Segmentation of multiple sclerosis lesions in MR images: a review

Introduction  

Multiple sclerosis (MS) is an inflammatory demyelinating disease that the parts of the nervous system through the lesions generated in the white matter of the brain. It brings about disabilities in different organs of the body such as eyes and muscles. Early detection of MS and estimation of its progression are critical for optimal treatment of the disease.     

慢性多发性硬化病灶:高场MRI特征

目的阐明多发性硬化(MS)病灶在磁敏感加权成像MR影像对比的机制以及评估铁及髓磷脂对产生MR影像对比的组织关联性。方法每例病人均提供了机构审查委员会批准的个人受试者协议的书面同意书。21例MS病人进行了     

Evolution of multiple sclerosis lesions on serial contrast-enhanced T1-weighted and magnetization-transfer MR images

BACKGROUND AND PURPOSE: Magnetization-transfer imaging is a technique that could provide indirect evidence of the characteristics of multiple sclerosis (MS) lesions. The purpose of this work was to study the evolution of MS lesions on T1-weighted MR images over time and to investigate changes in magnetization-transfer ratio (MTR) values of MS lesions with different initial appearances on contrast-enhanced T1-weighted images. METHODS: Eleven patients with relapsing-remitting MS were studied with MR imaging. The MTRs were calculated for 47 lesions that had been classified according to their appearance on contrast-enhanced T1-weighted images. Each patient was examined at four time points over a 1-year period. The MTR changes observed in the selected lesions were compared with their initial T1-weighted appearance. RESULTS: The lowest MTR values were initially found in hypointense nonenhancing lesions and in ring-enhancing lesions, with both types showing a hypointense center. Changes in MTR values were more dynamic and reversible in ring-enhancing than in hypointense nonenhancing plaques. Nodular-enhancing lesions had slightly lower initial MTRs than did isointense non-enhancing lesions. CONCLUSION: The absence or presence of contrast uptake may indicate a different pathologic basis for hypointense MS lesions on T1-weighted MR images. These differences should be kept in mind when considering T1 lesion load as a surrogate marker of disability in MS.     

Artificial multiple sclerosis lesions on simulated FLAIR brain MR images: echo time and observer performance in detection

The institutional review board approved the described HIPAA-compliant study, which was performed to prospectively evaluate observer performance in the detection of artificial multiple sclerosis (MS) lesions that were randomly distributed supra- and infratentorially on simulated fluid-attenuated inversion-recovery (FLAIR) magnetic resonance (MR) images obtained at different echo times (TEs). MR parametric maps were derived from mixed multi-echo inversion-recovery images obtained in a 40-year-old healthy male volunteer and in a patient with MS, both of whom gave informed consent. Pseudo-randomly distributed artificial MS lesions of varying size, number, and location were equally represented on the FLAIR images (11 000/100-200/2600 [repetition time msec/TE msec/inversion time msec]). Twelve images obtained in both regions at each of 11 TEs spaced 10 msec apart were rated by seven neuroradiologists by using a four-point scale. Observer performance in the detection of MS lesions on the FLAIR images, as estimated by using areas under the alternative free-response receiver operating characteristic curve, was highest and most consistent at the 100-msec TE, both supratentorially (93.0% +/- 8.6 [standard error of the mean]) and infratentorially (87.4% +/- 10.0).     

Corpus callosum and subcallosal-periventricular lesions in multiple sclerosis: detection with MR

Examination with magnetic resonance imaging of 40 patients with confirmed diagnoses of multiple sclerosis showed that corpus callosum involvement is common. Thirty percent of the patients had focal callosal lesions similar to those described in the pathology literature. Long, inner callosal-subcallosal lesions were found in 55% of patients. These lesions had signal characteristics similar to those of noncallosal periventricular lesions. Diffuse moderate to severe atrophy of the corpus callosum was noted in 40% of patients, with one exception concurrent with inner callosal lesions. The nature of the inner callosal lesions is not known, since these lesions are not typically described in the literature. These lesions may represent demyelination or increased water content and may be the precursor to atrophy that progresses from the ependymal surface toward the outer fibers of the corpus callosum.     

