DOI: 10.1148/radiol.2373041630
Reproducibility of Functional MR Imaging: Preliminary Results of Prospective Multi-institutional Study Performed by Biomedical Informatics Research Network1
Kelly H. Zou, PhD,
Douglas N. Greve, PhD,
Meng Wang, MSE,
Steven D. Pieper, PhD,
Simon K. Warfield, PhD,
Nathan S. White, BS,
Sanjay Manandhar, MS,
Gregory G. Brown, PhD,
Mark G. Vangel, PhD,
Ron Kikinis, MD,
William M. Wells, III, PhD, For the FIRST BIRN Research Group
1 From the Surgical Planning Laboratory, Dept of Radiology, Brigham and Women's Hosp (K.H.Z., M.W., S.D.P., S.K.W., S.M., R.K., W.M.W.); Dept of Health Care Policy (K.H.Z.); and Athinoula A. Martinos Center for Biomedical Imaging, Dept of Radiology, Massachusetts Gen Hosp (D.N.G., N.S.W., M.G.V.), Harvard Medical School, 75 Francis St, L-2, Boston, MA 02115; Isomics, Cambridge, Mass (S.D.P.); Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, Mass (S.K.W., W.M.W.); Computational Radiology Laboratory, Dept of Radiology, Brigham and Women's Hosp, Boston, Mass (S.K.W.); Dept of Radiology, Children's Hosp, Boston, Mass (S.K.W.); Laboratory of Cognitive Imaging, Dept of Psychiatry, Univ of California, San Diego, La Jolla, Calif (G.G.B.); and Veterans Affairs San Diego Health Care System, San Diego, Calif (G.G.B.). Received Sep 21, 2004; revision requested Nov 29; revision received Jan 24, 2005; accepted Feb 24. The BIRN study at two sites supported by NIH grants NCRR P41RR13218 and P41RR14075. Supported in part by NIH grants R01LM007861-01A1, R03HS013234-01, R21MH67054, and R21CA89449-01.
Address correspondence to K.H.Z. (e-mail: zou{at}bwh.harvard.edu).
View larger version (100K):
[in a new window]
|
Figure 1a. Subject 3. Registered MR imaging activation maps of anatomic data obtained during visit 1 at the study sites with (a) 1.5-T, (b) 3.0-T, and (c) 4.0-T MR imaging units.
|
|
View larger version (99K):
[in a new window]
|
Figure 1b. Subject 3. Registered MR imaging activation maps of anatomic data obtained during visit 1 at the study sites with (a) 1.5-T, (b) 3.0-T, and (c) 4.0-T MR imaging units.
|
|
View larger version (100K):
[in a new window]
|
Figure 1c. Subject 3. Registered MR imaging activation maps of anatomic data obtained during visit 1 at the study sites with (a) 1.5-T, (b) 3.0-T, and (c) 4.0-T MR imaging units.
|
|
View larger version (20K):
[in a new window]
|
Figure 2a. (a) Mean activation fractions (x-axis values) in left (left) and right (right) hemispheres at 1.5-, 3.0-, and 4.0-T MR imaging, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm. (b) Mean sensitivities (x-axis values) of 1.5-, 3.0-, and 4.0-T MR imaging in left (left) and right (right) hemispheres, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm. (c) Mean specificities (x-axis values) of 1.5-, 3.0-, and 4.0-T MR imaging in left (left) and right (right) hemispheres, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm.
|
|
View larger version (23K):
[in a new window]
|
Figure 2b. (a) Mean activation fractions (x-axis values) in left (left) and right (right) hemispheres at 1.5-, 3.0-, and 4.0-T MR imaging, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm. (b) Mean sensitivities (x-axis values) of 1.5-, 3.0-, and 4.0-T MR imaging in left (left) and right (right) hemispheres, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm. (c) Mean specificities (x-axis values) of 1.5-, 3.0-, and 4.0-T MR imaging in left (left) and right (right) hemispheres, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm.
|
|
View larger version (26K):
[in a new window]
|
Figure 2c. (a) Mean activation fractions (x-axis values) in left (left) and right (right) hemispheres at 1.5-, 3.0-, and 4.0-T MR imaging, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm. (b) Mean sensitivities (x-axis values) of 1.5-, 3.0-, and 4.0-T MR imaging in left (left) and right (right) hemispheres, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm. (c) Mean specificities (x-axis values) of 1.5-, 3.0-, and 4.0-T MR imaging in left (left) and right (right) hemispheres, derived according to subject (y-axis) and field strength by using step 1 in the expectation-maximization algorithm.
|
|
View larger version (24K):
[in a new window]
|
Figure 3a. ROC curves for left (LH) and right (RH) hemispheres in subject 2 at (a) 1.5-T and (b) 3.0-T MR imaging, derived according to examination visit. Solid lines represent visit 1 and dashed lines represent visit 2. FPR = false-positive rate, TPR = true-positive rate. (c) ROC curves for subject 2 at 4.0-T MR imaging.
|
|
View larger version (20K):
[in a new window]
|
Figure 3b. ROC curves for left (LH) and right (RH) hemispheres in subject 2 at (a) 1.5-T and (b) 3.0-T MR imaging, derived according to examination visit. Solid lines represent visit 1 and dashed lines represent visit 2. FPR = false-positive rate, TPR = true-positive rate. (c) ROC curves for subject 2 at 4.0-T MR imaging.
|
|
View larger version (16K):
[in a new window]
|
Figure 3c. ROC curves for left (LH) and right (RH) hemispheres in subject 2 at (a) 1.5-T and (b) 3.0-T MR imaging, derived according to examination visit. Solid lines represent visit 1 and dashed lines represent visit 2. FPR = false-positive rate, TPR = true-positive rate. (c) ROC curves for subject 2 at 4.0-T MR imaging.
|
|
Copyright © 2005 by the Radiological Society of North America.
