Investigating the state-of-the-art in whole-body MR-based attenuation correction: an intra-individual,inter-system,inventory study on three clinical PET/MR systems

Objective

We assess inter- and intra-subject variability of magnetic resonance (MR)-based attenuation maps (MRμMaps) of human subjects for state-of-the-art positron emission tomography (PET)/MR imaging systems.

Materials and methods

Four healthy male subjects underwent repeated MR imaging with a Siemens Biograph mMR, Philips Ingenuity TF and GE SIGNA PET/MR system using product-specific MR sequences and image processing algorithms for generating MRμMaps. Total lung volumes and mean attenuation values in nine thoracic reference regions were calculated. Linear regression was used for comparing lung volumes on MRμMaps. Intra- and inter-system variability was investigated using a mixed effects model.

Results

Intra-system variability was seen for the lung volume of some subjects, (p = 0.29). Mean attenuation values across subjects were significantly different (p < 0.001) due to different segmentations of the trachea. Differences in the attenuation values caused noticeable intra-individual and inter-system differences that translated into a subsequent bias of the corrected PET activity values, as verified by independent simulations.

Conclusion

Significant differences of MRμMaps generated for the same subjects but different PET/MR systems resulted in differences in attenuation correction factors, particularly in the thorax. These differences currently limit the quantitative use of PET/MR in multi-center imaging studies.

     

A novel alternative to classify tissues from <Emphasis Type="Italic">T</Emphasis><Subscript>1</Subscript> and <Emphasis Type="Italic">T</Emphasis><Subscript>2</Subscript> relaxation times for prostate MRI

Objective

To segment and classify the different attenuation regions from MRI at the pelvis level using the T ₁ and T ₂ relaxation times and anatomical knowledge as a first step towards the creation of PET/MR attenuation maps.

Materials and methods

Relaxation times were calculated by fitting the pixel-wise intensities of acquired T ₁- and T ₂-weighted images from eight men with inversion-recovery and multi-echo multi-slice spin-echo sequences. A decision binary tree based on relaxation times was implemented to segment and classify fat, muscle, prostate, and air (within the body). Connected component analysis and an anatomical knowledge-based procedure were implemented to localize the background and bone.

Results

Relaxation times at 3 T are reported for fat (T ₁ = 385 ms, T ₂ = 121 ms), muscle (T ₁ = 1295 ms, T ₂ = 40 ms), and prostate (T ₁ = 1700 ms, T ₂ = 80 ms). The relaxation times allowed the segmentation–classification of fat, prostate, muscle, and air, and combined with anatomical knowledge, they allowed classification of bone. The good segmentation–classification of prostate [mean Dice similarity score (mDSC) = 0.70] suggests a viable implementation in oncology and that of fat (mDSC = 0.99), muscle (mDSC = 0.99), and bone (mDSCs = 0.78) advocates for its implementation in PET/MR attenuation correction.

Conclusion

Our method allows the segmentation and classification of the attenuation-relevant structures required for the generation of the attenuation map of PET/MR systems in prostate imaging: air, background, bone, fat, muscle, and prostate.

     

Segmentation and characterization of interscapular brown adipose tissue in rats by multi-parametric magnetic resonance imaging

Objective

The aim was to auto-segment and characterize brown adipose, white adipose and muscle tissues in rats by multi-parametric magnetic resonance imaging with validation by histology and UCP1.

Materials and methods

Male Wistar rats were randomized into two groups for thermoneutral (n = 8) and cold exposure (n = 8) interventions, and quantitative MRI was performed longitudinally at 7 and 11 weeks. Prior to imaging, rats were maintained at either thermoneutral body temperature (36 ± 0.5 °C), or short term cold exposure (26 ± 0.5 °C). Neural network based automatic segmentation was performed on multi-parametric images including fat fraction, T ₂ and T ₂* maps. Isolated tissues were subjected to histology and UCP1 analysis.

