Chordoma and chondrosarcoma share common radiographic characteristics yet are distinct clinically. A radiomic machine learning model differentiating these tumors preoperatively would help plan surgery. MR images were acquired from 57 consecutive patients with chordoma (N = 32) or chondrosarcoma (N = 25) treated at the University of Tokyo Hospital between September 2012 and February 2020. Preoperative T1-weighted images with gadolinium enhancement (GdT1) and T2-weighted images were analyzed. Datasets from the first 47 cases were used for model creation, and those from the subsequent 10 cases were used for validation. Feature extraction was performed semi-automatically, and 2438 features were obtained per image sequence. Machine learning models with logistic regression and a support vector machine were created. The model with the highest accuracy incorporated seven features extracted from GdT1 in the logistic regression. The average area under the curve was 0.93 ± 0.06, and accuracy was 0.90 (9/10) in the validation dataset. The same validation dataset was assessed by 20 board-certified neurosurgeons. Diagnostic accuracy ranged from 0.50 to 0.80 (median 0.60, 95% confidence interval 0.60 ± 0.06%), which was inferior to that of the machine learning model (p = 0.03), although there are some limitations, such as the risk of overfitting and the lack of an extramural cohort for truly independent final validation. In summary, we created a novel MRI-based machine learning model to differentiate skull base chordoma and chondrosarcoma from multiparametric signatures.

Mapping of MR fiber g-ratio, which is the ratio of the diameter of the axon to the diameter of the neuronal fiber, is introduced in this article. We investigated the MR fiber g-ratio, the axon volume fraction (AVF) and the myelin volume fraction (MVF) to evaluate microstructural changes in the spinal cord in patients with cervical spondylotic myelopathy (CSM) in vivo, using atlas-based analysis. We used diffusion MRI data acquired with a new simultaneous multi-slice accelerated readout-segmented echo planar imaging sequence for diffusion analysis for AVF calculation and magnetization transfer saturation imaging for MVF calculation. The AVFs of fasciculus gracilis in the affected side spinal cord, fasciculus cuneatus and lateral corticospinal tracts (LSCT) in the affected and unaffected side spinal cord were significantly lower (P = 0.019, 0.001, 0019, 0.000, and 0.002, respectively) than those of normal controls. No difference was found in the MVFs. The fiber g-ratio of LSCT was significantly lower (P = 0.040) in the affected side spinal cords than in the normal controls. The pathological microstructural changes in the spinal cord in patients with CSM, presumably partial axonal degenerations with preserved myelin. This technique has the potential to be a clinical biomarker in patients with CSM in vivo.

To evaluate the potential of diffusion kurtosis imaging (DKI) analysis with the breath-hold technique to assess the stage or classify hepatic fibrosis. Patients (n=67) suspected of having a disease of the hepatobiliary system examined by diffusion-weighted imaging (DWI) using a 3.0-T magnetic resonance imaging unit were enrolled in this study. To evaluate hepatic fibrosis, mean kurtosis, Mean apparent diffusion (MD) and apparent diffusion coefficient (ADC) values were compared between groups with varying fibrosis; F0-F1, F2-F3, and F4. The Steel-Dwass test was used for overall comparisons. Correlations between the fibrosis stage and mean kurtosis, MD or ADC values were assessed using Spearman's rank correlation. Discriminative capacities of DKI were evaluated using receiver operating characteristic (ROC) analysis. There were significant differences in ADC, MD and mean kurtosis values between non-cirrhosis and cirrhosis groups. Moreover, the mean kurtosis value was statistically different between the F0-F1 and F2-F3, F0-F1 and F4, and F2-F3 and F4 groups (all P<0.05). MD value was statistically different between the F0-F1 and F4 groups, and F2-F3 and F4 groups (all P0.05). In addition, mean kurtosis and MD values significantly correlated with the extent of hepatic fibrosis staging (Spearman's rank correlation coefficient, ρ=0.851 and -0.672; P<0.0001). However, ADC values did not reveal a correlation with the extent of hepatic fibrosis staging (ρ=-0.227; P=0.078). According to the ROC analysis for the assessment of no fibrosis (F0), fibrosis (≥F1), and advanced fibrosis (≥F2) and liver cirrhosis, the DKI cut-off values were 0.923, 0.955, and 1.11, respectively. Using the DKI method with the breath-hold technique in the liver, the stage of hepatic fibrosis can be classified into normal and early hepatic fibrosis, substantial stages, and advanced hepatic fibrosis.

BACKGROUND AND PURPOSE: The hemodynamics associated with cerebral AVMs have a significant impact on their clinical presentation. This study aimed to evaluate the hemodynamic features of AVMs using 3D phase-contrast MR imaging with dual velocity-encodings. MATERIALS AND METHODS: Thirty-two patients with supratentorial AVMs who had not received any previous treatment and had undergone 3D phase-contrast MR imaging were included in this study. The nidus diameter and volume were measured for classification of AVMs (small, medium, or large). Flow parameters measured included apparent AVM inflow, AVM inflow index, apparent AVM outflow, AVM outflow index, and the apparent AVM inflow-to-outflow ratio. Correlation coefficients between the nidus volume and each flow were calculated. The flow parameters between small and other AVMs as well as between nonhemorrhagic and hemorrhagic AVMs were compared. RESULTS: Patients were divided into hemorrhagic (n = 8) and nonhemorrhagic (n = 24) groups. The correlation coefficient between the nidus volume and the apparent AVM inflow and outflow was .83. The apparent AVM inflow and outflow in small AVMs were significantly smaller than in medium AVMs (P < .001 for both groups). The apparent AVM inflow-to-outflow ratio was significantly larger in the hemorrhagic AVMs than in the nonhemorrhagic AVMs (P = .02). CONCLUSIONS: The apparent AVM inflow-to-outflow ratio was the only significant parameter that differed between nonhemorrhagic and hemorrhagic AVMs, suggesting that a poor drainage system may increase AVM pressure, potentially causing cerebral hemorrhage.