Metric to quantify white matter damage on brain magnetic resonance images

Automated brain segmentation and volumetry in dementia diagnostics: a narrative review with emphasis on FreeSurfer

Background

Dementia can be caused by numerous different diseases that present variable clinical courses and reveal multiple patterns of brain atrophy, making its accurate early diagnosis by conventional examinative means challenging. Although highly accurate and powerful, magnetic resonance imaging (MRI) currently plays only a supportive role in dementia diagnosis, largely due to the enormous volume and diversity of data it generates. AI-based software solutions/algorithms that can perform automated segmentation and volumetry analyses of MRI data are being increasingly used to address this issue. Numerous commercial and non-commercial software solutions for automated brain segmentation and volumetry exist, with FreeSurfer being the most frequently used.

Objectives

This Review is an account of the current situation regarding the application of automated brain segmentation and volumetry to dementia diagnosis.

Methods

We performed a PubMed search for “FreeSurfer AND Dementia” and obtained 493 results. Based on these search results, we conducted an in-depth source analysis to identify additional publications, software tools, and methods. Studies were analyzed for design, patient collective, and for statistical evaluation (mathematical methods, correlations).

Results

In the studies identified, the main diseases and cohorts represented were Alzheimer’s disease (n = 276), mild cognitive impairment (n = 157), frontotemporal dementia (n = 34), Parkinson’s disease (n = 29), dementia with Lewy bodies (n = 20), and healthy controls (n = 356). The findings and methods of a selection of the studies identified were summarized and discussed.

Conclusion

Our evaluation showed that, while a large number of studies and software solutions are available, many diseases are underrepresented in terms of their incidence. There is therefore plenty of scope for targeted research.

Imaging Biomarkers of VCI: A Focused Update

Vascular cognitive impairment is common after stroke, in memory clinics, medicine for the elderly services, and undiagnosed in the community. Vascular disease is said to be the second most common cause of dementia after Alzheimer disease, yet vascular dysfunction is now known to predate cognitive decline in Alzheimer disease, and most dementias at older ages are mixed. Neuroimaging has a major role in identifying the proportion of vascular versus other likely pathologies in patients with cognitive impairment. Here, we aim to provide a pragmatic but evidence-based summary of the current state of potential imaging biomarkers, focusing on magnetic resonance imaging and computed tomography, which are relevant to diagnosing, estimating prognosis, monitoring vascular cognitive impairment, and incorporating our own experiences. We focus on markers that are well-established, with a known profile of association with cognitive measures, but also consider more recently described, including quantitative tissue markers of vascular injury. We highlight the gaps in accessibility and translation to more routine clinical practice.

Paranasal sinuses opacification on magnetic resonance imaging in relation to brain health in sporadic small vessel disease - Systematic review and pilot analysis

BACKGROUND

The paranasal sinus mucosal thickening, visible in magnetic resonance imaging (MRI), maybe a source of inflammation in microvessels, but its relationship with small vessel disease (SVD) is unclear. We reviewed the literature and analysed a sample of patients with sporadic SVD to identify any association between paranasal sinus opacification severity and SVD neuroimaging markers.

METHODS

We systematically reviewed MEDLINE and EMBASE databases up to April 2020 for studies on paranasal sinus mucosal changes in patients with SVD, cerebrovascular disease (CVD), and age-related neurodegenerative diseases. We analysed clinical and MRI data from 100 participants in a prospective study, the Mild Stroke Study 3 (ISRCTN 12113543) at 1-3, 6 and 12 months following a minor stroke to test key outcomes from the literature review. We used multivariate linear regression to explore associations between modified Lund-Mackay (LM) scores and brain, white matter hyperintensities (WMH), enlarged perivascular spaces (PVS) volumes at each time point, adjusted for baseline age, sex, diabetes, hypercholesterolaemia, hypertension and smoking.

RESULTS

The literature review, after screening 3652 publications, yielded 11 primary studies, for qualitative synthesis with contradictory results, as positive associations/higher risk from 5/7 CVD studies were contradicted by the two studies with largest samples, and data from dementia studies was equally split in their outcome. From the pilot sample of patients analysed (female N = 33, mean age 67.42 (9.70) years), total LM scores had a borderline negative association with PVS in the centrum semiovale at baseline and 6 months (B = -0.25, SE = 0.14, p = 0.06) but were not associated with average brain tissue, WMH or normal-appearing white matter volumes.

CONCLUSION

The inconclusive results from the literature review and empirical study justify larger studies between PVS volume and paranasal sinuses opacification in patients with sporadic SVD.

