1. Multiparametric MRI Tissue Characterization in Clinical Stroke With Correlation to Clinical Outcome
by Michael A. Jacobs, Panayiotis Mitsias, Hamid Soltanian-Zadeh, Sunitha Santhakumar, Amir Ghanei, Rabih Hammond, Donald J. Peck, Michael Chopp, and Suresh Patel
Stroke
Volume 32(4):
April 1, 2001
Copyright © American Heart Association, Inc. All rights reserved.

2. Representative coregistered and warped acute (≤12 hours) inputted MR data used for the ISODATA algorithm.
Representative coregistered and warped acute (≤12 hours) inputted MR data used for the ISODATA algorithm. A, Proton density–weighted image (TE=30 ms); B, T2WI (TE=90 ms); C, T1WI; D and E, DWI (D, b=600 s/mm2; E, b=1000 s/mm2); F, T2 map (arrow depicts old infarction on left frontal lobe near inferior frontal gyrus; and G, ADC map (2 arrows show location of new areas of ischemic tissue).
Michael A. Jacobs et al. Stroke. 2001;32:
Copyright © American Heart Association, Inc. All rights reserved.

3. Demonstration of tissue signature vectors constructed for different tissue types on MR data from a representative patient.
Demonstration of tissue signature vectors constructed for different tissue types on MR data from a representative patient. A tissue signature vector is defined for each separate tissue class and assigned to the cluster that closely resembles its vector elements in the ISODATA algorithm. The tissue signature vector is a mathematical descriptor for a tissue class. The tissue signature vector is defined as \mathbf|<||<|\FINLINE-FIGLINK LOCATOR="cjs1164"/INLINE-FIG/F|<|S|>||>||>|_|<|\mathbf|<|ij|>||>|\mathrm|<||<|=|>||<|[|>||>|\mathit|<|S|>|_|<|\mathrm|<|ij1|>||>|\mathrm|<|,|>|\mathit|<|S|>|_|<|\mathrm|<|ij2|>||>|\mathrm|<| |<|\ldots|>| S|>|_|<|\mathrm|<|ijk|>||>|\mathrm|<||<|]|>||>|^|<|\mathrm|<|T|>||>|\mathrm|<|,|>| where Sijk is the k-th element of the ij-th pixel vector, ie, the mean gray level of the ij-th pixel in the k-th image in the image sequence for each cluster, and T is the transpose of the matrix.
Michael A. Jacobs et al. Stroke. 2001;32:
Copyright © American Heart Association, Inc. All rights reserved.

4. Demonstration of the multiparametric ISODATA segmentation methodology and corresponding DWI (b=1000 s/mm2), ADC map, and T2 map at different times after stroke.
Demonstration of the multiparametric ISODATA segmentation methodology and corresponding DWI (b=1000 s/mm2), ADC map, and T2 map at different times after stroke. A 63-year-old woman presented with a right hemispheric infarct extending from the frontal to the occipital lobe, and she had no therapeutic intervention. The multiparametric ISODATA approach clearly segments the lesion into different components, which are not obvious from the DWI and ADC and T2 maps. See Results for detailed description. CSF indicates cerebrospinal fluid.
Michael A. Jacobs et al. Stroke. 2001;32:
Copyright © American Heart Association, Inc. All rights reserved.

5. Illustration of the multiparametric (2 T2WI, 1 T1WI, and 2 DWI) ISODATA segmentation and corresponding MR images at different time points after ictus in a 68-year-old woman after a left hemispheric ischemic stroke with recombinant tissue plasminogen activator treatment.
Illustration of the multiparametric (2 T2WI, 1 T1WI, and 2 DWI) ISODATA segmentation and corresponding MR images at different time points after ictus in a 68-year-old woman after a left hemispheric ischemic stroke with recombinant tissue plasminogen activator treatment. Acutely (<8 hours), the ISODATA segmentation defines a larger homogeneous region of tissue damage in the parietal lobe than defined by the DWI (b=1000 s/mm2) and the ADC map. Subtle signal intensity changes are noted on the T2 maps. However, subacutely (3 days), the multiparametric ISODATA clearly demonstrates new heterogeneous regions within the lesion extending into the frontal lobe and the corona radiata. These different areas of tissue damage are not visualized on the corresponding DWI and T2 maps, although there is evidence of heterogeneity on the ADC map. The discrimination of different regions of tissue damage demonstrates the power of the ISODATA technique to segment the ischemic area, as shown at the chronic time point. At the chronic (3 months) time point, the ISODATA segmentation clearly separates the ischemic infarct into several regions that are not as well defined from DWI and ADC and T2 maps. CSF indicates cerebrospinal fluid.
Michael A. Jacobs et al. Stroke. 2001;32:
Copyright © American Heart Association, Inc. All rights reserved.