Sign-up for our newsletter
MAIN
Event Calendar
Awardee Reports
ABOUT DIACOMP
Citing DiaComp
Contact
Committees
Institutions
Awardee Reports
Publications
Bioinformatics
RESOURCES
Protocols & Methods
Reagents & Resources
Mouse Diet
Breeding Schemes
Validation Criteria
IMPC / KOMP Data
Publications
Bioinformatics
CONTACT
PARTICIPANT AREA
Login
▹
Publications
▹
Home
Publication
Spatially Aware Transformer Networks for Contextual Prediction of Diabetic
Nephropathy Progression from Whole Slide Images.
Authors
Shickel B, Lucarelli N, Rao AS, Yun D, Moon KC, Han SS, Sarder P
Submitted By
Submitted Externally on 10/16/2023
Status
Published
Journal
medRxiv : the preprint server for health sciences
Year
2023
Date Published
2/1/2023
Volume : Pages
Not Specified
:
Not Specified
PubMed Reference
36865174
Abstract
Diabetic nephropathy (DN) in the context of type 2 diabetes is the leading cause
of end-stage renal disease (ESRD) in the United States. DN is graded based on
glomerular morphology and has a spatially heterogeneous presentation in kidney
biopsies that complicates pathologists’ predictions of disease progression.
Artificial intelligence and deep learning methods for pathology have shown
promise for quantitative pathological evaluation and clinical trajectory
estimation; but, they often fail to capture large-scale spatial anatomy and
relationships found in whole slide images (WSIs). In this study, we present a
transformer-based, multi-stage ESRD prediction framework built upon nonlinear
dimensionality reduction, relative Euclidean pixel distance embeddings between
every pair of observable glomeruli, and a corresponding spatial self-attention
mechanism for a robust contextual representation. We developed a deep
transformer network for encoding WSI and predicting future ESRD using a dataset
of 56 kidney biopsy WSIs from DN patients at Seoul National University Hospital.
Using a leave-one-out cross-validation scheme, our modified transformer
framework outperformed RNNs, XGBoost, and logistic regression baseline models,
and resulted in an area under the receiver operating characteristic curve (AUC)
of 0.97 (95% CI: 0.90-1.00) for predicting two-year ESRD, compared with an AUC
of 0.86 (95% CI: 0.66-0.99) without our relative distance embedding, and an AUC
of 0.76 (95% CI: 0.59-0.92) without a denoising autoencoder module. While the
variability and generalizability induced by smaller sample sizes are
challenging, our distance-based embedding approach and overfitting mitigation
techniques yielded results that suggest opportunities for future spatially aware
WSI research using limited pathology datasets.
Complications
All Complications
Bioinformatics
Bone
Cardiomyopathy
Cardiovascular
Gastro-Intestinal (GI)
Nephropathy
Neuropathy & Neurocognition
Pediatric Endocrinology
Retinopathy
Uropathy
Wound Healing
Welcome to the DiaComp Login / Account Request Page.
Email Address:
Password:
Note: Passwords are case-sensitive.
Please save my Email Address on this machine.
Not a member?
If you are a funded DiaComp investigator, a member of an investigator's lab,
or an External Scientific Panel member to the consortium, please
request an account.
Forgot your password?
Enter your Email Address and
click here.
ERROR!
There was a problem with the page:
User Info
User Confirm
Please acknowledge all posters, manuscripts or scientific materials that were generated in part or whole using funds from the Diabetic Complications Consortium(DiaComp) using the following text:
Financial support for this work provided by the NIDDK Diabetic Complications Consortium (RRID:SCR_001415, www.diacomp.org), grants DK076169 and DK115255
Citation text and image have been copied to your clipboard. You may now paste them into your document. Thank you!