MRI Criteria Standardize Healing Assessment in Perianal Crohn’s Fistulas, Expert Review Finds

Perianal fistulising Crohn’s disease affects up to one-third of patients with Crohn’s disease during their lifetime, producing pain and discharge that can substantially diminish quality of life. Assessing whether treatment is actually working has long depended on a simple clinical test: a clinician applies gentle finger compression and checks whether pus drains from the external opening. Absence of drainage counts as remission. Yet fistulas whose external openings have closed can still harbor an active internal tract, a state without deep healing that causes anal pain and goes unnoticed at the bedside. Magnetic resonance imaging, by contrast, easily reveals such hidden activity, and a new review published in European Radiology Abdomen lays out how MRI should be used to monitor treatment response in this condition, drawing on recently established consensus criteria from a global expert group.

The review, authored by T. J. van Croonenburg, C. C. Huisman, K. Horsthuis and J. Stoker, notes that MRI is the reference standard for evaluating perianal fistulising Crohn’s disease and has become essential in monitoring treatment response. Until recently, however, response monitoring was subjective and heterogeneous among radiologists, with no clear guidance or radiological criteria. The prior best guideline came from an ECCO-ESGAR topical review, which limited activity assessment to three features that could be interpreted ambiguously. To address this gap, a global expert consensus group called TOpCLASS, comprising 84 experts including 32 radiologists, 30 gastroenterologists and 22 surgeons, established consensus definitions for healing and healed fistulas. The review argues these definitions provide a framework that represents an important step toward greater consistency in clinical practice and facilitates communication between radiologists and clinicians globally.

On MRI, active perianal fistulas appear as hyperintense tracts on T2-weighted and fat-saturated T2-weighted sequences, with the bright signal representing fluid, pus or granulation tissue. A post-contrast fat-saturated T1-weighted sequence with gadolinium distinguishes between them: a T2-hyperintense tract that shows complete enhancement contains granulation tissue, while a central region without enhancement indicates fluid or pus within the tract. The clinical relevance of making this distinction depends on whether it influences management decisions. Subtle active disease, such as in tracts only a few millimeters in diameter or in the final phase of healing, may be difficult to distinguish from a completely fibrotic tract on T2-weighted sequences alone, and post-contrast images help separate subtle active disease from fibrotic disease.

According to the TOpCLASS criteria, radiological healing is defined by the presence of at least one criterion on follow-up MRI compared with baseline. These criteria include reduction in T1- and/or T2-hyperintensity, diameter, length, decrease or complete absence of previously present abscesses, and increasingly fibrotic tracts. The consensus group stated that reduction expressed in percentages cannot be supported by literature, so terms like increasingly and unequivocal reduction are used instead. Radiologists are advised to state in their reports which criteria showed unequivocal change, and the use of the term unequivocal enforces a higher threshold for interpretation, improving reproducibility and consistency. The review cautions, however, that the absence of a precise threshold for reduction still leads to subjective assessment, and further research is required to define what constitutes unequivocal change.

Each criterion carries interpretive subtleties. Unequivocal reduction in T2-hyperintensity, and T1-hyperintensity when contrast is used, may be the earliest sign of improvement, but caution is warranted to ensure the apparent decrease reflects a true change rather than a windowing effect; if decreased signal is the only notable improvement, an additional follow-up MRI is recommended to confirm further change over time. Reductions in diameter and length must be measured only in the active, hyperintense portion of the tract, excluding completely hypointense fibrotic segments, and the diameter change must be substantial considering the total tract volume. An increasingly fibrotic tract is part of the healing process, but in the case of only increased fibrosis with persistent hyperintense areas, a consecutive MRI is needed to determine whether additional signs of healing emerge or whether the fibrosis represents a non-responding chronic fistula. Inflammation in fat surrounding the tract, though frequently observed, was deliberately excluded from the new definition because its incremental value is considered limited and its assessment relatively subjective.

A radiologically healed fistula is defined more strictly: the absence of hyperintensity on all sequences, reflecting a completely fibrotic tract. Through a Delphi consensus process, the TOpCLASS group established that this means absence of T2-hyperintensity with a completely fibrotic tract and, if available, absence of contrast enhancement on the post-contrast fat-saturated T1-weighted sequence. Some fistulas leave little to no residual fibrosis. A fistula cannot be considered healed while a seton, the surgical thread placed through the tract to optimize drainage, remains in place, since the seton precludes complete healing; it must be removed and time allowed for healing before a healed fistula can be diagnosed on MRI.

