ASN's Mission

To create a world without kidney diseases, the ASN Alliance for Kidney Health elevates care by educating and informing, driving breakthroughs and innovation, and advocating for policies that create transformative changes in kidney medicine throughout the world.

learn more

Contact ASN

1401 H St, NW, Ste 900, Washington, DC 20005

email@asn-online.org

202-640-4660

The Latest on X

Kidney Week

Abstract: TH-PO564

An Unbiased and Automated Approach: Artificial Intelligence (AI)-Based Pipeline for Glomerulosclerosis Scoring in Rodent Models of CKD

Session Information

Category: Glomerular Diseases

  • 1401 Glomerular Diseases: Mechanisms, including Podocyte Biology

Authors

  • Frias Hernandez, Alex, Gubra AS, Horsholm, Hovedstaden, Denmark
  • Ougaard, Maria Katarina, Gubra AS, Horsholm, Hovedstaden, Denmark
  • Jensen, Ditte Marie, Gubra AS, Horsholm, Hovedstaden, Denmark
  • Christensen, Michael, Gubra AS, Horsholm, Hovedstaden, Denmark
Background

Glomerulosclerosis is a kidney disease that involves the formation of scar tissue in the glomeruli in patients with chronic kidney disease (CKD). An accurate histopathological evaluation of glomerulosclerosis plays a crucial role in the diagnosis, prognosis and treatment of people suffering from CKD. Several rodent models mimic glomerulosclerosis and aid in preclinical target identification and drug development for CKD. To accelerate the objective and efficient assessment of glomerulosclerosis, we have designed an AI-powered scoring system using deep learning algorithms, adapted to murine and rat CKD models.

Methods

A deep learning model was trained on a large sample test set (4293 glomeruli images) of mouse and rat PAS-stained kidney sections and compared to expert histopathologist-verified manual glomerulosclerosis scoring. AI model performance was validated using an independent kidney sample set from three different mouse models of CKD (adeno-associated virus-mediated renin overexpression (ReninAAV) in uninephrectomized (UNx) db/db mice as a model of diabetic kidney disease, anti-GBM induced nephritis, and adriamycin-induced nephropathy) as well as 5/6 nephrectomy (Nx)-induced CKD in the rat.

Results

Our AI-based method accurately identify and quantify the severity of glomerulosclerosis (ranging from absence to advanced or global sclerosis, scored from 0 to 4) with exceptional sensitivity and specificity in murine and rat kidney sections. It showed promising efficacy in monitoring the progression of glomerulosclerosis not only in murine models (db/db UNx ReninAAV, anti-GBM and adriamycin mice), but also in nephrectomy 5/6-induced CKD model in rats.

Conclusion

Our AI-based glomerulosclerosis scoring method offers unbiased, accurate and automated glomerulosclerosis assessment in rodent models of CKD. The pipeline is optimized to accommodate both mouse and rat kidney sections, highlighting the applicability to the wide range of rodent models of CKD used in preclinical drug discovery.

Funding

  • Commercial Support – Gubra