The American Society of Nephrology (ASN) Kidney Week 2020 has just finished. This year, a reimagined, fully virtual event gathered people from all around the world to discuss and explore the field of kidney disease.
Collaboration between Antaros Medical and researchers at Uppsala University has resulted in a deep-learning pipeline for automated measurements of parenchymal kidney volume in Magnetic Resonance Images (MRI), generating large-scale data which was presented at this year’s ASN Kidney Week.
Kidney volume is a well-known marker for the state of the kidney and has been associated with several aspects for kidney disease, including:
- In autosomal dominant polycystic kidney disease (ADPKD) kidney volume is an important biomarker for diagnosis and prognosis
- In chronic kidney disease (CKD), kidney volume decreases with kidney function as a cause of the ongoing pathophysiological processes including fibrosis
- The early stage of type 2 diabetes (T2D) is associated with an increased kidney size which can lead to diabetic nephropathy
There is, however, a lot more to be explored for these relationships as well as other confounders. The presented large-scale MRI kidney volume analysis (n=~40,000) investigated relationships between kidney volume and multiple parameters such as age, body size/composition measurements as well as T2D. As previously known, kidney volume declines with age, but the analysis showed that males have an increase rate of decline with age. Interestingly, kidney volume was larger in middle-aged T2D subjects than non-T2D, possibly explained by increased hyperfiltration and oxidative stress in T2D. Future work might elucidate further insights of the biological importance of kidney volume.
To find out more, have a look at the following contributions at the ASN Kidney Week 2020.
Kidney Segmentation with Deep Learning in MRI of 40,000 UK Biobank Subjects
Authors: T Langner, A Östling, L Maldonis, A Karlsson, D Olmo, D Lindgren, A Wallin, L
Lundin, R Strand, H Ahlström, J Kullberg
Authors: J Kullberg, T Langner, D Rivas, I Friedli, T Fall, R Strand, H Ahlström, L Johansson
Authors: L Johansson, T Langner, P Hockings, L Jarl, L Maldonis, T Fall, H Ahlström, J Kullberg
This work was performed using the UK Biobank under Application no. 14237.