Automatic quantification of adipose tissue in autosomal dominant polycystic kidney disease: a new tool for better understanding the disease pathophysiological mechanisms

Autosomal Dominant Polycystic Kidney Disease is the most common inherited nephropathy, characterized by progressive cyst growth that enlarges the kidneys and ultimately leads to renal failure. While Total Kidney Volume is a well established marker of disease progression, recent studies suggest that body fat distribution, particularly Visceral Adipose Tissue (VAT), may influence metabolic and inflammatory mechanisms contributing to cyst development. However VAT estimation using conventional MRI-based methods relying on single slices at fixed L3-L4 vertebrae, may be unreliable in ADPKD due to severe abdominal distortion caused by organomegaly. This thesis describes the development of an automatic and fully reproducible framework for volumetric quantification of abdominal adipose tissue in ADPKD using T1-weighted dual-echo axial MRI data. The algorithm, implemented in MATLAB, integrates adaptive image selection and dedicated correction steps for liver and intestine exclusion, enabling precise 3D segmentation of Subcutaneous and Visceral Adipose Tissue. Applied to MRI scans from 24 ADPKD patients, the method identified the slice at the inferior margin of the L3 vertebra as the most representative of total VAT volume in this population. A strong correlation was observed between VAT and Abdominal Circumference, confirming it as a more reliable indicator of visceral adiposity than Body Mass Index. No significant association emerged between VAT and renal function markers, suggesting that visceral fat primarily reflects systemic metabolic alterations rather than direct kidney impairment. Overall, this work introduces a robust and anatomically consistent tool for the objective assessment of adipose tissue distribution in ADPKD, supporting more accurate metabolic evaluation and personalized monitoring of disease progression.