Sleep quality, perivascular spaces and brain health markers in ageing - A longitudinal study in the Lothian Birth Cohort 1936

BACKGROUND

Sleep is thought to play a major role in brain health and general wellbeing. However, few longitudinal studies have explored the relationship between sleep habits and imaging markers of brain health, particularly markers of brain waste clearance such as perivascular spaces (PVS), of neurodegeneration such as brain atrophy, and of vascular disease, such as white matter hyperintensities (WMH). We explore these associations using data collected over 6 years from a birth cohort of older community-dwelling adults in their 70s.

METHOD

We analysed brain MRI data from ages 73, 76 and 79 years, and self-reported sleep duration, sleep quality and vascular risk factors from community-dwelling participants in the Lothian Birth Cohort 1936 (LBC1936) study. We calculated sleep efficiency (at age 76), quantified PVS burden (at age 73), and WMH and brain volumes (age 73 to 79), calculated the white matter damage metric, and used structural equation modelling (SEM) to explore associations and potential causative pathways between indicators related to brain waste cleaning (i.e., sleep and PVS burden), brain and WMH volume changes during the 8th decade of life.

RESULTS

Lower sleep efficiency was associated with a reduction in normal-appearing white matter (NAWM) volume (β = 0.204, P = 0.009) from ages 73 to 79, but not concurrent volume (i.e. age 76). Increased daytime sleep correlated with less night-time sleep (r = -0.20, P < 0.001), and with increasing white matter damage metric (β = -0.122, P = 0.018) and faster WMH growth (β = 0.116, P = 0.026). Shorter night-time sleep duration was associated with steeper 6-year reduction of NAWM volumes (β = 0.160, P = 0.011). High burden of PVS at age 73 (volume, count, and visual scores), was associated with faster deterioration in white matter: reduction of NAWM volume (β = -0.16, P = 0.012) and increasing white matter damage metric (β = 0.37, P < 0.001) between ages 73 and 79. On SEM, centrum semiovale PVS burden mediated 5% of the associations between sleep parameters and brain changes.

CONCLUSION

Sleep impairments, and higher PVS burden, a marker of impaired waste clearance, were associated with faster loss of healthy white matter and increasing WMH in the 8th decade of life. A small percentage of the effect of sleep in white matter health was mediated by the burden of PVS consistent with the proposed role for sleep in brain waste clearance.

Fluid-attenuated inversion recovery magnetic resonance imaging textural features as sensitive markers of white matter damage in midlife adults

White matter hyperintensities are common radiological findings in ageing and a typical manifestation of cerebral small vessel disease. White matter hyperintensity burden is evaluated by quantifying their volume; however, subtle changes in the white matter may not be captured by white matter hyperintensity volumetry. In this cross-sectional study, we investigated whether magnetic resonance imaging texture of both white matter hyperintensities and normal appearing white matter was associated with reaction time, white matter hyperintensity volume and dementia risk in a midlife cognitively normal population. Data from 183 cognitively healthy midlife adults from the PREVENT-Dementia study (mean age 51.9 ± 5.4; 70% females) were analysed. White matter hyperintensities were segmented from 3 Tesla fluid-attenuated inversion recovery scans using a semi-automated approach. The fluid-attenuated inversion recovery images were bias field corrected and textural features (intensity mean and standard deviation, contrast, energy, entropy, homogeneity) were calculated in white matter hyperintensities and normal appearing white matter based on generated textural maps. Textural features were analysed for associations with white matter hyperintensity volume, reaction time and the Cardiovascular Risk Factors, Aging and Dementia risk score using linear regression models adjusting for age and sex. The extent of normal appearing white matter surrounding white matter hyperintensities demonstrating similar textural associations to white matter hyperintensities was further investigated by defining layers surrounding white matter hyperintensities at increments of 0.86 mm thickness. Lower mean intensity within white matter hyperintensities was a significant predictor of longer reaction time (t = −3.77, P < 0.01). White matter hyperintensity volume was predicted by textural features within white matter hyperintensities and normal appearing white matter, albeit in opposite directions. A white matter area extending 2.5 – 3.5 mm further from the white matter hyperintensities demonstrated similar associations. White matter hyperintensity volume was not related to reaction time, although interaction analysis revealed that participants with high white matter hyperintensity burden and less homogeneous white matter hyperintensity texture demonstrated slower reaction time. Higher Cardiovascular Risk Factors, Aging, and Dementia score was associated with a heterogeneous normal appearing white matter intensity pattern. Overall, greater homogeneity within white matter hyperintensities and a more heterogeneous normal appearing white matter intensity profile were connected to a higher white matter hyperintensity burden, while heterogeneous intensity was related to prolonged reaction time (white matter hyperintensities of larger volume) and dementia risk (normal appearing white matter). Our results suggest that the quantified textural measures extracted from widely used clinical scans, might capture underlying microstructural damage and might be more sensitive to early pathological changes compared to white matter hyperintensity volumetry.

