Intravoxel incoherent motion (IVIM) and non-Gaussian diffusion MRI of the lactating breast

Abstract

PurposeTo investigate the effect of breastfeeding on IVIM and non-Gaussian diffusion MRI in the breast. Materials and methodsAn IRB approved prospective study enrolled seventeen volunteers (12 in lactation and 5 with post-weaning, range 31–43 years; mean 35.4 years). IVIM (fIVIM and D*) and non-Gaussian diffusion (ADC0 and K) parameters using 16 b values, plus synthetic apparent diffusion coefficients (sADCs) from 2 key b values (b = 200 and 1500 s/mm2) were calculated using regions of interest. ADC0 maps of the whole breast were generated and their contrast patterns were evaluated by two independent readers using retroareolar and segmental semi-quantitative scores. To compare the diffusion and IVIM parameters, Wilcoxon signed rank tests were used between pre- and post-breastfeeding and Mann-Whitney tests were used between post-weaning and pre- or post-breastfeeding. ResultsADC0 and sADC values significantly decreased post-breastfeeding (1.90 vs. 1.72 × 10−3 mm2/s, P < 0.001 and 1.39 vs. 1.25 × 10−3 mm2/s, P < 0.001) while K values significantly increased (0.33 vs. 0.44, P < 0.05). fIVIM values significantly increased after breastfeeding (1.97 vs. 2.97%, P < 0.01). No significant difference was found in D* values. There was significant heterogeneity in ADC0 maps post-breastfeeding, both in retroareolar and segmental scores (P < 0.0001 and =0.0001). ConclusionIVIM and non-Gaussian diffusion parameters significantly changed between pre- and post-breastfeeding status, and care needs to be taken in interpreting diffusion-weighted imaging (DWI) data in lactating breasts.