In our study, we determined the fractional blood volume in skeletal muscle tissue with histological and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) methods.We compare histological results with results of various DCE-MRI methods. We determined the percentage tissue-blood volume fraction (v(b)) in the medial thigh tissue of twelve female pigs using of a 3D gradient echo sequence with k-space-sharing. Two contrast agents were used: a rapidly extravasating, low-molecular-weight contrast agent (LMCA, gadoterate meglumine) and an albumin-binding, slowly extravasating blood pool contrast agent (BPCA, gadofosveset trisodium)¹. For analysis of the LMCA data, the extended Tofts model (ETM)² and the 2-compartmental exchange Model (2CXM) were used³. For analysis of the BPCA data, we first used a bolus convolution method⁴ (thereby also taking into account the bolus dispersion⁵). Secondly, we applied an equilibrium MRI method (Eq-MRI) by calculating the ratio of the amount of blood in the tissue curve at steady state (averaged over 20 sec starting 40 sec after bolus arrival) to the height of the same area of the arterial input function (AIF). The AIF was measured in full blood voxels in the abdominal aorta.The model fits with the best quality were obtained at an acquisition time of 10 min. The results are summarized in table 1 (Fig. 1). Histological measurements (exemplary section shown in Fig. 2) yielded median values (25%-75% quartile) of v(b)=1.0 (0.8-1.2)% compared with 2.2 (1.5-2.8)% for BPCA bolus deconvolution, 2.5 (1.9-3.6)% for BPCA Eq-MRI, 2.0 (1.3-3.1)% for LMCA-2CXM, and 0.3 (0.2-0.3)% LMCA-ETM. Figure 3 lists the correlations between the different MR techniques. The closest match was found between the two BPCA methods (r = 0.93). Moreover, relatively high correlations were found between LMCA-2CXM and BPCA bolus deconvolution (BPCA Eq-MRI) with r = 0.87 (0.84). Low correlation was found for v(b) results between the ETM and the 2CXM (r = 0.47). The correlation between the results of ETM and the BPCA methods was not significant. Fig. 4 presents two examples of curve fits for the ETM and the 2CXM including the fit results for the individual compartments. The 2CXM resulted in a far better curve fit and was accompanied by distinct bolus broadening of the vascular compartment curve.The difference of approx. 1% between histological determination of v(b) and that of the MRI methods might be attributable to a collapse of blood vessels in the tissue biopsies. For the MRI methods, variations in v(b) estimation might arise from the uncertainty of systemic and microvascular hematocrit. Based on published data⁶ , we assumed a systemic hematocrit of 0.4 and a microvascular of 0.2 for our animals. In contrast to the LMCA models, the bolus method corrects for bolus dispersion, which should yield a higher specific blood volume. The ETM underestimates v(b) values, resulting in approximately only one third (compared to histologic estimates) to one tenth (Eq-MRI) of the blood volumes estimated with the other methods. The reason for the high accuracy of blood volume determination with the 2CXM might be that this model accounts for widening of the bolus in the blood compartment, which leads to an increase in the area under the curve for the vascular compartment.Even in low-perfused and low-vascularized muscle tissue, there is fairly good agreement between the v(b) estimates obtained with the different MR techniques. The only notable exception is the ETM, which markedly underestimates blood volume. Based on its very good fit quality, the 2CXM is a promising alternative to the Tofts models also in routine clinical practice.