MR multispectral analysis of multiple sclerosis lesions

Although quantification of the lesion burden from serial MR examinations of patients with multiple sclerosis (MS) is a common technique to assess disease activity in clinical trials, pathologic change may occur within a lesion without a corresponding change in volume. Therefore, measures of lesion volume and composition may improve the sensitivity of detecting disease activity. A new technique has been developed that provides information about the intensity composition of MS lesions in standard spin-echo MR examinations. The new technique is based on the multispectral “feature space” intensity distributions of the lesions and normal tissues. Analysis of MR examinations of materials with known T1 and T2 times showed that feature space position from spin-echo examinations is largely determined from proton density (ρ), T2, and the interecho delay. Information about intensity composition was obtained by reducing the multidimensional intensity distribution to one dimension while minimizing the loss of information. This technique was used to analyze eight lesions in standard spin-echo MR examinations of three patients with MS. Lesion distributions were compared between examinations by first calibrating the examinations based on the intensity distributions of cerebrospinal fluid (CSF), an internal reference tissue. Many of the lesion distributions had a distinctive peak at low intensity, corresponding to normal-appearing white matter (WM). Within the lesion distributions, increases in high intensity peaks generally were accompanied by reductions in the WM peak. Serial analysis of the lesion distributions revealed some dramatic fluctuations, even when lesion volume remained constant.     

Volume measurement of multiple sclerosis lesions with magnetic resonance images

Summary The ability to visualise multiple selerosis lesions in vivo with magnetic resonance imaging suggests and important role in monitoring the course of the disease. In order to help the long-term assessment of prospective treatments, a semi-automated technique for measuring lesion volume has been developed to provide a quantitative index of disease progression. Results are presented from a preliminary study with a single patient and compared to measurements taken from lesion outlines traced by a neuroradiologist, two neurologists and a technician. The semi-automated technique achieved a precision of 6% compared to a range of 12–33% for the manual tracing method. It also reduced the human interaction time from at least 60 min to 15 min.     

Quantitative MR in the diagnosis of multiple sclerosis.

In patients with multiple sclerosis (MS), the apparently uninvolved cerebral white matter between demyelinated plaques may have biochemical abnormalities. To what degree the changes in the white matter contribute to symptomatology in MS is unknown. In 39 patients with multiple sclerosis, and in 39 age-matched nondiseased volunteers, T1 and T2 were calculated from spin-echo images in four regions of apparently uninvolved white matter. In three of four white matter areas, the average T1 and T2 were significantly longer in the patients than in the controls. The T1 correlated with the disability, measured by the Kurtzke Extended Disability Status Scale, although the correlation was marginally significant. The results suggest that in patients with MS, white matter disease that is not visualized in MR as distinct foci of abnormal signal intensity may contribute to disease burden and disability.     

Demyelinating plaques in relapsing-remitting and secondary-progressive multiple sclerosis: assessment with diffusion MR imaging

BACKGROUND AND PURPOSE: Conventional MR imaging does not provide specific information that can be reliably associated with the pathologic substrate and clinical status of patients with multiple sclerosis (MS). Our goals were 1) to determine whether the orientationally averaged water diffusion coefficient () can be used to distinguish between plaques of different severity in these patients and 2) to assess possible correlations between values and disease duration, Expanded Disability Status Scale (EDSS) score, and signal intensity on T1-weighted MR images. METHODS: Twenty patients (10 with relapsing-remitting MS and 10 with secondary-progressive MS) and 11 healthy volunteers underwent a combined conventional and diffusion-weighted MR study of the brain. , a parameter that is proportional to the trace of the diffusion tensor, was computed by averaging the apparent diffusion coefficients measured in the x, y, and z directions. measurements were obtained for selected areas of white matter plaques. Differences in among the three groups were tested using analysis of variance. RESULTS: was significantly higher (1.445 +/- 0.129 x 10(-3) mm2/s) in secondary-progressive lesions than in relapsing-remitting lesions (0.951 +/- 0.08), and both values were higher than in normal white matter (0.732 +/- 0.02). There was a significant negative correlation between and the degree of hypointensity on T1-weighted images, and a positive correlation between and both EDSS score and disease duration. CONCLUSION: Our findings suggest that is useful for distinguishing MS lesions of different severities, which are associated with different degrees of clinical disability.     

The evolution of multiple sclerosis lesions on serial MR.

PURPOSETo characterize temporal changes in signal intensity patterns of multiple sclerosis lesions on serial MR.METHODST1-, T2-, proton density-, and contrast-enhanced T1-weighted MR was performed on five patients with relapsing-remitting multiple sclerosis at least 22 times in the course of 1 year.RESULTSForty-three enhancing lesions and 1 new lesion that never showed enhancement were detected and followed for periods ranging from approximately 4 weeks to 1 year (total of 702 time points). At first detection the center of new lesions was brighter than the periphery (20 of 24 new lesions on proton density-weighted and 19 of 23 new lesions on contrast-enhanced images). On contrast-enhanced images, ring hyperintensity was predominant at time points later than 29 days. As lesions aged, a residual rim of "nonenhancing" hyperintensity often was noted on contrast-enhanced images. Some older lesions (> 1 year) showed similar appearance on unenhanced T1-weighted images. On proton density-weighted images ring hyperintensity was most frequent 2 to 4 months after lesion detection. The estimated average duration of gadopentetate dimeglumine enhancement was 1 to 2 months.CONCLUSIONSA lesion evolution pattern relevant to MR was inferred. We believe that specific information about the histopathologic evolution of a lesion may be extracted not only from contrast-enhanced but also from nonenhanced serial MR. Assessment of drugs targeting specific phases of lesion evolution could benefit from quantitative pattern analysis of routine MR images.     

