Multiparametric MRI ISODATA Ischemic Lesion Analysis

Abstract

Background and Purpose-The purpose of this study was to show that the computer segmentation algorithm Iterative Self-Organizing Data Analysis Technique (ISODATA), which integrates multiple MRI parameters (diffusion-weighted imaging [DWI], T2-weighted imaging [T2WI], and T1-weighted imaging [T1WI]) into a single composite image, is capable of defining the ischemic lesion in a time-independent manner equally as well as the MRI techniques considered the best for each phase after stroke onset (ie, perfusion weighted imaging [PWI] and DWI for the acute phase and T2WI for the outcome phase). Methods-We measured MRI parameters of PWI, DWI, T2WI, and T1WI from patients at the acute phase (Ͻ30 hours) and DWI, T2WI, and T1WI at the outcome phase (3 months) of ischemic stroke. The clinical neurological deficit was graded with the National Institutes of Health Stroke Scale (NIHSS). We compared the ISODATA lesion size with the PWI, DWI, and T2WI lesion sizes measured within the same slice at each phase. The lesion sizes were also correlated with NIHSS score of each phase. Results-We included 11 patients; 9 (82%) were women, and 7 (64%) were black. The meanϮSD age was 65.5Ϯ9.3 years (range, 45 to 82 years). The median NIHSS score was 15 (minimum, 4; maximum, 24)at the acute phase and 3 (minimum, 0; maximum, 22) at the outcome phase. The median time interval from stroke symptom onset to the acute MRI study was 10 hours (range, 6 to 29 hours), and the mean time interval to the outcome study was 93Ϯ11 days (range, 72 to 106 days). In the acute phase, the ISODATA lesion size had high correlation with the PWI lesion size (rϭ0.95; 95% CI, 0.89 to 0.98; PϽ0.0001), DWI lesion size (rϭ0.83; 95% CI, 0.66 to 0.92; PϽ0.0001), and T2WI lesion size (rϭ0.67; 95% CI, 0.39 to 0.84; Pϭ0.008) and moderate correlation with NIHSS score (rϭ0.59; 95% CI, 0.02 to 0.88; Pϭ0.06). In the outcome phase, the ISODATA lesion size had high correlation with the T2WI lesion size (rϭ0.97; 95% CI, 0.94 to 0.99; PϽ0.0001) and NIHSS score (rϭ0.78; 95% CI, 0.34 to 0.94; Pϭ0.004). Conclusions-The integrated ISODATA method can identify and characterize the ischemic lesion independently of time elapsed since stroke onset. The ISODATA lesion size highly correlates with the PWI and DWI lesion size in the acute phase and with the T2WI lesion size in the outcome phase of ischemic stroke, as well as with the clinical neurological status of the patient.

Statistical Prediction of Tissue Fate in Acute Ischemic Brain Injury

“…Region-ofinterest analysis thereby inadvertently mixes the characteristics that one is trying to resolve. The complex temporal and spatial evolution of focal cerebral ischemia has prompted the use of more sophisticated analysis methods (Welch et al, 1995;Jiang et al, 1997;Carano et al, 1998Carano et al, , 2000Jacobs et al, 2000Jacobs et al, , 2001aWu et al, 2001;Mitsias et al, 2002) to stage stroke outcome. Multiparametric analysis using K-mean and Fuzzy c-mean clustering techniques have been used to classify ischemic tissue fate based on CBF index maps, T 2 and ADC maps in a rat stroke model (Carano et al, , 2000.…”

“…This approach requires the number of tissue clusters to be assigned a priori, which is generally unknown during the evolution of cerebral ischemia. Jacobs et al (2001a) and Mitsias et al (2002) eloquently incorporated the iterative self-organizing data analysis algorithm (ISODATA) (Ball and Hall, 1965) for analyzing T 1 -, T 2 -and diffusion-weighted images in human stroke and found that the multiparametric ISODATA analysis outperformed analysis using any single parameter alone. This approach has also been applied to analyze T 1 -, T 2 -and diffusion-weighted images in an animal stroke models during the subacute and chronic phase (Jacobs et al, 2001b).…”

Predicting final extent of ischemic infarction using an artificial neural network analysis of multiparametric MRI in patients with stroke

“…Figure 6 compares the results of the ISODATA technique with those of the ANN. As shown in this figure the lesion areas in the ISODATA maps are clustered to different signatures (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). Signatures range between one (corresponding to normal tissue) to twelve (corresponding to CSF or cavitated tissue).…”

“…Stop the gradient descent training at an appropriate point. 3. Add noise to the training patterns to smooth out the data points.…”

“…Consequently, during the acute phase of ischemic stroke, a fast and reliable identification of the ischemic lesion, its extent, and a prediction as to its fate, might aid clinical decision-making and help to maximize benefit and minimize side effects of a therapeutic intervention [5,6]. Many studies, by correlating with follow-up imaging or neurological status, have shown the potential for diffusion weighted imaging (DWI), perfusion weighted imaging (PWI), and/or T 2 -weighted (T2WI) magnetic resonance imaging (MRI) used together for staging stroke outcome [3,7,8]. Most of these analyses were performed using as an outcome metric the volume of a region-of-interest (ROI) defined by the difference between an acute data set and the outcome image [9,10,11].…”