Three-dimensional endoanal ultrasound is accurate and reproducible in determining type and height of anal fistulas

Aim Surgical treatment of high anal fistulas is associated with the potential risk of faecal incontinence and recurrence. The primary aim of this study was to determine the accuracy of three-dimensional endoanal ultrasound (3D-EAUS) in the assessment of height and type of anal fistulas, compared to the intra-operative findings (gold standard). The secondary aim was to evaluate the inter-observer reproducibility of 3D-EAUS. Method The study design was a prospective analysis of retrospective data. 299 patients (202 men), mean age 45.3 years, who underwent surgery for anal fistulas, were included. All patients were preoperatively assessed by 3D-EAUS. Two readers independently reviewed the volumes to determine the type and height of fistulas. Sensitivity, specificity, positive and negative predictive values, proportion of agreements and Cohen's kappa coefficient (k) were calculated for both examiners. Ultrasound findings were compared with intra-operative data (reference standard), evaluated blindly by the surgeons. Results At surgery, 201 (67%) were transsphincteric, 49 (16%) suprasphincteric, 47 ( 16%) intersphincteric and two (1%) extrasphincteric fistulas. Intra-operatively, 177 (59%) were low and 122 (41%) high fistulas. The overall accuracy of 3D-EAUS was 91% for fistula type (271/299 fistulas: 97% transsphincteric, 100% intersphincteric, 57% suprasphincteric, 0% extrasphincteric) and 92% for fistula height (275/299 fistulas: 80% high and 100% low). Both readers reported very good agreement with surgery in the assessment of fistula type (proportion of agreement 0.88, k = 0.89) and height (proportion of agreement 0.90, k = 0.91). Conclusions 3D-EAUS is an accurate and reproducible modality for the assessment of type and height of anal fistulas.