Intraobserver and interobserver variability in schemes for estimating volume of brain lesions on MR images in multiple sclerosis.

PURPOSEOur goal was to evaluate the intraobserver and interobserver reproducibility of measurements of brain lesion load in multiple sclerosis (MS) by using two proposed acquisition schemes.METHODSThree-millimeter-thick conventional spin-echo (CSE) and fast fluid-attenuated inversion-recovery (FLAIR) sequences were obtained and the lesions segmented using a semiautomated technique based on local thresholding to calculate intraobserver and interobserver reproducibility. These were compared with images obtained from two separate MR units in which 5-mm CSE sequences were obtained and segmented by using the local thresholding technique and also by manual outlining.RESULTSThe intraobserver coefficient of variation was 4.0% (95% confidence interval [CI], 3.0% to 4.5%) for the 5-mm CSE sequence measured with manual outlining, 3.1% (95% CI, 2.5% to 3.2%) and 5.1% (95% CI, 4.1% to 5.6%) for the two sets of 5-mm CSE sequences measured using the local thresholding technique, 5.7% (95% CI, 3.9% to 6.6%) for the 3-mm CSE sequence, and 2.6% (95% CI, 2.1% to 2.7%) for the fast FLAIR sequence. The interobserver coefficient of variation was 7.1% (95% CI, 4.9% to 8.7%) and 8.3% (95% CI, 6.4% to 9.6%) for the two sets of 5-mm CSE sequences, 7.3% (95% CI, 4.7% to 9.1%) for the 3-mm CSE sequence, and 2.9% (95% CI, 2.3% to 3.3%) for the fast FLAIR sequence. The intraobserver and interobserver reproducibility of measurements obtained with the fast FLAIR technique was significantly better than those obtained with the other techniques.CONCLUSIONSOur data indicate that the intraobserver and interobserver variability in quantifying MS lesions can be reduced significantly with the use of fast FLAIR sequences, while no significant improvement is gained by reducing the section thickness from 5 mm to 3 mm.     

Progressive multifocal leukoencephalopathy in patients with AIDS: appearance on MR images

Progressive multifocal leukoencephalopathy (PML) is an uncommon demyelinating disease that occurs in immunocompromised patients. The authors evaluated magnetic resonance (MR) images of 10 patients with pathologically proved PML and clinically diagnosed acquired immunodeficiency syndrome (AIDS) to determine the MR characteristics of this disorder. All patients had asymmetric cerebral involvement. Lesions were distributed throughout the brain, including the brain stem and basal ganglia. White matter was affected in all 10 patients; gray matter was also involved in five. In one patient the lesion enlarged and crossed the corpus callosum and contained focal hemorrhage. The authors conclude that, contrary to reported findings on computed tomographic scans, PML in patients with AIDS has a variable appearance on MR images and has many characteristics that differ from those previously thought to be typical on imaging studies.     

Multiple sclerosis: Hyperintense lesions in the brain on T1‐weighted MR images assessed by diffusion tensor imaging

Purpose:

To evaluate retrospectively quantitative diffusion tensor imaging (DTI) values of hyperintense lesions on nonenhanced T1‐weighted magnetic resonance (MR) images in patients with multiple sclerosis (MS) to elucidate the degree of demyelination or remyelination associated with T1 hyperintense lesions and study their relationship to MR markers of tissue damage (brain atrophy).

Materials and Methods:

Institutional review board approval was obtained; informed consent was waived for this HIPAA‐compliant study, including 76 patients with MS and 20 healthy control subjects without evidence of MS clinically or on imaging. T1 lesions were compared with normal white matter on nonenhanced images and judged to be hyperintense. Quantitative DTI metrics of T1 hyperintense lesions were examined, and the relationship between DTI parameters and brain atrophy were investigated in this study.

