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 Twitter

Kidney Week

Abstract: TH-PO566

Computational Assessment of Early Diabetic Nephropathy

Session Information

  • Pathology and Lab Medicine
    November 03, 2022 | Location: Exhibit Hall, Orange County Convention Center‚ West Building
    Abstract Time: 10:00 AM - 12:00 PM

Category: Pathology and Lab Medicine

  • 1700 Pathology and Lab Medicine

Authors

  • Claus, Lindsey, Boston University, Boston, Massachusetts, United States
  • Zhang, Yichi, Boston University, Boston, Massachusetts, United States
  • Zheng, Yi, Boston University, Boston, Massachusetts, United States
  • Surendran, Tejus, Boston University, Boston, Massachusetts, United States
  • Chitalia, Vipul C., Boston University, Boston, Massachusetts, United States
  • Walker, Patrick D., Arkana Laboratories, Little Rock, Arkansas, United States
  • Cassol, Clarissa Araujo, Arkana Laboratories, Little Rock, Arkansas, United States
  • Kolachalama, Vijaya B., Boston University, Boston, Massachusetts, United States
Background

Patients with diabetic nephropathy (DN) often have glomerular, tubular, interstitial, and vascular lesions on biopsy, though a recently proposed DN classification focuses on glomerular features. While it is important to quantify the extent of these histopathological manifestations across all stages of DN, we sought to investigate whether these findings can be identified in early DN patients.

Methods

A deep learning framework known as a feature pyramid network (FPN) was implemented to classify digitized renal biopsies as class I or II DN. PAS-stained whole slide images (WSIs) of cases with class I (n=23) or class II (n=77) DN were divided into training, validation, and test sets in 3:1:1 ratio. This process was repeated five times to achieve a 5-fold cross validation. The FPN expected two inputs, 224x224-pixel patches cropped from WSIs and 16x downsampled WSIs, to learn and combine features at two levels of magnification. Class activation maps (CAMs) were generated to visualize informative regions on the WSIs.

Results

Our FPN model achieved an accuracy of 0.8 and an F1 score of 0.6. On a representative set of test images, we found the majority (~84%) of patches identified by CAMs as highly informative for model prediction contained tubules or interstitium, not just glomerular regions (Fig. 1).

Conclusion

Our study identified several regions on the biopsy images as informative for prediction of class I vs II DN. Further analysis can elucidate the importance of various histopathological features of early stage DN.

(A and B) Whole slide images and their class labels (Class I & II DN). (C and D) Class activation maps generated on the images, indicating the regions highly associated with the corresponding label. The warmer the color, the higher the probability of a region contributing to the DN class prediction.

Funding

  • NIDDK Support