Combining Multifrequency Magnetic Resonance Elastography With Automatic Segmentation to Assess Renal Function in...

Area of Science:

• Medical Imaging
• Nephrology
• Biophysics

Background:

• Multifrequency MR elastography (mMRE) quantifies renal stiffness noninvasively in chronic kidney disease (CKD).
• Manual kidney segmentation in mMRE is time-consuming and increases interobserver variability.

Purpose of the Study:

• To evaluate the performance of mMRE combined with automatic segmentation for assessing CKD severity.
• To determine if automated segmentation improves the efficiency and reliability of mMRE in CKD patients.

Main Methods:

• A prospective study included 179 participants (95 healthy, 84 CKD).
• Renal stiffness (shear wave speed, SWS) was measured using mMRE with nnU-Net deep-learning for automatic segmentation.
• SWS values were correlated with estimated glomerular filtration rate (eGFR) and compared between healthy and CKD subgroups.

Main Results:

• Automatic segmentation achieved high Dice similarity coefficients (0.901–0.970) compared to manual segmentation.
• Automatically quantified SWS strongly correlated with eGFR (r=0.620–0.640).
• CKD patients (stage 1-2) showed significantly lower cortical SWS than healthy volunteers, even with mild impairment.

Conclusions:

• mMRE combined with automatic segmentation effectively reveals abnormal renal stiffness in CKD patients, including those with mild impairment.
• This automated approach enhances diagnostic efficiency and reduces subjectivity in assessing renal stiffness for CKD management.