Results:

At least one T1 hyperintense lesion was found in 16 patients (total, 28 lesions). Hyperintense lesions on T1‐weighted imaging (T1WI) had lower mean diffusion (MD) than others signal intensity lesions on T1WI but higher MD than normal white matter (F = 3.931; P < 0.001); Fractional anisotropy (FA; F = 3.24; P < 0.001) and volume ratio (VR; F = 1.664; P < 0.001) were higher in hyperintense lesions on T1WI than hypointense/isointense lesions on T1WI, but were lower than normal‐appearing white matter (NAWM) and normal white matter in controls. There was correlation between FA and VR (r = 0.678; P < 0.001) and inverse correlation between FA and MD (r = ?0.437; P = 0.02), MD and VR (r = ?0.423; P 0.025) for T1 hyperintense lesion. The MD values of T1 hyperintense lesions (r = ?0.304; P < 0.001) and the VR values of T1 hyperintense lesions (r = 0.096; P = 0.042) were significantly (negative) correlated with Brain parenchymal fraction (BPF; higher BPF score); the FA values of T1 hyperintense lesions (r = ?0.111; P = 0.018), the MD values of T1 hyperintense lesions (r = 0.379; P < 0.001) and the VR values of T1 hyperintense lesions (r = ?0.142; P = 0.003) were significantly correlated with third ventricular width (lower width). However, the FA value of T1 hyperintense lesions was not significantly associated with BPF(r = 0.083; P = 0.08).

Conclusion:

The quantitative DTI values of T1 hyperintense MS plaques were between hypo‐/isointense lesions and NAWM or normal white matter, and correlated with BPF and third ventricular width. Our results supports the notion that axonal remyelination may be the reason for T1 hyperintense lesions. J. Magn. Reson. Imaging 2010;31:789–795. ©2010 Wiley‐Liss, Inc.

     

Cortical lesions in multiple sclerosis: combined postmortem MR imaging and histopathology

BACKGROUND AND PURPOSE: Cortical lesions constitute a substantial part of the total lesion load in multiple sclerosis (MS) brain. They have been related to neuropsychological deficits, epilepsy, and depression. However, the proportion of purely cortical lesions visible on MR images is unknown. The aim of this study was to determine the proportion of intracortical and mixed gray matter (GM)-white matter (WM) lesions that can be visualized with postmortem MR imaging. METHODS: We studied 49 brain samples from nine cases of chronic MS. Tissue sections were matched to dual-echo T2-weighted spin-echo (T2SE) MR images. MS lesions were identified by means of myelin basic protein immunostaining, and lesions were classified as intracortical, mixed GM-WM, deep GM, or WM. Investigators blinded to the histopathologic results scored postmortem T2SE and 3D fluid-attenuated inversion recovery (FLAIR) images. RESULTS: Immunohistochemistry confirmed 70 WM, eight deep GM, 27 mixed GM-WM, and 63 purely cortical lesions. T2SE images depicted only 3% of the intracortical lesions, and 3D FLAIR imaging showed 5%. Mixed GM-WM lesions were most frequently detectable on T2SE and 3D FLAIR images (22% and 41%, respectively). T2SE imaging showed 13% of deep GM lesions versus 38% on 3D FLAIR. T2SE images depicted 63% of the WM lesions, whereas 3D FLAIR images depicted 71%. Even after side-by-side review of the MR imaging and histopathologic results, many of the intracortical lesions could not be identified retrospectively. CONCLUSION: In contrast to WM lesions and mixed GM-WM lesions, intracortical lesions remain largely undetected with current MR imaging resolution.     

In vivo MR imaging of hippocampal lesions in multiple sclerosis

PURPOSE: To investigate whether a recently improved version of the three-dimensional double inversion-recovery (3D-DIR) technique enables the in vivo detection of hippocampal lesions in multiple sclerosis (MS). MATERIALS AND METHODS: Magnetic resonance images of 16 patients and nine healthy control subjects were acquired at 1.5T. Lesions were scored on 3D-DIR images and were anatomically classified as white matter (WM), cortical, or hippocampal lesions. Associations between hippocampal, cortical, and WM lesion numbers were evaluated. Also, hippocampal lesions were retrospectively assessed on 3D-T2 and hippocampal and brain volumes were measured. RESULTS: No hippocampal lesions were detected in control subjects. By contrast, 14 out of 16 MS patients had at least one hippocampal lesion. The mean number (+/-SD) of hippocampal lesions detected with 3D-DIR was 2.6 +/- 1.8 in MS patients; only 56% of these lesions could be observed on 3D-T2. CONCLUSION: Hippocampal lesions can be visualized in vivo with 3D-DIR and occur frequently in MS. The ability to visualize hippocampal lesions in vivo is of fundamental importance to future studies focusing on the role of gray matter (GM) damage in cognitive deficits, which are common in MS.     

薄层双反转恢复MRI上Brodmann 4区三层结构的表现

目的 探讨3．0T的薄层轴向双反转恢复（DIR）MRI能否根据“三层结构”的显示与否区分初级运动皮质（PMC或Brodmann 4区）与其他所选皮质区域包括初级感觉皮质（PSC或Brodmann 1-3区）。方法 该前瞻性研究获机构审查委员会批准和病人知情同意。研究包括3．0T MRI表现正常的191例病人（女94例，年龄5～80岁；男97例，年龄5～76岁）。