[R01] Distributed foundational models for multi-task learning in diabetic retinopathy
Ente: National Eye Institute
Scadenza: 2030-03-31
Importo max: 617.244 EUR
Paese: US
Descrizione
Abstract: This project aims to establish distributed federated learning (FL) approaches for multi-task training of
foundational machine learning (ML) models for diabetic retinopathy (DR), using multi-modal, real-world optical
coherence tomography (OCT) data (OCT cross-section, OCT angiography (OCTA), and OCT enface). DR is one
of the leading causes of severe vision loss. Early detection, prompt intervention, and reliable assessment of
treatment outcomes are essential to prevent irreversible vision loss from DR. However, there are major
challenges towards developing clinically relevant holistic algorithms that can perform multi-tasks, i.e., multi-class
classification of disease stages (diagnosis), prediction of onset and progression of disease stages (prognosis),
and assessment of treatment outcomes. They require large amounts of well curated and labelled datasets from
a diverse sub-population for robust performance. Moreover, efforts towards large, centralized datasets for ML
research are hindered by significant barriers to data sharing and privacy concerns. In this project, we propose to
develop foundational ML models that allow efficient learning of feature representations from a large corpus of
ophthalmic imaging data for various downstream tasks – breaking the task-specific paradigm of current ML
models. We also establish novel federated ML approaches, where the model training is distributed across
institutions instead of sharing patient data. Our first aim is to establish and validate a domain adaptive FL
framework for DR diagnosis across four independent institutions. We propose a novel ophthalmic adaptive
personalized FL (optho-APFL) technique to tackle domain shift caused by heterogeneous data distribution at
different institutions (due to different sub-population density and OCT devices/imaging protocols). We will
conduct experiments on the FL deployment in a clinical setting and integrate a granular differential privacy (DP)
algorithm into our FL framework to provide ‘patient-level’ data privacy. Key success criterion is to deploy the FL
framework and validate FL-trained ML models against state-of-the-art models for DR diagnosis. The second aim
is to develop foundational ML models with self-supervised learning (SSL) to learn multiple tasks within the same
framework from label invariant OCT/OCTA data, where different institutions don’t need to have labeled data for
each of the tasks. We will train these foundational models in a centralized and FL framework for comparative
analysis. Key success criterion is to i) validate foundational model performance for multi-task learning (MTL) (DR
staging, prediction of NPDR to PDR progression, and prediction of DME treatment evaluation) on new clinical
data (centralized and FL approach, and ii) identify task-specific quantitative OCT/OCTA (mean and artery-vein
specific) features. As an alternative approach, we propose diffusion probabilistic modeling (DPM) for SSL to
learn holistic representations f
Istituzione: UNIVERSITY OF NORTH CAROLINA CHARLOTTE
PI: Minhaj Nur Alam
Progetto: 1R01EY037828-01
Settori: National Eye Institute
Vai al bando originale
Registrati gratis su Bandolo per trovare bandi compatibili con la tua azienda.