Interactive algorithms are an attractive approach to the accurate segmentation of 3D brain scans as they potentially improve the reliability of fully automated segmentation while avoiding the labour intensiveness and inaccuracies of manual segmentation. We present a 3D image analysis package (MIDAS) with a novel architecture enabling highly interactive segmentation algorithms to be implemented as add on modules. Interactive methods based on intensity thresholding, region growing and the constrained application of morphological operators are also presented. The methods involve the application of constraints and freedoms on the algorithms coupled with real time visualisation of the effect. This methodology has been applied to the segmentation, visualisation and measurement of the whole brain and a small irregular neuroanatomical structure, the hippocampus. We demonstrate reproducible and anatomically accurate segmentations of these structures. The efficacy of one method in measuring volume loss (atrophy) of the hippocampus in Alzheimer's disease is shown and is compared to conventional methods.
Laquinimod modulates CNS inflammatory pathways thought to be involved in the pathology of Huntington's disease. Studies with laquinimod in transgenic rodent models of Huntington's disease suggested improvements in motor function, reduction of brain volume loss, and prolonged survival. We aimed to evaluate the safety and efficacy of laquinimod in improving motor function and reducing caudate volume loss in patients with Huntington's disease.
LEGATO-HD was a multicentre, double-blind, placebo-controlled, phase 2 study done at 48 sites across ten countries (Canada, Czech Republic, Germany, Italy, Netherlands, Portugal, Russia, Spain, UK, and USA). Patients aged 21–55 years with a cytosine-adenosine-guanine (CAG) repeat length of between 36 and 49 who had symptomatic Huntington's disease with a Unified Huntington's Disease Rating Scale-Total Motor Score (UHDRS-TMS) of higher than 5 and a Total Functional Capacity score of 8 or higher were randomly assigned (1:1:1:1) by centralised interactive response technology to laquinimod 0·5 mg, 1·0 mg, or 1·5 mg, or to matching placebo, administered orally once daily over 52 weeks; people involved in the randomisation had no other role in the study. Participants, investigators, and study personnel were masked to treatment assignment. The 1·5 mg group was discontinued before recruitment was finished because of cardiovascular safety concerns in multiple sclerosis studies. The primary endpoint was change from baseline in the UHDRS-TMS and the secondary endpoint was percent change in caudate volume, both comparing the 1·0 mg group with the placebo group at week 52. Primary and secondary endpoints were assessed in the full analysis set (ie, all randomised patients who received at least one dose of study drug and had at least one post-baseline UHDRS-TMS assessment). Safety measures included adverse event frequency and severity, and clinical and laboratory examinations, and were assessed in the safety analysis set (ie, all randomised patients who received at least one dose of study drug). This trial is registered with ClinicalTrials.gov, NCT02215616, and EudraCT, 2014–000418–75, and is now complete.
Between Oct 28, 2014, and June 19, 2018, 352 adults with Huntington's disease (179 [51%] men and 173 [49%] women; mean age 43·9 [SD 7·6] years and 340 [97%] White) were randomly assigned: 107 to laquinimod 0·5 mg, 107 to laquinimod 1·0 mg, 30 to laquinimod 1·5 mg, and 108 to matching placebo. Least squares mean change from baseline in UHDRS-TMS at week 52 was 1·98 (SE 0·83) in the laquinimod 1·0 mg group and 1·2 (0·82) in the placebo group (least squares mean difference 0·78 [95% CI –1·42 to 2·98], p=0·4853). Least squares mean change in caudate volume was 3·10% (SE 0·38) in the 1·0 mg group and 4·86% (0·38) in the placebo group (least squares mean difference –1·76% [95% CI –2·67 to –0·85]; p=0·0002). Laquinimod was well tolerated and there were no new safety findings. Serious adverse events were reported by eight (7%) patients on placebo, seven (7%) on laquinimod 0·5 mg, five (5%) on laquinimod 1·0 mg, and one (3%) on laquinimod 1·5 mg. There was one death, which occurred in the placebo group and was unrelated to treatment. The most frequent adverse events in all laquinimod dosed groups (0·5 mg, 1·0 mg, and 1·5 mg) were headache (38 [16%]), diarrhoea (24 [10%]), fall (18 [7%]), nasopharyngitis (20 [8%]), influenza (15 [6%]), vomiting (13 [5%]), arthralgia (11 [5%]), irritability (ten [4%]), fatigue (eight [3%]), and insomnia (eight [3%]).
Laquinimod did not show a significant effect on motor symptoms assessed by the UHDRS-TMS, but significantly reduced caudate volume loss compared with placebo at week 52. Huntington's disease has a chronic and slowly progressive course, and this study does not address whether a longer duration of laquinimod treatment could have produced detectable and meaningful changes in the clinical assessments.
We investigated differences due to sex in brain structural volume and cortical thickness in older cognitively normal (N=742), cognitively impaired (MCI; N=540) and Alzheimer’s Dementia (AD; N=402) individuals from the ADNI and AIBL datasets (861 Males and 823 Females). General linear models were used to control the effect of relevant covariates including age, intracranial volume, magnetic resonance imaging (MRI) scanner field strength and scanner types. Significant volumetric differences due to sex were observed within different cortical and subcortical regions of the cognitively normal group. The number of significantly different regions was reduced in the MCI group, and no region remained different in the AD group. Cortical thickness was overall thinner in males than females in the cognitively normal group, and likewise, the differences due to sex were reduced in the MCI and AD groups. These findings were sustained after including cerebrospinal fluid (CSF) Tau and phosphorylated tau (pTau) as additional covariates.
