Investigation of breast cancer microstructure and microvasculature from time-dependent DWI and CEST in correlation with histological biomarkers.

We investigated the associations of time-dependent DWI, non-Gaussian DWI, and CEST parameters with histological biomarkers in a breast cancer xenograft model. 22 xenograft mice (7 MCF-7 and 15 MDA-MB-231) were scanned at 4 diffusion times [T = 2.5/5 ms with 11 b-values (0-600 s/mm) and T = 9/27.6 ms with 17 b-values (0-3000 s/mm), respectively]. The apparent diffusion coefficient (ADC) was estimated using 2 b-values in different combinations (ADC using b = 0 and 600 s/mm and shifted ADC [sADC] using b = 200 and 1500 s/mm) at each of those diffusion times. Then the change (Δ) in ADC/sADC between diffusion times was evaluated. Non-Gaussian diffusion and intravoxel incoherent motion (IVIM) parameters (ADC, the virtual ADC at b = 0; K, Kurtosis from non-Gaussian diffusion; f, the IVIM perfusion fraction) were estimated. CEST images were acquired and the amide proton transfer signal intensity (APT SI) were measured. The ΔsADC (between [Formula: see text] and [Formula: see text] and ΔADC_sADC (between [Formula: see text] and [Formula: see text]) was significantly larger for MCF-7 groups, and ΔADC_sADC was positively correlated with Ki67 and APT SI. ADC decreased significantly in MDA-MB-231 group and K increased significantly with T in MCF-7 group. APT SI and cellular area had a moderately strong positive correlation in MDA-MB-231 and MCF-7 tumors combined, and there was a positive correlation in MDA-MB-231 tumors. There was a significant negative correlation between APT SI and the Ki-67-positive ratio in MDA-MB-231 tumors and when combined with MCF-7 tumors. The associations of ΔADC_sADC and API SI with Ki-67 parameters indicate that the T-dependent DW and CEST parameters are useful to predict the histological markers of breast cancers.