Objectives To determine whether MRI findings are associated with tumor behavior in desmoid-type fibromatosis (DF) during active surveillance. Materials and methods This single-center retrospective study included patients with histopathologically confirmed DF who underwent active surveillance between January 2008 and March 2024. Tumor behavior was classified by percentage monthly change in tumor volume (PMCV) as growth ( > 1%), stable ( −1% to +1%), or shrinkage (<−1%). Longitudinal analysis included 61 observation intervals derived from 85 MRI examinations; a separate patient-based analysis included baseline MRI examinations from 24 patients. MRI findings included qualitative features, whole-tumor T2 ratio, whole-tumor apparent diffusion coefficient (ADC), and semiautomated three-dimensional volumetry. Mixed-effects logistic regression models were used for longitudinal analysis, and interobserver agreement was assessed. Results Twenty-four patients with 85 MRI examinations were included. PMCV was calculated for 61 observation intervals: 9 shrinkage, 13 stable, and 39 growth. No qualitative MRI feature, including margin definition, extensive T2 hyperintensity, band sign on T1- or T2-weighted imaging, tail sign, Yin–Yang sign, or dynamic contrast-enhanced MRI pattern, was significantly associated with interval tumor behavior. Neither whole-tumor T2 ratio nor whole-tumor ADC was significantly associated with growth or shrinkage relative to stability. Interobserver agreement ranged from substantial to almost perfect for qualitative features and was excellent for quantitative measurements. Conclusion No evaluated MRI feature was significantly associated with tumor behavior in DF during active surveillance, suggesting limited value of MRI as a standalone predictor. Key Points <Emphasis Type="BoldItalic">Question</Emphasis> Can MRI help predict tumor behavior during active surveillance in patients with DF, given the lack of established imaging predictors? <Emphasis Type="BoldItalic">Findings</Emphasis> No evaluated MRI feature, including qualitative findings, whole-tumor T2 ratio, or whole-tumor ADC value, was significantly associated with tumor behavior during active surveillance. <Emphasis Type="BoldItalic">Clinical relevance</Emphasis> Although MRI remains essential for monitoring DF, its standalone value for predicting tumor behavior during active surveillance may be limited. Graphical Abstract

MRI assessment of tumor behavior in desmoid-type fibromatosis during active surveillance

期刊论文
发表日期：

2026年7月17日

查看全文

Shuji Nagata,
Aiko Yamaguchi,
Yuki Shouji,
Miyuki Sawano,
Yusuke Uchiyama,
Hideo Arai,
Hiroshi Nishimura,
Shuichi Tanoue

摘要

Objectives

To determine whether MRI findings are associated with tumor behavior in desmoid-type fibromatosis (DF) during active surveillance.

Materials and methods

This single-center retrospective study included patients with histopathologically confirmed DF who underwent active surveillance between January 2008 and March 2024. Tumor behavior was classified by percentage monthly change in tumor volume (PMCV) as growth ( > 1%), stable ( −1% to +1%), or shrinkage (<−1%). Longitudinal analysis included 61 observation intervals derived from 85 MRI examinations; a separate patient-based analysis included baseline MRI examinations from 24 patients. MRI findings included qualitative features, whole-tumor T2 ratio, whole-tumor apparent diffusion coefficient (ADC), and semiautomated three-dimensional volumetry. Mixed-effects logistic regression models were used for longitudinal analysis, and interobserver agreement was assessed.

Results

Twenty-four patients with 85 MRI examinations were included. PMCV was calculated for 61 observation intervals: 9 shrinkage, 13 stable, and 39 growth. No qualitative MRI feature, including margin definition, extensive T2 hyperintensity, band sign on T1- or T2-weighted imaging, tail sign, Yin–Yang sign, or dynamic contrast-enhanced MRI pattern, was significantly associated with interval tumor behavior. Neither whole-tumor T2 ratio nor whole-tumor ADC was significantly associated with growth or shrinkage relative to stability. Interobserver agreement ranged from substantial to almost perfect for qualitative features and was excellent for quantitative measurements.

Conclusion

No evaluated MRI feature was significantly associated with tumor behavior in DF during active surveillance, suggesting limited value of MRI as a standalone predictor.

Key Points

Question Can MRI help predict tumor behavior during active surveillance in patients with DF, given the lack of established imaging predictors?

Findings No evaluated MRI feature, including qualitative findings, whole-tumor T2 ratio, or whole-tumor ADC value, was significantly associated with tumor behavior during active surveillance.

Clinical relevance Although MRI remains essential for monitoring DF, its standalone value for predicting tumor behavior during active surveillance may be limited.

Graphical Abstract