Characterization of the diffusion signal of breast tissues using multi‐exponential models

• DOI: 10.1002/mrm.29090
• Publication Date: 2021-12-14T11:44:58Z
• Authors: Ana E. Rodríguez‐...; Maren M. Sjaastad...; Lauren K. Fang; Christopher C. Co...; Helen H. Park; Grace S. Ahn; Hauke Bartsch; Joshua Kuperman; Igor Vidić; Haydee Ojeda‐Four...; Anne M. Wallace; Michael Hahn; Tyler M. Seibert; Neil Peter Jerome; Agnes Østlie; Tone Frost Bathen; Pål Erik Goa; Rebecca Rakow‐Penner; Anders M. Dale
• Keywords: Breast MRI; SIGNAL (programming language)

Purpose Restriction spectrum imaging (RSI) decomposes the diffusion‐weighted MRI signal into separate components of known apparent diffusion coefficients (ADCs). The number and optimal ADCs for RSI are organ‐specific determined empirically. purpose this work was to determine model breast tissues. Methods described using a linear combination multiple exponential components. A set ADC values estimated fit voxels in cancer control ROIs. Later, contributions each component were these fixed values. Relative‐fitting residuals Bayesian information criterion assessed. Contrast‐to‐noise ratio between fibroglandular tissue RSI‐derived contribution maps compared DCE imaging. Results total 74 women with scanned at 3.0 Tesla MRI. fitting conventional suggest that 3‐component improves characterization over biexponential model. Estimated triexponential D 1,3 = 0, 2,3 1.5 × 10 −3 , 3,3 10.8 mm 2 /s. slower larger tumors than Further, contrast‐to‐noise specificity 80% sensitivity subset equivalent. Conclusion Breast best Tumor conspicuity is comparable without use exogenous contrast. These data may be used as differential features healthy malignant