IVIM MRI shows great potential in noninvasive assessment of renal fibrosis induced by UUO. J. Magn. Reson. Imaging 2016;44:698-706.

Background COVID-19 has overwhelmed health systems worldwide. It is important to identify severe cases as early as possible, such that resources can be mobilized and treatment can be escalated. Objective This study aims to develop a machine learning approach for automated severity assessment of COVID-19 based on clinical and imaging data. Methods Clinical data—including demographics, signs, symptoms, comorbidities, and blood test results—and chest computed tomography scans of 346 patients from 2 hospitals in the Hubei Province, China, were used to develop machine learning models for automated severity assessment in diagnosed COVID-19 cases. We compared the predictive power of the clinical and imaging data from multiple machine learning models and further explored the use of four oversampling methods to address the imbalanced classification issue. Features with the highest predictive power were identified using the Shapley Additive Explanations framework. Results Imaging features had the strongest impact on the model output, while a combination of clinical and imaging features yielded the best performance overall. The identified predictive features were consistent with those reported previously. Although oversampling yielded mixed results, it achieved the best model performance in our study. Logistic regression models differentiating between mild and severe cases achieved the best performance for clinical features (area under the curve [AUC] 0.848; sensitivity 0.455; specificity 0.906), imaging features (AUC 0.926; sensitivity 0.818; specificity 0.901), and a combination of clinical and imaging features (AUC 0.950; sensitivity 0.764; specificity 0.919). The synthetic minority oversampling method further improved the performance of the model using combined features (AUC 0.960; sensitivity 0.845; specificity 0.929). Conclusions Clinical and imaging features can be used for automated severity assessment of COVID-19 and can potentially help triage patients with COVID-19 and prioritize care delivery to those at a higher risk of severe disease.

Objective: Patients with diabetes mellitus, diabetic nephropathy (DN) and healthy donor were analyzed to test whether the early DN patients can be detected using both blood oxygenation level dependent (BOLD) and diffusion tensor imaging. Methods: This study was approved by the Ethics Committee of our hospital. MR images were acquired on a 3.0-Tesla MR system (Discovery MR750, General Electric, Milwaukee, WI). 30 diabetic patients were divided into NAU (normal to mildly increased albuminuria, N = 15) and MAU (moderately increased albuminuria, N = 15) group based on the absence or presence of microalbuminuria. 15 controls with sex- and age-matched were enrolled in the study. Prior to MRI scan, all participants were instructed to collect their fresh morning urine samples for quantitative measurement of urinary microalbumin and urinary creatinine. Then, the estimations of serum creatinine, serum uric acid, HbAlc and fasting plasma glucose as well as fundus examinations were performed in all subjects. Then, the values of albumin–creatinine ratio (ACR) and estimated glomerular filtration rate were also calculated. All subjects underwent renal diffusion tensor imaging (DTI) and BOLD acquisition after fasting for 4 h. Regions of interest were placed in renal medulla and cortex for evaluating apparent diffusion coefficient (ADC), fractional anisotropy (FA) and R2* values by two experienced radiologists. The consistency between the two observations was estimated using intragroup correlation coefficients. To test differences in ADC, FA and R2* values across the three groups, the data were analyzed using separate one-way ANOVAs. Post-hoc pair wise comparisons were then performed using t-test. To investigate the clinical relevance of imaging parameters in both regions across the three groups, the correlations of values of the ACR/estimated glomerular filtration rate and of the ADC/FA/R2* were calculated. Results: There was a high level of consistency of those ADC, FA and R2* values across the three groups on both renal cortex and medulla measured by the two doctors. The FA value of medulla in MAU group was lower than that in control (p < 0.01). The R2* value of medulla in the NAU group was higher than that in the control (p < 0.01), and the R2* value of medulla in the MAU group was lower than that in the control (p = 0.009) . Moreover, the current study revealed a decreasing trend in FA values of the renal medulla from the control group to NAU and MAU groups. Finally, a weak negatively correlation between medullary R2* and ACR was found in current study. Conclusion: Medullary R2* value might be a new more sensitive predictor of early DN. Meanwhile, BOLD imaging detected the medullary hypoxia at the simply diabetic stage, while DTI didn’t identify the medullary directional diffusion changes at this stage. Based on our assumption mentioned above, it’s presumable that BOLD imaging may be more sensitive for assessment of the early renal function changes than DTI. These imaging techniques are more accurate and practical than conventional tests. Advances in knowledge: Non-invasive MRI was used to detect renal function changes at early DN stage.

IVIM DWI might be helpful in noninvasively identifying early-stage DN. The IVIM parametric values are more sensitive than the ACR in detecting early-stage kidney changes.