TY - GEN
T1 - Addressing imaging accessibility by cross-modality transfer learning
AU - Zheng, Zhiyang
AU - Su, Yi
AU - Chen, Kewei
AU - Weidman, David A.
AU - Wu, Teresa
AU - Lo, Ben
AU - Lure, Fleming
AU - Li, Jing
N1 - Funding Information:
This work is not being, or has been, submitted for publication or presentation elsewhere. Research reported in this publication was supported by the National Institute on Aging of the National Institutes of Health under Award Number 2R42AG053149-02A1. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Publisher Copyright:
© 2022 SPIE.
PY - 2022
Y1 - 2022
N2 - Multi-modality images usually exist for diagnosis/prognosis of a disease, such as Alzheimer's Disease (AD), but with different levels of accessibility and accuracy. MRI is used in the standard of care, thus having high accessibility to patients. On the other hand, imaging of pathologic hallmarks of AD such as amyloid-PET and tau-PET has low accessibility due to cost and other practical constraints, even though they are expected to provide higher diagnostic/prognostic accuracy than standard clinical MRI. We proposed Cross-Modality Transfer Learning (CMTL) for accurate diagnosis/prognosis based on standard imaging modality with high accessibility (mod-HA), with a novel training strategy of using not only data of mod-HA but also knowledge transferred from the model based on advanced imaging modality with low accessibility (mod-LA). We applied CMTL to predict conversion of individuals with Mild Cognitive Impairment (MCI) to AD using the Alzheimer's Disease Neuroimaging Initiative (ADNI) datasets, demonstrating improved performance of the MRI (mod-HA)-based model by leveraging the knowledge transferred from the model based on tau-PET (mod-LA).
AB - Multi-modality images usually exist for diagnosis/prognosis of a disease, such as Alzheimer's Disease (AD), but with different levels of accessibility and accuracy. MRI is used in the standard of care, thus having high accessibility to patients. On the other hand, imaging of pathologic hallmarks of AD such as amyloid-PET and tau-PET has low accessibility due to cost and other practical constraints, even though they are expected to provide higher diagnostic/prognostic accuracy than standard clinical MRI. We proposed Cross-Modality Transfer Learning (CMTL) for accurate diagnosis/prognosis based on standard imaging modality with high accessibility (mod-HA), with a novel training strategy of using not only data of mod-HA but also knowledge transferred from the model based on advanced imaging modality with low accessibility (mod-LA). We applied CMTL to predict conversion of individuals with Mild Cognitive Impairment (MCI) to AD using the Alzheimer's Disease Neuroimaging Initiative (ADNI) datasets, demonstrating improved performance of the MRI (mod-HA)-based model by leveraging the knowledge transferred from the model based on tau-PET (mod-LA).
KW - Alzheimer s disease
KW - Knowledge distillation
KW - Mild cognitive impairment
KW - Multi-modality images
KW - Transfer learning
UR - http://www.scopus.com/inward/record.url?scp=85132823538&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85132823538&partnerID=8YFLogxK
U2 - 10.1117/12.2611791
DO - 10.1117/12.2611791
M3 - Conference contribution
AN - SCOPUS:85132823538
T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE
BT - Medical Imaging 2022
A2 - Drukker, Karen
A2 - Iftekharuddin, Khan M.
PB - SPIE
T2 - Medical Imaging 2022: Computer-Aided Diagnosis
Y2 - 21 March 2022 through 27 March 2022
ER -