Deep Learning-Based MRI Volumetry for Living Kidney Donor Assessment: A New Tool for Predicting Post-Donation Renal...

Area of Science:

• Nephrology
• Radiology
• Medical Imaging

Background:

• Living kidney donation is vital for addressing organ shortages.
• Donor safety relies on precise preoperative assessment of kidney anatomy and function.
• Deep learning MRI volumetry is explored as a novel assessment tool.

Purpose of the Study:

• To evaluate deep learning MRI volumetry for kidney volume estimation in living donors.
• To compare MRI volumetry with intraoperative measurements and renal scintigraphy.
• To assess the correlation between MRI-derived metrics and post-donation estimated glomerular filtration rate (eGFR).

Main Methods:

• Retrospective analysis of 178 living kidney donors.
• Comparison of deep learning MRI volumetry with water displacement method for kidney volume.
• Correlation of MRI-based volume ratios with scintigraphy-based split renal function ratios.

Main Results:

• Deep learning MRI volumetry showed strong correlation with intraoperative kidney volumes (r=0.7671).
• MRI-based volume ratios had moderate correlation with scintigraphy ratios (r=0.4798).
• MRI volumetry correlated better with 1-year post-donation eGFR than renal scintigraphy (r=0.6829 vs. 0.6191).

Conclusions:

• Deep learning MRI volumetry is a reliable, non-invasive tool for estimating kidney volume in living donors.
• It offers a radiation-free alternative for preoperative assessment.
• MRI volumetry shows potential for predicting post-donation eGFR, supporting its role in donor evaluation.