From Picoscale Pathology to Decascale Disease: Image Registration with a Scattering Transform and Varifolds for Manipulating Multiscale Data

Stouffer, Kaitlin; Wang, Zhenzhen; Xu, Eileen; Lee, Karl; Lee, Paige; Miller, Michael I.; Tward, Daniel J.

doi:10.1007/978-3-030-89847-2_1

Cited by 2 publications

(2 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As described previously [52, 27], we used a convolutional neural network to model and predict probabilities of being part of a tau tangle for each pixel in a digital histology image. To capture larger contextual features as well as local information for producing per pixel probabilities at high resolutions, we trained UNETs [53] with the architecture described in Table 1.…”

Section: Methodsmentioning

confidence: 99%

Projective LDDMM: Mapping Molecular Digital Pathology with Tissue MRI

Stouffer

Tward

Miller

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Reconstructing dense 3D anatomical coordinates from 2D projective measurements has become a central problem in digital pathology for both animal models and human studies. We describe a new family of diffeomorphic mapping technologies called Projective LDDMM which generate diffeomorphic mappings of dense human MRI atlases at tissue scales onto sparse measurements at micron scales associated with histological and more general optical imaging modalities. We solve the problem of dense mapping surjectively onto histological sections by incorporating new technologies for crossing modalities that use non-linear scattering transforms to represent multiple radiomic-like textures at micron scales and incorporating a Gaussian mixture-model frame-work for modelling tears and distortions associated to each section. We highlight the significance of our method through incorporation of neuropathological measures and MRI, as relevant to the development of biomarkers for Alzheimer’s disease and one instance of the integration of imaging data across the scales of clinical imaging and digital pathology.

show abstract

Section: Methodsmentioning

confidence: 99%

Projective LDDMM: Mapping Molecular Digital Pathology with Tissue MRI

Stouffer

Tward

Miller

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…As described previously [62,36], we used a convolutional neural network to model and predict probabilities of being part of a tau tangle for each pixel in a digital histology image. To capture larger contextual features as well as local information for producing per pixel probabilities at high resolutions, we trained UNETs [63] with the architecture described in Table F2 (see Appendix F).…”

Section: Density Determination Of Nftsmentioning

confidence: 99%

Projective LDDMM: Spatially Reconstructing a Story of Rostrally-Dominant Tau in Alzheimer’s Disease

Stouffer

Chen

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Since Braak's initial histological observations, it has been recognized that Alzheimer's disease (AD) neurofibrillary tangles (NFTs) appear in the medial temporal lobe (MTL) of the brain very early in the disease course. MRI-based shape diffeomorphometry markers have demonstrated preclinical AD changes in the MTL but it has not been possible to confirm that these MRI changes correspond to the presence of NFTs. Here, we present a method termed Projective LDDMM for aligning sparse measurement profiles of AD pathology (i.e., 2D digital histology images) with 3D MRI. We reconstruct measures of 2D NFT density in the dense metric of 3D MRI, using the Mai Paxinos Atlas coordinates for two cases of advanced AD. Analyses reveal the highest levels of NFT density in the rostral third (10-15 mm) of the hippocampus and the adjoining regions of the entorhinal cortex and amygdala. These findings emphasize the selective vulnerability of MTL subregions in AD, and suggest that high resolution MRI methods might benefit from focusing on the rostral MTL to more closely link these MRI images to AD neuropathology.

show abstract

From Picoscale Pathology to Decascale Disease: Image Registration with a Scattering Transform and Varifolds for Manipulating Multiscale Data

Cited by 2 publications

References 23 publications

Projective LDDMM: Mapping Molecular Digital Pathology with Tissue MRI

Projective LDDMM: Mapping Molecular Digital Pathology with Tissue MRI

Projective LDDMM: Spatially Reconstructing a Story of Rostrally-Dominant Tau in Alzheimer’s Disease

Contact Info

Product

Resources

About