The healed definition has been put to the test. In an international multi-centre study of 40 patients with perianal fistulising Crohn’s disease from three tertiary referral centers, all prospectively diagnosed in local MRI reports as healed before the TOpCLASS definition was published, three experienced gastrointestinal radiologists reviewed the scans using the new criteria. Only 14 patients, or 35 percent, met the TOpCLASS healed definition, while 26, or 65 percent, did not. Of the 14 who met the criteria, 13 achieved sustained clinical remission. The definition showed excellent inter-rater reliability, with Cohen’s kappa values of 0.89 to 0.95 and total agreement of 95.0 to 97.5 percent. The study’s authors note the tertiary care setting produced a cohort with highly complex disease, which may account for the low rate of observed radiologically healed fistulas, and the findings highlight the need to inform radiologists about the definition to prevent recurrent disease.

The stakes of strict adherence are illustrated by a case from that validation study: one fistula initially considered radiologically healed was found on retrospective review to have a very subtle residual hyperintensity in the intersphincteric space, a finding most likely the cause of the patient’s recurrence. The review’s authors strongly advise thoroughly evaluating the MRI for any subtle residual hyperintense areas before declaring a tract healed, since these are associated with the potential risk of fistula recurrence, and adhering strictly to the criterion of a completely hypointense signal on all sequences. The PISA-II study underscores why deep healing matters: recurrence occurred in 36 percent of patients, 27 out of 76, who achieved clinical closure alone, whereas no recurrences, 0 out of 25, were observed in patients with radiological healing.

Timing is another crucial element. Studies on medical treatment show little radiological change in the initial weeks up to 6 to 12 weeks, so early imaging may not accurately reflect treatment response. Short-term radiological follow-up after surgery may reveal blood, oedema or granulation tissue related to surgical trauma, appearing hyperintense on T2-weighted and post-contrast T1-weighted sequences, along with gas or susceptibility artefacts that appear as hypointense spots. Delaying postoperative MRI to a minimum of 12 weeks gives a better overview since most postoperative changes will have resolved, though earlier scans remain useful for evaluating abscess drainage or new fistula tracts. One study showed that radiological improvement observed at 6 months following anti-TNF treatment persisted through 12 to 18 months of follow-up, and literature indicates radiological healing lags 12 months behind clinical remission in patients treated with anti-TNF combined with thiopurine. The TOpCLASS consensus therefore recommends a range of 6 to 12 months after definitive surgery or medical therapy, with 6 months more appropriate for treatment evaluations and 12 months for evaluating fistula closure.

Beyond the consensus criteria, the review surveys quantitative tools under investigation. Established composite scores such as the modified van Assche index and the MAGNIFI-CD index, which yields scores from 0 to 25 with external validation establishing cut-offs of a decrease of at least 2 points for treatment response and 6 or fewer points for remission, are considered time-consuming and impractical for routine clinical use, though the TOpCLASS group suggested them for research and clinical trials. A simpler Fibrosis score categorizing tracts from 0 to 100 percent fibrosis is potentially suitable for routine practice but has not yet been formally validated. A retrospective study of 52 patients found an association between a higher MAGNIFI-CD index at baseline and unfavorable long-term clinical outcome, and a high degree of fibrosis above 80 percent combined with a low MAGNIFI-CD index of 13 or less showed predictive value for better clinical outcome, though no significant association was found between individual fibrosis categories and outcome assessed separately. Three-dimensional volumetric measurement of the active fistula on fat-saturated T2-weighted sequences has been shown feasible and reliable, and some studies suggest MRI-based fistula volume may serve as a potential biomarker for monitoring response after medical and surgical interventions, but validation in larger prospective cohorts is warranted before routine clinical implementation. Emerging techniques including dynamic contrast enhancement, diffusion-weighted imaging and T2* mapping require advanced sequences and protocol adjustments while the evidence remains limited, which does not justify widespread implementation currently. The review concludes that the healing and healed criteria offer a practical framework for daily practice, with the healed criteria robust and associated with sustained clinical remission, while the healing criteria remain subject to further research and possible modification in coming years.

Subject of Research: MRI-based response monitoring criteria for perianal fistulising Crohn's disease

Article Title: Response monitoring of perianal fistula in Crohn’s disease on MRI

Article References: van Croonenburg, T. J., Huisman, C. C., Horsthuis, K., & Stoker, J. (2026). Response monitoring of perianal fistula in Crohn’s disease on MRI. European Radiology Abdomen, 1(1), Article 3. https://doi.org/10.1007/s44501-026-00004-6

Image Credits: AI Generated

DOI: 10.1007/s44501-026-00004-6

Keywords: Crohn's disease, perianal fistula, MRI, TOpCLASS consensus, radiological healing, treatment response, fibrosis, 3D volumetry, MAGNIFI-CD, deep healing, gastrointestinal radiology, Response