Results

Multi-parametric approach showed precise delineation of the interscapular brown adipose tissue (iBAT), white adipose tissue (WAT) and muscle regions. Neural network based segmentation results were compared with manually drawn regions of interest, and showed 96.6 and 97.1 % accuracy for WAT and BAT respectively. Longitudinal assessment of the iBAT volumes showed a reduction at 11 weeks of age compared to 7 weeks. The cold exposed group showed increased iBAT volume compared to thermoneutral group at both 7 and 11 weeks. Histology and UCP1 expression analysis supported our imaging results.

Conclusion

Multi-parametric MR based neural network auto-segmentation provides accurate separation of BAT, WAT and muscle tissues in the interscapular region. The cold exposure improves the classification and quantification of heterogeneous BAT.

     

Automatic gallbladder segmentation using combined 2D and 3D shape features to perform volumetric analysis in native and secretin-enhanced MRCP sequences

Objectives

We aimed to develop the first fully automated 3D gallbladder segmentation approach to perform volumetric analysis in volume data of magnetic resonance (MR) cholangiopancreatography (MRCP) sequences. Volumetric gallbladder analysis is performed for non-contrast-enhanced and secretin-enhanced MRCP sequences.

Materials and methods

Native and secretin-enhanced MRCP volume data were produced with a 1.5-T MR system. Images of coronal maximum intensity projections (MIP) are used to automatically compute 2D characteristic shape features of the gallbladder in the MIP images. A gallbladder shape space is generated to derive 3D gallbladder shape features, which are then combined with 2D gallbladder shape features in a support vector machine approach to detect gallbladder regions in MRCP volume data. A region-based level set approach is used for fine segmentation. Volumetric analysis is performed for both sequences to calculate gallbladder volume differences between both sequences.

Results

The approach presented achieves segmentation results with mean Dice coefficients of 0.917 in non-contrast-enhanced sequences and 0.904 in secretin-enhanced sequences.

Conclusion

This is the first approach developed to detect and segment gallbladders in MR-based volume data automatically in both sequences. It can be used to perform gallbladder volume determination in epidemiological studies and to detect abnormal gallbladder volumes or shapes. The positive volume differences between both sequences may indicate the quantity of the pancreatobiliary reflux.

     

Cardiac magnetic resonance T1 and extracellular volume mapping with motion correction and co-registration based on fast elastic image registration

Objective

Our aim was to investigate the technical feasibility of a novel motion compensation method for cardiac magntic resonance (MR) T1 and extracellular volume fraction (ECV) mapping.

Materials and methods

Native and post-contrast T1 maps were obtained using modified look-locker inversion recovery (MOLLI) pulse sequences with acquisition scheme defined in seconds. A nonrigid, nonparametric, fast elastic registration method was applied to generate motion-corrected T1 maps and subsequently ECV maps. Qualitative rating was performed based on T1 fitting-error maps and overlay images. Local deformation vector fields were produced for quantitative assessment. Intra- and inter-observer reproducibility were compared with and without motion compensation.

Results

Eighty-two T1 and 39 ECV maps were obtained in 21 patients with diverse myocardial diseases. Approximately 60% demonstrated clear quality improvement after motion correction for T1 mapping, particularly for the poor-rating cases (23% before vs 2% after). Approximately 67% showed further improvement with co-registration in ECV mapping. Although T1 and ECV values were not clinically significantly different before and after motion compensation, there was improved intra- and inter-observer reproducibility after motion compensation.

Conclusions

Automated motion correction and co-registration improved the qualitative assessment and reproducibility of cardiac MR T1 and ECV measurements, allowing for more reliable ECV mapping.