Accurate 3D Reconstruction of White Matter Hyperintensities Based on Attention-Unet

White matter hyperintensities (WMH), also known as white matter osteoporosis, have been clinically proven to be associated with cognitive decline, the risk of cerebral infarction, and dementia. The existing computer automatic measurement technology for the segmentation of patients' WMH does not have a good visualization and quantitative analysis. In this work, the author proposed a new WMH quantitative analysis and 3D reconstruction method for 3D reconstruction of high signal in white matter. At first, the author using ResUnet achieves the high signal segmentation of white matter and adds the attention mechanism into ResUnet to achieve more accurate segmentation. Afterwards, this paper used surface rendering to reconstruct the accurate segmentation results in 3D. Data experiments are conducted on the dataset collected from Shandong Province Third Hospital. After training, the Attention-Unet proposed in this paper is superior to other segmentation models in the segmentation of high signal in white matter and Dice coefficient and MPA reached 92.52% and 92.43%, respectively, thus achieving accurate 3D reconstruction and providing a new idea for quantitative analysis and 3D reconstruction of WMH.

Development of a protocol to assess within-subject, regional white matter hyperintensity changes in aging and dementia

BACKGROUND

White matter hyperintensities (WMH), associated with both dementia risk and progression, can individually progress, remain stable, or even regress influencing cognitive decline related to specific cerebrovascular-risks. This study details the development and validation of a registration protocol to assess regional, within-subject, longitudinal WMH changes (ΔWMH) that is currently lacking in the field.

NEW METHOD

3D-FLAIR images (baseline and one-year-visit) were used for protocol development and validation. The method was validated by assessing the correlation between forward and reverse longitudinal registration, and between summated regional progression-regression volumes and Global ΔWMH. The clinical relevance of growth-regression ΔWMH were explored in relation to an executive function test.

RESULTS

MRI scans for 79 participants (73.5 ± 8.8 years) were used in this study. Global ΔWMH vs. summated regional progression-regression volumes were highly associated (r2 = 0.90; p-value < 0.001). Bi-directional registration validated the registration method (r2 = 0.999; p-value < 0.001). Growth and regression, but not overall ΔWMH, were associated with one-year declines in performance on Trial-Making-Test-B.

COMPARISON WITH EXISTING METHOD(S)

This method presents a unique registration protocol for maximum tissue alignment, demonstrating three distinct patterns of longitudinal within-subject ΔWMH (stable, growth and regression).

CONCLUSIONS

These data detail the development and validation of a registration protocol for use in assessing within-subject, voxel-level alterations in WMH volume. The methods developed for registration and intensity correction of longitudinal within-subject FLAIR images allow regional and within-lesion characterization of longitudinal ΔWMH. Assessing the impact of associated cerebrovascular-risks and longitudinal clinical changes in relation to dynamic regional ΔWMH is needed in future studies.

Changes of Brain Structure in Patients With Metastatic Non-Small Cell Lung Cancer After Long-Term Target Therapy With EGFR-TKI

Purpose

Epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI) therapy is the routine treatment for patients with metastatic non-small cell lung cancer (NSCLC) harboring positive EGFR mutations. Patients who undergo such treatment have reported cognitive decline during follow-up. This study, therefore, aimed to evaluate brain structural changes in patients receiving EGFR-TKI to increase understanding of this potential symptom.

Method

The medical records of 75 patients with metastatic NSCLC (without brain metastasis or other co-morbidities) who received EGFR-TKI therapy from 2010 to 2017 were reviewed. The modified Scheltens Visual Scale and voxel-based morphometry were used to evaluate changes in white matter lesions (WML) and gray matter volume (GMV), respectively.

Results

The WML scores were higher at the 12-month [8.65 ± 3.86; 95% confidence interval (CI), 1.60–2.35; p < 0.001] and 24-month follow-ups (10.11 ± 3.85; 95% CI, 2.98–3.87; p < 0.001) compared to baseline (6.68 ± 3.64). At the 24-month follow-up, the visual scores were also significantly higher in younger patients (3.89 ± 2.04) than in older patients (3.00 ± 1.78; p = 0.047) and higher in female patients (3.80 ± 2.04) than in male patients (2.73 ± 1.56; p = 0.023). Additionally, significant GMV loss was observed in sub-regions of the right occipital lobe (76.71 voxels; 95% CI, 40.740–112.69 voxels), left occipital lobe (93.48 voxels; 95% CI, 37.48–149.47 voxels), and left basal ganglia (37.57 voxels; 95% CI, 21.58–53.57 voxels) (all p < 0.005; cluster-level false discovery rate < 0.05).

