Introduction The use of anti-TNF therapy for the treatment of fistulising perianal Crohn’s disease relies heavily on clinical and radiological assessments in order to assess response. We have shown that a third of patients sustain radiological remission at 3-years follow-up. It is difficult to predict response based on MRI images as they are difficult to interpret. Having a three-dimensional (3-D) fistula model and being able to quantify the changes in the fistula volume would be more beneficial than relying on complex and subjective MRI reports.

We aim to use computer software and MRI segmentation to design 3D models of Crohn’s anal fistulae, measure baseline and 3 years post-biologics treatment fistula volumes and compare with clinical and MRI healing.

Method Three-Dimensional fistula models were independently created by the surgeon and radiologist using the previously validated surgical computer software. Thirty-eight patients were selected and volumes were measured at baseline and 3-years post biological therapy.

Kappa statistics were used to assess correlation between the healing categories for the three assessment techniques: volume measurements, clinical and MRI healing. Binary logistical regression analysis was used to assess predictors of changes in 3-D anal fistula volumes.

Results A change in baseline and post-treatment volumes was observed. The three assessment tools were clinical healing, radiological healing and change in 3-D anal fistula volumes. The agreement between each pair of assessment techniques were measured using Kappa values. There was fair agreement between 3-D volumes and MRI healing (K = 0.34; 95% CI = 0.11, 0.57). Poor agreement was noted between 3-D volumes and clinical healing (K = 0.05; 95% CI = -0.15, 0.25) and the agreement was also poor for clinical and MRI healing (K = 0.10; 95% CI = -0.12, 0.32).

On binary logistical regressional analysis, the single independent factor predictive of 3-D volume change was the duration of the disease. An increased duration of disease was associated with a reduced likelihood of an improvement in volumes (OR = 0.57; 95% CI = 0.33, 0.9, p = 0.045). Five-year increase in the duration of disease was associated with two-fold decrease in the odds of improvement.

Conclusion MRI volume measurements of 3D Crohn’s fistula models provide an accurate assessment tool. The measurement of perianal fistulae volumes is more is useful than using MRI reports alone as it gives a quantitative measure that can be used to assess the efficacy of therapy. Serial fistula volume measurements should be used for monitoring the response to biological therapy for perianal Crohn’s fistula patients.

Disclosure of interest None Declared.