     

A method for the automatic segmentation of brown adipose tissue

Objective

Brown adipose tissue (BAT) plays a key role for thermogenesis in mammals and infants. Recent confirmation of BAT presence in adult humans has aroused great interest for its potential to initiate weight-loss and normalize metabolic disorders in diabetes and obesity. Reliable detection and differentiation of BAT from the surrounding white adipose tissue (WAT) and muscle is critical for assessment/quantification of BAT volume. This study evaluates magnetic resonance (MR) acquisition for BAT and the efficacy of different automated methods for MR features-based BAT segmentation to identify the best suitable method.

Materials and methods

Multi-point Dixon and multi-echo T₂ spin-echo images were acquired from 12 mice using an Agilent 9.4T scanner. Four segmentation methods: multidimensional thresholding (MTh); region-growing (RG); fuzzy c-means (FCM) and neural-network (NNet) were evaluated for the interscapular region and validated against manually defined BAT, WAT and muscle.

Results

Statistical analysis of BAT segmentation yielded a median Dice-Statistical-Index, and sensitivity of 89. 92 % for NNet, 82. 86 % for FCM, 72. 74 % for RG, and 72. 70 %, for MTh, respectively.

Conclusion

This study demonstrates that NNet improves the specificity to BAT from surrounding tissue based on 3-point Dixon and T₂ MRI. This method facilitates quantification and longitudinal measurement of BAT in preclinical-models and human subjects.

     

Theory of MRI contrast in the annulus fibrosus of the intervertebral disc

Objective

Here we develop a three-dimensional analytic model for MR image contrast of collagen lamellae in the annulus fibrosus of the intervertebral disc of the spine, based on the dependence of the MRI signal on collagen fiber orientation.

Materials and methods

High-resolution MRI scans were performed at 1.5 and 7 T on intact whole disc specimens from ovine, bovine, and human spines. An analytic model that approximates the three-dimensional curvature of the disc lamellae was developed to explain inter-lamellar contrast and intensity variations in the annulus. The model is based on the known anisotropic dipolar relaxation of water in tissues with ordered collagen.

Results

Simulated MRI data were generated that reproduced many features of the actual MRI data. The calculated inter-lamellar image contrast demonstrated a strong dependence on the collagen fiber angle and on the circumferential location within the annulus.

Conclusion

This analytic model may be useful for interpreting MR images of the disc and for predicting experimental conditions that will optimize MR image contrast in the annulus fibrosus.

     

Endoluminal high-resolution MR imaging protocol for colon walls analysis in a mouse model of colitis

Objective

An endoluminal magnetic resonance (MR) imaging protocol including the design of an endoluminal coil (EC) was defined for high-spatial-resolution MR imaging of mice gastrointestinal walls at 4.7 T.

Materials and methods

A receive-only radiofrequency single-loop coil was developed for mice colon wall imaging. Combined with a specific protocol, the prototype was first characterized in vitro on phantoms and on vegetables. Signal-to-noise ratio (SNR) profiles were compared with a quadrature volume birdcage coil (QVBC). Endoluminal MR imaging protocol combined with the EC was assessed in vivo on mice.

Results

The SNR measured close to the coil is significantly higher (10 times and up to 3 mm of the EC center) than the SNR measured with the QVBC. The gain in SNR can be used to reduce the in-plane pixel size up to 39 × 39 µm² (234 µm slice thickness) without time penalty. The different colon wall layers can only be distinguished on images acquired with the EC.

Conclusion

Dedicated EC provides suitable images for the assessment of mice colon wall layers. This proof of concept provides gains in spatial resolution and leads to adequate protocols for the assessment of human colorectal cancer, and can now be used as a new imaging tool for a better understanding of the pathology.

     

A novel MR-compatible sensor to assess active medical device safety: stimulation monitoring,rectified radio frequency pulses,and gradient-induced voltage measurements

Purpose

To evaluate the function of an active implantable medical device (AIMD) during magnetic resonance imaging (MRI) scans. The induced voltages caused by the switching of magnetic field gradients and rectified radio frequency (RF) pulse were measured, along with the AIMD stimulations.

