Quantitative and textural analysis of magnetization transfer and diffusion images in the early detection of brain metastases
Authors
Ainsworth, Nicola L
Honess, Davina J
Brown, Anna M
Harden, Susan V
Publication Date
2016-04-18Journal Title
Magnetic Resonance in Medicine
ISSN
0740-3194
Publisher
Wiley
Volume
77
Pages
1987-1995
Language
English
Type
Article
This Version
VoR
Metadata
Show full item recordCitation
Ainsworth, N. L., McLean, M., McIntyre, D., Honess, D. J., Brown, A. M., Harden, S. V., & Griffiths, J. (2016). Quantitative and textural analysis of magnetization transfer and diffusion images in the early detection of brain metastases. Magnetic Resonance in Medicine, 77 1987-1995. https://doi.org/10.1002/mrm.26257
Abstract
Purpose: The sensitivity of the magnetization transfer ratio (MTR) and apparent diffusion coefficient (ADC) for early detection of brain metastases was investigated in mice and humans.
Methods: Mice underwent MRI twice weekly for up to 31 days following intra-cardiac injection of the brain-homing breast cancer cell line MDA-MB231-BR. Patients with small cell lung cancer underwent quarterly MRI for a year. MTR and ADC were measured in regions of metastasis and matched contralateral tissue at the final time-point and in registered regions at earlier time-points. Texture analysis and linear discriminant analysis were performed to detect metastasis-containing slices.
Results: Compared with contralateral tissue, mouse metastases had significantly lower MTR and higher ADC at the final time-point. Some lesions were visible at earlier time-points on the MTR and ADC maps: 24% of these were not visible on corresponding T2-weighted images. Texture analysis using the MTR maps showed 100% specificity and 98% sensitivity for metastasis at the final time-point, with 77% sensitivity 2-4 days earlier and 46% 5-8 days earlier. Only 2/16 patients developed metastases, and their penultimate scans were normal.
Conclusion: Some brain metastases may be detected earlier on MTR than conventional T2; however, the small gain is unlikely to justify ‘predictive’ MRI.
Keywords
MRI, cancer, magnetization transfer, diffusion, brain metastasis, texture analysis
Sponsorship
The authors gratefully acknowledge the Cambridge Institute Biological Resources Unit for expert animal care and technical assistance, the Histopathology Core Facility, Drs Joe Frank and Diane Palmieri for providing the cell line, the advice of Dr. Dan Tozer, and the support of Cancer Research UK [grant number C14303/A17197], the Brian Cross Memorial Trust, the Addenbrooke’s Charitable Trust, the University of Cambridge, Hutchison Whampoa Ltd, the Cambridge Experimental Cancer Medicine Centre, and the NIHR Cambridge Biomedical Research Centre.
Funder references
Cancer Research UK (C14303_do not transfer)
Cancer Research UK (CB4110)
?Brian Cross Memorial Trust (GF3079)
Embargo Lift Date
2100-01-01
Identifiers
External DOI: https://doi.org/10.1002/mrm.26257
This record's URL: https://www.repository.cam.ac.uk/handle/1810/255113
Rights
Attribution 4.0 International, Attribution 4.0 International, Attribution 4.0 International, Attribution 4.0 International