To estimate the variation of the vertebral microvascular permeability and vertebral fat deposition in alloxan-induced diabetic rabbits using DCE-MRI and IDEAL-IQ.Twelve young New Zealand White rabbits were randomly assigned to alloxan-induced diabetic group (n=6) and control group (n=6). The rabbits in alloxan-induced diabetic group were injected with a total amount of 100mg/kg of alloxan. All rabbits underwent sagittal magnetic resonance imaging (IDEAL-IQ, DCE-MRI) of lumbar on a 3.0T scanner (GE Discovery 750 Plus) with an 8 channel knee coil at fixed time points (0, 4, 8, 12, 16 week). Imaging parameters for IDEAL- IQ were set as follows: FA = 6°; TE = 1.2, 3.2, 5.2, 7.2, 9.2 and 11.2 ms; TR = 19.6 ms; FOV = 16 x 12.8 cm²; slice thickness = 3 mm; the dynamic contrast-enhanced imaging was performed using a LAVA sequence with following parameters: TE = 2.3 ms, TR = 4.9 ms, FA = 15°, FOV = 16 x 12.8 mm². The contrast agent with a concentration of 0.2 mmol/kg was injected after one pre-contrast frame was acquired. L7 was chosen to measure vertebral fat fraction and DCE-MRI parameters. Fat fraction(FF) was measured on the fat fraction map produced by IDEAL-IQ to estimate the variation of vertebral fat deposition. The DCE parameters were measured using GE Omni kinetics software. The Extended Tofts model was used to estimate the quantitative parameters including the transfer constant (K^trans) the rate contrast (k_ep), fractional EES volume (v_e) and fractional plasma volume (v_p)¹ in the ROI shown in Fig. 1a (right). The Arterial Input Function (AIF) was determined from a circular ROI in the center of abdominal aorta at the plane of L7 (left in Fig. 1a). Vertebral fat fraction was measured in fat fraction map in the ROI shown in Fig1. b. At the week 16, all rabbits were sacrificed, after which L7 sampling, HE staining and immunoperoxidase CD31 labeling . HE staining and immunoperoxidase CD31 labeling were performed as gold standards to estimate the variation of bone marrow cell and vertebral microvascular respectively. To count microvessel density (MVD), a quantitative estimation was performed using a light microscopy in a ×200 field. Repetitive measure analysis of variance (ANOVA) was applied in analyzing FF and DCE-MRI parameters at different time points. Pearson correlations of DCE-MRI parameters with FF and with MVD were analyzed, respectively. All of the thresholds of statistical significance were set at P < 0.05. HE staining and immunohistochemistry showed that in comparison to the control group, the amount of the fat cell in diabetic group increased while the bone marrow cells and microvessel reduced (Fig. 2 and 3). The FF, K^trans and k_ep increased while v_e and v_p decreased in the diabetic group (Fig. 4). Only the alloxan-induced diabetic group had statistically significant differences in vertebral fat fraction and DCE parameters at each time point (FF: F=50.387, P<0.01; K^trans: F=63.694, P<0.05; k_ep: F=5.04, P<0.05; v_p: F=4.403, P<0.05; v_p: F=9.751, P<0.05). As shown in Fig.5, MVD showed negative correlation with K^trans and k_ep and positive correlation with v_e and v_p (K^trans: r=-0.901, P<0.05; k_ep: r=-0.731, P<0.05; v_e: r=0.741, P<0.05; v_p: r=0.593, P<0.05) (Fig. 5a-d) while FF showed positive correlation with K^trans and k_ep and negative correlation with v_e and v_p (K^trans: r=0.863, P<0.05; k_ep: r=0.641, P<0.05; v_e: r=-0.629, P<0.05; v_p: r=-0.665, P<0.05) (Fig. 5e-h).The results of this study demonstrated diabetes mellitus could cause the variation of the vertebral microvascular permeability, reduce the bone marrow cells and cause fat accumulation. In this study, MVD showed negative correlation with K^trans and k_ep. It inferred that the increase of Ktrans and kep could indicate a destruction of sinusoids and capillaries. A consistent reduction of sinusoids and arterioles can cause the decrease of v_p which represent plasma volume. FF showed negative correlation with v_e and v_p. The excess fat in vertebral bone marrow may replace hematological cells² and squeeze microvascular, which may be the explanation of decreased interstitial volume (v_e). Compared with previous study used marrow blood flow by Oikawa³, our method achieved consistent results in a noninvasive and simpler way. DCE-MRI and IDEAL-IQ can evaluate quantitatively the variation trends of thevertebral microvascular permeability and vertebral fat deposition in alloxan-induced diabetic rabbits. The variety of vertebral microvascular permeability is strongly associated with the increasing vertebral fat deposition.