Evaluation of renal dysfunction using texture analysis based on DWI, BOLD, and susceptibility-weighted imaging

Area of Science:

• Radiology and Medical Imaging
• Nephrology
• Biomedical Engineering

Background:

• Renal dysfunction assessment relies on traditional markers like estimated glomerular filtration rate (eGFR).
• Novel imaging biomarkers are needed for more sensitive evaluation of kidney function.
• Texture analysis offers quantitative insights into tissue heterogeneity from MRI data.

Purpose of the Study:

• To evaluate the utility of texture analysis from diffusion-weighted imaging (DWI), blood oxygen level-dependent (BOLD), and susceptibility-weighted imaging (SWI) in assessing renal dysfunction.
• To identify specific texture features that can differentiate between varying degrees of renal impairment.
• To compare the effectiveness of different MRI sequences for texture-based renal function assessment.

Main Methods:

• Seventy-two patients were categorized into severe renal impairment (sRI), non-severe renal impairment (non-sRI), and control (CG) groups based on eGFR.
• MRI scans were acquired using DWI, BOLD, and SWI sequences.
• Thirteen texture features were extracted and analyzed to identify those most valuable for group differentiation, with ROC curve analysis used for comparative assessment.

Main Results:

• Six key texture features were identified: 0.25Quantile_DWI, 0.5Quantile_DWI, Homogeneity_DWI, Entropy_BOLD, Skewness_SWI, and Correlation_SWI.
• DWI-derived features (0.25Quantile_DWI, 0.5Quantile_DWI, Homogeneity_DWI) were significantly lower in sRI compared to non-sRI.
• Entropy_BOLD and Correlation_SWI were lower in non-sRI compared to CG, indicating potential for early-stage detection.

Conclusions:

• Texture analysis of DWI, BOLD, and SWI MRI is a valuable tool for assessing renal dysfunction.
• Texture features derived from BOLD and SWI show particular promise for evaluating early-stage renal function impairment.
• SWI demonstrated comparable efficacy to BOLD in assessing renal dysfunction.