Multiparametric MRI for Staging of Bowel Inflammatory Activity in Crohn's Disease with MUSE-IVIM and DCE-MRI: A Preliminary Study
Liangqiang Mao,Yan Li,Bota Cui,Lin Lu,Weiqiang Dou,Dmytro Pylypenko,Jianguo Zhu,Haige Li
DOI: https://doi.org/10.1016/j.acra.2023.08.028
IF: 5.482
2023-09-20
Academic Radiology
Abstract:Rationale and Objectives To investigate if the combination of multishot diffusion imaging-based multiplexed sensitivity encoding intravoxel incoherent motion (MUSE-IVIM) and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is feasible for staging Crohn's disease (CD) activity. Materials and Methods A total of 65 CD patients were enrolled and analyzed in this retrospective study. The simplified endoscopic score for Crohn's disease (SES-CD) and magnetic resonance index of activity (MaRIA) were used as the reference. The MUSE-IVIM and DCE-MRI data were acquired at 3.0-T MRI scanner and processed by two radiologists. Three MUSE-IVIM parameters: fast apparent diffusion coefficient (ADC fast ), slow apparent diffusion coefficient (ADC slow ), and the fractional perfusion (Fraction of ADC fast ), as well as four DCE-MRI parameters: volume transfer constant ( K trans ), rate constant ( K ep ), extravascular extracellular volume fraction ( V e ), and plasma volume fraction ( V p ) were generated. Intraclass correlation coefficient (ICC), non-parametric test (Kruskal–Wallis H and Mann–Whitney U), logistic regression, receiver operating characteristic analysis, Delong test, and Spearman's correlation test were performed. Results According to SES-CD, 116 ileocolonic segments with CD lesions were identified as: inactive, mild, and moderate to severe. With multivariable logistic regression analysis, ADC fast ( p < 0.001), Fraction of ADC fast ( p = 0.005), K trans ( p < 0.001) and K ep ( p = 0.003) were identified as significant factors for differentiating among the three groups. Binary logistic analyses identified ADC fast ( p = 0.001), K trans ( p = 0.014), and K ep ( p = 0.029) as independent predictors for the active status. The combination of ADC fast , K trans , and K ep performed better than MaRIA score ( p = 0.028), for differentiating inactive and active status. MaRIA score was positively correlated with ADC fast ( p < 0.001), K trans ( p < 0.001), K ep ( p < 0.001), and V e ( p = 0.001), however, negatively correlated with Fraction of ADC fast ( p < 0.001). Conclusion The combination of MUSE-IVIM and DCE-MRI has been demonstrated to accurately stage inflammatory activity in CD.
radiology, nuclear medicine & medical imaging