MRI-Adapted M-RECIST Better Characterizes Treatment Response After Cryoablation in Desmoid-Type Fibromatosis

Clinical Summary:

• Design/Population: This retrospective study compared RECIST 1.1 with MRI-adapted RECIST (M-RECIST) for evaluating response following percutaneous cryoablation in patients with progressing extra-abdominal desmoid-type fibromatosis. 
• Key Outcomes: M-RECIST identified substantially more objective responses than RECIST 1.1, including complete responses that conventional dimensional criteria frequently classified as stable disease. Both approaches showed similar non-progression rates and complete agreement in identifying progressive disease. 
• Clinical Relevance: These findings suggest that incorporating MRI-based assessment of residual viable tumor may provide a more nuanced measure of response following cryoablation, although further studies are needed to determine whether M-RECIST response categories better predict clinical outcomes. 

Results from a retrospective study demonstrated that MRI-adapted response evaluation criteria in solid tumors (M-RECIST) may improve radiologic response assessment compared with RECIST 1.1 following percutaneous cryoablation in patients with progressing extra-abdominal desmoid-type fibromatosis. 

In this single-center analysis, investigators retrospectively evaluated consecutive patients with progressing extra-abdominal desmoid-type fibromatosis who underwent percutaneous cryoablation between July 2021 and April 2025. The final cohort included 34 patients who underwent 37 procedures. RECIST 1.1 assessments were available for 35 procedures and M-RECIST assessments for 34.

Unlike RECIST 1.1, which assesses changes in the longest overall tumor diameter, M-RECIST was designed to assess the longest diameter of residual viable tumor using multiparametric MRI. Viability was determined through combined evaluation of T1-weighted, T2-weighted, and diffusion-weighted imaging, as well as contrast-enhanced T1-weighted imaging when available. 

At a median follow-up of 15.7 months, RECIST 1.1 classified 28.6% of evaluable procedures as partial responses, 62.9% as stable disease, and 8.6% as progressive disease. The resulting overall response rate (ORR) was 28.6%, with a non-progression rate of 91.4%. No complete responses were identified using RECIST 1.1. 

Using M-RECIST, the ORR increased to 79.4% at a median follow-up of 16 months. Complete responses were observed in 44.1% of evaluable procedures and partial responses in 35.3%, while 11.8% were classified as stable disease and 8.8% as progressive disease.  The non-progression rate remained similar at 91.2%.

Despite the substantial difference in response categorization, RECIST 1.1 and M-RECIST were fully concordant in identifying progressive disease. Overall agreement across complete response, partial response, stable disease, and progressive disease categories was negligible, however, indicating that most discordance occurred in the classification of tumors that had not progressed.

The difference was also evident when treatment response was assessed continuously. The median reduction in longest tumor diameter according to RECIST 1.1 was 15.1%, compared with a 75% reduction in the longest viable tumor diameter using M-RECIST. Agreement between the two measurements was poor, further demonstrating the difference between dimensional and viability-based response assessment.

Notably, patients classified as having a complete response by M-RECIST at their first post-treatment MRI, performed approximately 30 to 40 days after cryoablation, maintained complete response during subsequent follow-up. However, the study was not designed to determine whether M-RECIST response categories were associated with improved clinical outcomes.

No severe postprocedural complications were reported. The most frequently observed symptoms following cryoablation were edema of the subcutaneous tissue in the treated area and pain.

The authors noted that conventional RECIST 1.1 may underestimate response following locoregional therapy because treatment-induced tissue changes can occur without immediate tumor shrinkage. By incorporating MRI characteristics of residual viable disease, M-RECIST may distinguish complete and partial responses that would otherwise be categorized as stable disease based on tumor dimensions alone.

The study was limited by its retrospective, single-center design, relatively small cohort, and absence of a prespecified correlation between imaging response and clinical outcomes. Further prospective validation will be needed to determine whether M-RECIST provides greater prognostic value than conventional RECIST 1.1 following cryoablation.

Source:

Vanzulli A, Saggiante L, Sciacqua LV, et al. Radiologic response assessment in patients with desmoid-type fibromatosis treated with percutaneous cryoablation. Eur Radiol Exp. Published online: July 1, 2026. doi: 10.1186/s41747-026-00767-2