potential as a screening tool after primary repair of OASIS. Copyright © 2016 ISUOG. Published by John Wiley & Sons Ltd.