Conclusions

An increase in WMLs and loss of GMV were observed in patients with metastatic NSCLC undergoing long-term EGFR-TKI treatment. This might reflect an unknown side-effect of EGFR-TKI treatment. Further prospective studies are necessary to confirm our findings.

Assessment of White Matter Integrity after Pediatric Traumatic Brain Injury

White matter (WM) abnormalities, such as atrophy and hyperintensities (WMH), can be accessed via magnetic resonance imaging (MRI) after pediatric traumatic brain injury (TBI). Several methods are available to classify WM abnormalities (i.e., total WM volumes and WMHs), but automated and manual volumes and clinical ratings have yet to be compared in pediatric TBI. In addition, WM integrity has been associated reliably with processing speed. Consequently, methods of assessing WM integrity should relate to processing speed to have clinical application. This study had two goals: (1) to compare Scheltens rating scale, manual tracing, FreeSurfer, and NeuroQuant® methods of assessing WM abnormalities, and (2) to relate WM methods to processing speed scores. We report findings from the Social Outcomes of Brain Injury in Kids (SOBIK) study, a multi-center study of 60 children with chronic TBI (65% male) from ages 8-13. Scheltens WMH ratings had good to excellent agreement with WMH volumes for both NeuroQuant (ICC = 0.62; r = 0.29, p = 0.005) and manual tracing (ICC = 0.82; r = 0.50, p = 0.000). NeuroQuant WMH volumes did not correlate with manually traced WMH volumes (r = 0.12, p = 0.21) and had poor agreement (ICC = 0.24). NeuroQuant and FreeSurfer total WM volumes correlated (r = 0.38, p = 0.004) and had fair agreement (ICC = 0.52). The WMH assessment methods, both ratings and volumes, were associated with processing speed scores. In contrast, total WM volume was not related to processing speed. Measures of WMH may hold clinical utility for predicting cognitive functioning after pediatric TBI.

Automated lesion segmentation with BIANCA: Impact of population-level features, classification algorithm and locally adaptive thresholding

White matter hyperintensities (WMH) or white matter lesions exhibit high variability in their characteristics both at population- and subject-level, making their detection a challenging task. Population-level factors such as age, vascular risk factors and neurodegenerative diseases affect lesion load and spatial distribution. At the individual level, WMH vary in contrast, amount and distribution in different white matter regions. In this work, we aimed to improve BIANCA, the FSL tool for WMH segmentation, in order to better deal with these sources of variability. We worked on two stages of BIANCA by improving the lesion probability map estimation (classification stage) and making the lesion probability map thresholding stage automated and adaptive to local lesion probabilities. Firstly, in order to take into account the effect of population-level factors, we included population-level lesion probabilities, modelled with respect to a parametric factor (e.g. age), in the classification stage. Secondly, we tested BIANCA performance when using four alternative classifiers commonly used in the literature with respect to K-nearest neighbour algorithm (currently used for lesion probability map estimation in BIANCA). Finally, we propose LOCally Adaptive Threshold Estimation (LOCATE), a supervised method for determining optimal local thresholds to apply to the estimated lesion probability map, as an alternative option to global thresholding (i.e. applying the same threshold to the entire lesion probability map). For these experiments we used data from a neurodegenerative cohort, a vascular cohort and the cohorts available publicly as a part of a segmentation challenge. We observed that including population-level parametric lesion probabilities with respect to age and using alternative machine learning techniques provided negligible improvement. However, LOCATE provided a substantial improvement in the lesion segmentation performance, when compared to the global thresholding. It allowed to detect more deep lesions and provided better segmentation of periventricular lesion boundaries, despite the differences in the lesion spatial distribution and load across datasets. We further validated LOCATE on a cohort of CADASIL (Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) patients, a genetic form of cerebral small vessel disease, and healthy controls, showing that LOCATE adapts well to wide variations in lesion load and spatial distribution.

Identifying patients with neuronal intranuclear inclusion disease in Singapore using characteristic diffusion-weighted MR images

PURPOSE

Adult-onset neuronal intranuclear inclusion disease (NIID) is a rare neurodegenerative disorder described mainly in the Japanese population, with characteristic DWI abnormalities at the junction between gray and white matter. We identify possible cases of NIID in the picture archive and communication system (PACS) of a tertiary neurological referral hospital in Singapore and describe their radiological features.