Materials and methods

An MRI-compatible voltage probe with a bandwidth of 0–40 kHz was designed. Measurements were carried out both on the bench with an overvoltage protection circuit commonly used for AIMD and with a pacemaker during MRI scans on a 1.5 T (64 MHz) MR scanner.

Results

The sensor exhibits a measurement range of?±?15 V with an amplitude resolution of 7 mV and a temporal resolution of 10 µs. Rectification was measured on the bench with the overvoltage protection circuit. Linear proportionality was confirmed between the induced voltage and the magnetic field gradient slew rate. The pacemaker pacing was recorded successfully during MRI scans.

Conclusion

The characteristics of this low-frequency voltage probe allow its use with extreme RF transmission power and magnetic field gradient positioning for MR safety test of AIMD during MRI scans.

     

Glioma vessel abnormality quantification using time-of-flight MR angiography

Objectives

To differentiate between abnormal tumor vessels and regular brain vasculature using new quantitative measures in time-of-flight (TOF) MR angiography (MRA) data.

Materials and methods

In this work time-of-flight (TOF) MR angiography data are acquired in 11 glioma patients to quantify vessel abnormality. Brain vessels are first segmented with a new algorithm, efficient monte-carlo image-analysis for the location of vascular entity (EMILOVE), and are then characterized in three brain regions: tumor, normal-appearing contralateral brain, and the total brain volume without the tumor. For characterization local vessel orientation angles and the dot product between local orientation vectors are calculated and averaged in the 3 regions. Additionally, correlation with histological and genetic markers is performed.

Results

Both the local vessel orientation angles and the dot product show a statistically significant difference (p < 0.005) between tumor vessels and normal brain vasculature. Furthermore, the connection to both histology and the gene expression of the tumor can be found—here, the measures were compared to the proliferation marker Ki-67 [MIB] and genome-wide expression analysis. The results in a subgroup indicate that the dot product measure may be correlated with activated genetic pathways.

Conclusion

It is possible to define a measure of vessel abnormality based on local vessel orientation angles which can differentiate between normal brain vasculature and glioblastoma vessels.

     

Multi-atlas-based fully automatic segmentation of individual muscles in rat leg

Objective

To quantify individual muscle volume in rat leg MR images using a fully automatic multi-atlas-based segmentation method.

Materials and methods

We optimized a multi-atlas-based segmentation method to take into account the voxel anisotropy of numbers of MRI acquisition protocols. We mainly tested an image upsampling process along Z and a constraint on the nonlinear deformation in the XY plane. We also evaluated a weighted vote procedure and an original implementation of an artificial atlas addition. Using this approach, we measured gastrocnemius and plantaris muscle volumes and compared the results with manual segmentation. The method reliability for volume quantification was evaluated using the relative overlap index.

Results

The most accurate segmentation was obtained using a nonlinear registration constrained in the XY plane by zeroing the Z component of the displacement and a weighted vote procedure for both muscles regardless of the number of atlases. The performance of the automatic segmentation and the corresponding volume quantification outperformed the interoperator variability using a minimum of three original atlases.

Conclusion

We demonstrated the reliability of a multi-atlas segmentation approach for the automatic segmentation and volume quantification of individual muscles in rat leg and found that constraining the registration in plane significantly improved the results.

     

Quantitative T<Subscript>1</Subscript> and T<Subscript>2</Subscript> MRI signal characteristics in the human brain: different patterns of MR contrasts in normal ageing

Objective

The objective of this study was to examine age-dependent changes in both T₁-weighted and T₂-weighted image contrasts and spin-echo T₂ relaxation time in the human brain during healthy ageing.

Methods

A total of 37 participants between the ages of 49 and 87 years old were scanned with a 3 Tesla system, using T₁-weighted, T₂ weighted and quantitative spin-echo T₂ imaging. Contrast between image intensities and T₂ values was calculated for various regions, including between individual hippocampal subfields.

