Back to Search
Start Over
Novel magnetic resonance technique for characterizing mesoscale structure of trabecular bone
- Source :
- Royal Society open science, vol 5, iss 8, Royal Society Open Science, Vol 5, Iss 8 (2018)
- Publication Year :
- 2018
-
Abstract
- Osteoporosis, characterized by increased fracture risk and bone fragility, impacts millions of adults worldwide, but effective, non-invasive and easily accessible diagnostic tests of the disease remain elusive. We present a magnetic resonance (MR) technique that overcomes the motion limitations of traditional MR imaging to acquire high-resolution frequency-domain data to characterize the texture of biological tissues. This technique does not involve obtaining full two-dimensional or three-dimensional images, but can probe scales down to the order of 40 μm and in particular uncover structural information in trabecular bone. Using micro-computed tomography data of vertebral trabecular bone, we computationally validate this MR technique by simulating MR measurements of a ‘ratio metric’ determined from a few k -space values corresponding to trabecular thickness and spacing. We train a support vector machine classifier on ratio metric values determined from healthy and simulated osteoporotic bone data, which we use to accurately classify osteoporotic bone.
- Subjects :
- 0301 basic medicine
Aging
Computer science
Osteoporosis
030209 endocrinology & metabolism
Bioengineering
histomorphometry
magnetic resonance
03 medical and health sciences
0302 clinical medicine
medicine
Magnetic resonance technique
lcsh:Science
Multidisciplinary
medicine.diagnostic_test
Magnetic resonance imaging
medicine.disease
Mr imaging
osteoporosis
Trabecular bone
030104 developmental biology
trabecular bone
Musculoskeletal
Metric (mathematics)
Osteoporotic bone
Biomedical Imaging
lcsh:Q
Tomography
Biomedical engineering
Subjects
Details
- ISSN :
- 20545703
- Volume :
- 5
- Issue :
- 8
- Database :
- OpenAIRE
- Journal :
- Royal Society open science
- Accession number :
- edsair.doi.dedup.....8b4598223fc623c4bb610acb9810ab6a