METHODS

The neuroradiology imaging database was reviewed using keyword search of radiological reports to identify patients who had "subcortical U fibre" abnormalities on DWI. MRI were retrospectively reviewed, and those fulfilling inclusion criteria were invited for skin biopsy to detect nuclear inclusions by light and electron microscopy.

RESULTS

Twelve Chinese patients (nine female; median age 70.5 years) were enrolled. Seven patients were being assessed for dementia and five for other neurological indications. In all patients, DWI showed distinctive subcortical high signal with increased average apparent diffusion coefficient (ADC), involving frontal, parietal, and temporal more than occipital lobes; the corpus callosum and external capsule were affected in some patients. On T2-weighted images, cerebral and cerebellar atrophy and white matter hyperintensity of Fazekas grade 2 and above were seen in all patients. Three patients underwent skin biopsy; all were positive for intranuclear hyaline inclusion bodies on either p62 staining or electron microscopy, which are pathognomonic for NIID.

CONCLUSION

Previously undiagnosed patients with NIID can be identified by searching for abnormalities at the junction between gray and white matter on DWI in PACS and subsequently confirmed by skin biopsy. Radiologists should recognize the distinctive neuroimaging pattern of this dementing disease.

MRI Relaxometry for Quantitative Analysis of USPIO Uptake in Cerebral Small Vessel Disease

A protocol for evaluating ultrasmall superparamagnetic particles of iron oxide (USPIO) uptake and elimination in cerebral small vessel disease patients was developed and piloted. B₁-insensitive R₁ measurement was evaluated in vitro. Twelve participants with history of minor stroke were scanned at 3-T MRI including structural imaging, and R₁ and R₂* mapping. Participants were scanned (i) before and (ii) after USPIO (ferumoxytol) infusion, and again at (iii) 24–30 h and (iv) one month. Absolute and blood-normalised changes in R₁ and R₂* were measured in white matter (WM), deep grey matter (GM), white matter hyperintensity (WMH) and stroke lesion regions. R₁ measurements were accurate across a wide range of values. R₁ (p < 0.05) and R₂* (p < 0.01) mapping detected increases in relaxation rate in all tissues immediately post-USPIO and at 24–30 h. R₂* returned to baseline at one month. Blood-normalised R₁ and R₂* changes post-infusion and at 24–30 h were similar, and were greater in GM versus WM (p < 0.001). Narrower distributions were seen with R₂* than for R₁ mapping. R₁ and R₂* changes were correlated at 24–30 h (p < 0.01). MRI relaxometry permits quantitative evaluation of USPIO uptake; R₂* appears to be more sensitive to USPIO than R₁. Our data are explained by intravascular uptake alone, yielding estimates of cerebral blood volume, and did not support parenchymal uptake. Ferumoxytol appears to be eliminated at 1 month. The approach should be valuable in future studies to quantify both blood-pool USPIO and parenchymal uptake associated with inflammatory cells or blood-brain barrier leak.

Magnetic Resonance Imaging Markers of Cerebral Small Vessel Disease in Hematoma Expansion of Intracerebral Hemorrhage

BACKGROUND

Hematoma expansion is an independent risk factor of unfavorable outcome after intracerebral hemorrhage (ICH), which always occurs in the early phase after symptoms onset. The relationship between underlying small vessel disease (SVD) and hematoma expansion was inconsistent in patients with ICH. We aimed to investigate the relationship between magnetic resonance (MR) characteristics of SVD and hematoma expansion in patients with ICH within 72 hours after symptoms onset.

METHODS

Data were derived from a cohort of biological sample collection from April 2014 to April 2016. We recruited patients aged 18 years or older with a baseline and follow-up computed tomography within 72 hours after symptom onset, as well as an MR imaging within 3 months before or after ICH. Hematoma expansion was defined as an increase in volume between baseline and final hematoma volume exceeding 6 mL or 33% of the baseline volume. Multivariate logistic regression was used to explore the association between clinical characteristics, imaging markers, total SVD score, and hematoma expansion in patients with ICH.

RESULTS

A total of 103 patients experienced hematoma expansion among the 263 enrolled patients (mean age 53.4 ± 14.0 years, 76.4% male). Electrocardiogram abnormal rhythm, fewer non-lobar microbleeds, lower plasma homocysteine concentration, and smaller baseline hematoma volume independently predicted the risk of hematoma expansion (P = .004, .021, .001, and .024, respectively). Odds ratios ranged from 1.02 to 3.72.

CONCLUSIONS

Our study suggested that the use of MR markers revealing underlying SVD may help to identify patients with ICH with potential hematoma expansion.