Results

The T₁ contrast-to-noise (CNR) and gray:white signal intensity ratio (GWR) did not change in the hippocampus, but it declined in the cingulate cortex with age. In contrast, T₂ CNR and GWR declined in both brain regions. T₂ relaxation time was almost constant in gray matter and most (but not all) hippocampal subfields, but increased substantially in white matter, pointing to an age effect on water relaxation in white matter.

Conclusions

Changes in T₁ and T₂ MR characteristics influence the appearance of brain images in later life and should be considered in image analyses of aged subjects. It is speculated that alterations at the cell biology level, with concomitant alterations to the local magnetic environment, reduce dephasing and subsequently prolong spin-echo T₂ through reduced diffusion effects in later life.

     

Reproducibility of locus coeruleus and substantia nigra imaging with neuromelanin sensitive MRI

Objectives

The purpose of this study was to assess the reproducibility of substantia nigra pars compacta (SNpc) and locus coeruleus (LC) delineation and measurement with neuromelanin-sensitive MRI.

Materials and methods

Eleven subjects underwent two neuromelanin-sensitive MRI scans. SNpc and LC volumes were extracted for each scan. Reproducibility of volume and magnetization transfer contrast measurements in SNpc and LC was assessed using intraclass correlation coefficients (ICC) and dice similarity coefficients (DSC).

Results

SNpc and LC volume measurements showed excellent reproducibility (SNpc-ICC: 0.94, p < 0.001; LC-ICC: 0.96, p < 0.001). SNpc and LC were accurately delineated between scans (SNpc-DSC: 0.80 ± 0.03; LC-DSC: 0.63 ± 0.07).

Conclusion

Neuromelanin-sensitive MRI can consistently delineate SNpc and LC.

     

Validation of an active shape model-based semi-automated segmentation algorithm for the analysis of thigh muscle and adipose tissue cross-sectional areas

Objective

To validate a semi-automated method for thigh muscle and adipose tissue cross-sectional area (CSA) segmentation from MRI.

Materials and methods

An active shape model (ASM) was trained using 113 MRI CSAs from the Osteoarthritis Initiative (OAI) and combined with an active contour model and thresholding-based post-processing steps. This method was applied to 20 other MRIs from the OAI and to baseline and follow-up MRIs from a 12-week lower-limb strengthening or endurance training intervention (n = 35 females). The agreement of semi-automated vs. previous manual segmentation was assessed using the Dice similarity coefficient and Bland-Altman analyses. Longitudinal changes observed in the training intervention were compared between semi-automated and manual segmentations.

Results

High agreement was observed between manual and semi-automated segmentations for subcutaneous fat, quadriceps and hamstring CSAs. With strength training, both the semi-automated and manual segmentation method detected a significant reduction in adipose tissue CSA and a significant gain in quadriceps, hamstring and adductor CSAs. With endurance training, a significant reduction in adipose tissue CSAs was observed with both methods.

Conclusion

The semi-automated approach showed high agreement with manual segmentation of thigh muscle and adipose tissue CSAs and showed longitudinal training effects similar to that observed using manual segmentation.

     

Anti-EpCAM scFv gadolinium chelate: a novel targeted MRI contrast agent for imaging of colorectal cancer

Objectives

The development of targeted contrast agents for magnetic resonance imaging (MRI) facilitates enhanced cancer imaging and more accurate diagnosis. In the present study, a novel contrast agent was developed by conjugating anti-EpCAM humanized scFv with gadolinium chelate to achieve target specificity.

Materials and methods

The material design strategy involved site-specific conjugation of the chelating agent to scFv. The scFv monomer was linked to maleimide-DTPA via unpaired cysteine at the scFv C-terminus, followed by chelation with gadolinium (Gd). Successful scFv-DTPA conjugation was achieved at 1:10 molar ratio of scFv to maleimide-DTPA at pH 6.5. The developed anti-EpCAM-Gd-DTPA MRI contrast agent was evaluated for cell targeting ability, in vitro serum stability, cell cytotoxicity, relaxivity, and MR contrast enhancement.

