MRI-based standard may predict outcomes in perianal Crohn's

Share

Results from a small multicenter study suggest that a strict MRI-based definition of radiological healing may help predict long-term outcomes for patients with perianal fistulizing Crohn’s disease, a difficult-to-treat complication of inflammatory bowel disease.

For study known as VALIDATE-PERIANAL, presented by Sami Samaan, MD, a research fellow in the Deepak Lab and Research Group at Washington University School of Medicine in St. Louis, at the 2026 Crohn’s & Colitis Congress®, a partnership of AGA and the Crohn’s & Colitis Foundation, held in Las Vegas, evaluated whether a consensus-based MRI definition developed by an international group of experts known as the TopCLASS consortium could be reliably applied in real-world clinical settings and whether it correlated with improved patient outcomes.

According to Dr. Samaan, previous research has shown that clinical closure of perianal fistulas does not always reflect true healing beneath the surface. MRI studies have demonstrated that persistent inflammation may remain even when fistulas appear closed on physical exam, increasing the risk of recurrence. However, until now, there has been no standardized MRI definition of what constitutes true radiological healing.

The TopCLASS definition identifies healing based on three criteria: absence of T2 hyperintensity on MRI, a completely fibrotic fistula tract, and lack of contrast enhancement when contrast is used. Experts reached a 95% consensus on this definition.

To test it, researchers conducted a retrospective, multicenter study from 2021 to 2023 using data from Washington University in St. Louis, St. Mark’s Hospital in London, and Amsterdam University Medical Center, the Netherlands. Out of 977 patients screened, 40 met inclusion criteria after having MRIs locally reported as healed. Independent expert radiologists then re-reviewed the scans using the TopCLASS criteria.

Only 14 patients met the strict definition of radiological healing, highlighting how difficult the endpoint is to achieve. However, outcomes among these patients were markedly better: 93% achieved sustained clinical remission for at least one year, compared with 78% of those who did not meet the criteria for radiological healing. Fistular recurrence at one year was also less frequent in the radiologically healed group (7% vs. 22%).

Secondary outcomes also favored the radiologically healed group over the non-healed group, including lower rates of fistula-related antibiotic use (7% vs. 22%, respectively), seton placement (7% vs. 15%), proctectomy (7% vs. 11%), return to operating room (30% vs. 7%), defunctioning stoma (0% vs. 15%), and hospitalizations (7% vs. 19%).

The study found excellent inter-reader reliability, with agreement exceeding 95%. “This means that our definition was objective, easy to use, and highly reproducible across different sectors,” Dr. Samaan said.

He acknowledged limitations of the study, including its small sample size, retrospective design, and lack of comparison with data on quality of life.

“For future directions, we have ongoing funded work to validate this definition in a larger retrospective and prospective cohort,” Dr. Samaan said. “We're also incorporating radiomics and AI in collaboration with biomedical engineers to better understand fistula healing response on MRI.”

The researchers reported having no disclosures.

Summary content