Results

A high level of targeting efficacy of anti-EpCAM-Gd-DTPA to an EpCAM-overexpressing HT29 colorectal cell was demonstrated by confocal microscopy. Good stability of the contrast agent was obtained and no cytotoxicity was observed in HT29 cells after 48 h incubation with 25–100 µM of Gd. Favorable imaging was obtained using anti-EpCAM-Gd-DTPA, including 1.8-fold enhanced relaxivity compared with Gd-DTPA, and MR contrast enhancement observed after binding to HT29.

Conclusion

The potential benefit of this contrast agent for in vivo MR imaging of colorectal cancer, as well as other EpCAM positive cancers, is suggested and warrants further investigation.

     

Assessment of the myelin water fraction in rodent spinal cord using T2-prepared ultrashort echo time MRI

Objective

Multi-component T2 relaxation allows for assessing the myelin water fraction in nervous tissue, providing a surrogate marker for demyelination. The assessment of the number and distribution of different T2 components for devising exact models of tissue relaxation has been limited by T2 sampling with conventional MR methods.

Materials and methods

A T2-prepared UTE sequence was used to assess multicomponent T2 relaxation at 9.4 T of fixed mouse and rat spinal cord samples and of mouse spinal cord in vivo. For in vivo scans, a cryogenically cooled probe allowed for 78-µm resolution in 1-mm slices. Voxel-wise non-negative least square analysis was used to assess the number of myelin water-associated T2 components.

Results

More than one myelin water-associated T2 component was detected in only 12 % of analyzed voxels in rat spinal cords and 6 % in mouse spinal cords, both in vivo and in vitro. However, myelin water-associated T2 values of individual voxels varied between 0.1 and 20 ms. While in fixed samples almost no components below 1 ms were identified, in vivo, these contributed 14 % of the T2 spectrum. No significant differences in MWF were observed in mouse spinal cord in vivo versus ex vivo measurements.

Conclusion

Voxel-wise analysis methods using relaxation models with one myelin water-associated T2 component are appropriate for assessing myelin content of nervous tissue.

     

Semi-automated myocardial segmentation of bright blood multi-gradient echo images improves reproducibility of myocardial contours and T2* determination

Objectives

Early detection of iron loading is affected by the reproducibility of myocardial contour assessment. A novel semi-automatic myocardial segmentation method is presented on contrast-optimized composite images and compared to the results of manual drawing.

Materials and methods

Fifty-one short-axis slices at basal, mid-ventricular and apical locations from 17 patients were acquired by bright blood multi-gradient echo MRI. Four observers produced semi-automatic and manual myocardial contours on contrast-optimized composite images. The semi-automatic segmentation method relies on vector field convolution active contours to generate the endocardial contour. After creating radial pixel clusters on the myocardial wall, a combination of pixel-wise coefficient of variance (CoV) assessment and k-means clustering establishes the epicardial contour for each segment.

Results

Compared to manual drawing, semi-automatic myocardial segmentation lowers the variability of T2* quantification within and between observers (CoV of 12.05 vs. 13.86% and 14.43 vs. 16.01%) by improving contour reproducibility (P < 0.001). In the presence of iron loading, semi-automatic segmentation also lowers the T2* variability within and between observers (CoV of 13.14 vs. 15.19% and 15.91 vs. 17.28%).

Conclusion

Application of semi-automatic myocardial segmentation on contrast-optimized composite images improves the reproducibility of T2* quantification.

     

Vessel radius mapping in an extended model of transverse relaxation

Objectives

Spin dephasing of the local magnetization in blood vessel networks can be described in the static dephasing regime (where diffusion effects may be ignored) by the established model of Yablonskiy and Haacke. However, for small capillary radii, diffusion phenomena for spin-bearing particles are not negligible.

Material and methods

In this work, we include diffusion effects for a set of randomly distributed capillaries and provide analytical expressions for the transverse relaxation times T2* and T2 in the strong collision approximation and the Gaussian approximation that relate MR signal properties with microstructural parameters such as the mean local capillary radius.

Results

Theoretical results are numerically validated with random walk simulations and are used to calculate capillary radius distribution maps for glioblastoma mouse brains at 9.4 T. For representative tumor regions, the capillary maps reveal a relative increase of mean radius for tumor tissue towards healthy brain tissue of \(128 \pm 23 \%\) (p < 0.001).

Conclusion

The presented method may be used to quantify angiogenesis or the effects of antiangiogenic therapy in tumors whose growth is associated with significant microvascular changes.

     

Scan–rescan reproducibility of segmental aortic wall shear stress as assessed by phase-specific segmentation with 4D flow MRI in healthy volunteers

Objective

The aim was to investigate scan–rescan reproducibility and observer variability of segmental aortic 3D systolic wall shear stress (WSS) by phase-specific segmentation with 4D flow MRI in healthy volunteers.

Materials and methods

Ten healthy volunteers (age 26.5?±?2.6 years) underwent aortic 4D flow MRI twice. Maximum 3D systolic WSS (WSSmax) and mean 3D systolic WSS (WSSmean) for five thoracic aortic segments over five systolic cardiac phases by phase-specific segmentations were calculated. Scan–rescan analysis and observer reproducibility analysis were performed.

Results

Scan–rescan data showed overall good reproducibility for WSSmean (coefficient of variation, COV 10–15%) with moderate-to-strong intraclass correlation coefficient (ICC 0.63–0.89). The variability in WSSmax was high (COV 16–31%) with moderate-to-good ICC (0.55–0.79) for different aortic segments. Intra- and interobserver reproducibility was good-to-excellent for regional aortic WSSmax (ICC?≥?0.78; COV?≤?17%) and strong-to-excellent for WSSmean (ICC?≥?0.86; COV?≤?11%). In general, ascending aortic segments showed more WSSmax/WSSmean variability compared to aortic arch or descending aortic segments for scan–rescan, intraobserver and interobserver comparison.

Conclusions

Scan–rescan reproducibility was good for WSSmean and moderate for WSSmax for all thoracic aortic segments over multiple systolic phases in healthy volunteers. Intra/interobserver reproducibility for segmental WSS assessment was good-to-excellent. Variability of WSSmax is higher and should be taken into account in case of individual follow-up or in comparative rest–stress studies to avoid misinterpretation.

     

Reproducible segmentation of white matter hyperintensities using a new statistical definition

Objectives

We present a method based on a proposed statistical definition of white matter hyperintensities (WMH), which can work with any combination of conventional magnetic resonance (MR) sequences without depending on manually delineated samples.

Materials and methods

T1-weighted, T2-weighted, FLAIR, and PD sequences acquired at 1.5 Tesla from 119 subjects from the Kings Health Partners-Dementia Case Register (healthy controls, mild cognitive impairment, Alzheimer’s disease) were used. The segmentation was performed using a proposed definition for WMH based on the one-tailed Kolmogorov–Smirnov test.

Results

The presented method was verified, given all possible combinations of input sequences, against manual segmentations and a high similarity (Dice 0.85–0.91) was observed. Comparing segmentations with different input sequences to one another also yielded a high similarity (Dice 0.83–0.94) that exceeded intra-rater similarity (Dice 0.75–0.91). We compared the results with those of other available methods and showed that the segmentation based on the proposed definition has better accuracy and reproducibility in the test dataset used.

Conclusion

Overall, the presented definition is shown to produce accurate results with higher reproducibility than manual delineation. This approach can be an alternative to other manual or automatic methods not only because of its accuracy, but also due to its good reproducibility.