The automatic and accurate segmentation of the hippocampus in brain magnetic resonance (MR) images is important to study various neurological diseases. However, it is a challenging task due to the small structure size, the irregular shape, and the blurred boundaries between the hippocampus and its surrounding structures. In this chapter, we review the two most popularly used types of methods for this task: (1) multiatlas-based methods and (2) learning-based methods. We first review various existing patch-based multiatlas label fusion strategies. Then, we present a supervised metric learning-based label fusion in detail. This method learns a distance metric model from the atlases to keep the image patches of the same structure close to each other and those of different structures distant. For the learning-based methods, we present a multiatlas-based deep learning method and an end-to-end deep learning method in detail. Specifically, the multiatlas-based deep learning method applies a deep learning-based confidence estimation method to alleviate the potential effects of the registration errors in the traditional atlas-based methods. The end-to-end deep learning method directly learns the segmentation maps from the input images, by embedding a dilated dense network in the residual U-net. We present a comprehensive evaluation of the discussed methods compared with the state-of-the-art methods using the public datasets. In the end of the chapter, we include promising directions for related future research.
Neurofilament light chain (NfL) is a promising blood biomarker in genetic frontotemporal dementia, with elevated concentrations in symptomatic carriers of mutations in GRN, C9orf72, and MAPT. A better understanding of NfL dynamics is essential for upcoming therapeutic trials. We aimed to study longitudinal NfL trajectories in people with presymptomatic and symptomatic genetic frontotemporal dementia.
We recruited participants from 14 centres collaborating in the Genetic Frontotemporal Dementia Initiative (GENFI), which is a multicentre cohort study of families with genetic frontotemporal dementia done across Europe and Canada. Eligible participants (aged ≥18 years) either had frontotemporal dementia due to a pathogenic mutation in GRN, C9orf72, or MAPT (symptomatic mutation carriers) or were healthy at-risk first-degree relatives (either presymptomatic mutation carriers or non-carriers), and had at least two serum samples with a time interval of 6 months or more. Participants were excluded if they had neurological comorbidities that were likely to affect NfL, including cerebrovascular events. We measured NfL longitudinally in serum samples collected between June 8, 2012, and Dec 8, 2017, through follow-up visits annually or every 2 years, which also included MRI and neuropsychological assessments. Using mixed-effects models, we analysed NfL changes over time and correlated them with longitudinal imaging and clinical parameters, controlling for age, sex, and study site. The primary outcome was the course of NfL over time in the various stages of genetic frontotemporal dementia.
We included 59 symptomatic carriers and 149 presymptomatic carriers of a mutation in GRN, C9orf72, or MAPT, and 127 non-carriers. Nine presymptomatic carriers became symptomatic during follow-up (so-called converters). Baseline NfL was elevated in symptomatic carriers (median 52 pg/mL [IQR 24–69]) compared with presymptomatic carriers (9 pg/mL [6–13]; p<0·0001) and non-carriers (8 pg/mL [6–11]; p<0·0001), and was higher in converters than in non-converting carriers (19 pg/mL [17–28] vs 8 pg/mL [6–11]; p=0·0007; adjusted for age). During follow-up, NfL increased in converters (b=0·097 [SE 0·018]; p<0·0001). In symptomatic mutation carriers overall, NfL did not change during follow-up (b=0·017 [SE 0·010]; p=0·101) and remained elevated. Rates of NfL change over time were associated with rate of decline in Mini Mental State Examination (b=–94·7 [SE 33·9]; p=0·003) and atrophy rate in several grey matter regions, but not with change in Frontotemporal Lobar Degeneration-Clinical Dementia Rating scale score (b=–3·46 [SE 46·3]; p=0·941).
Our findings show the value of blood NfL as a disease progression biomarker in genetic frontotemporal dementia and suggest that longitudinal NfL measurements could identify mutation carriers approaching symptom onset and capture rates of brain atrophy. The characterisation of NfL over the course of disease provides valuable information for its use as a treatment effect marker.
2019, Handbook of Medical Image Computing and Computer Assisted Intervention
In the context of Alzheimer's Disease, neuroimaging can provide regional indicators of biological or pathogenic processes or responses to therapeutic intervention. These imaging biomarkers coming from different image modalities and acquisitions can be used for a large variety of purposes (diagnostic, safety assessment, inclusion/exclusion criteria, …). With the development of better and more complete and complex imaging acquisition techniques, the requirements over imaging analysis techniques and the need for more robust, more accurate and faster automation techniques have grown stronger. This chapter presents an overview of existing imaging biomarkers used in the context of Alzheimer's Disease and the development of the field towards the automation of the extraction of imaging measurements. The challenges associated to this automation as well as the practical use of imaging biomarkers is then presented followed by practical examples and their relevance in the context of longitudinal pathological evolution. Lastly, the impact of newly acquired computational capacities in the field of imaging biomarkers and current research trends will be explored.
2019, Progress in Molecular Biology and Translational Science
Huntington's disease (HD) gene-carriers show prominent neuronal loss by end-stage disease, and the use of magnetic resonance imaging (MRI) has been increasingly used to quantify brain changes during earlier stages of the disease. MRI offers an in vivo method of measuring structural and functional brain change. The images collected via MRI are processed to measure different anatomical features, such as brain volume, macro- and microstructural changes within white matter and functional brain activity. Structural imaging has demonstrated significant volume loss across multiple white and gray matter regions in HD, particularly within subcortical structures. There also appears to be increasing disorganization of white matter tracts and between-region connectivity with increasing disease progression. Finally, functional changes are thought to represent changes in brain activity underlying compensatory mechanisms in HD. This chapter will provide an overview of the principles of MRI and practicalities associated with using MRI in HD studies, and summarize findings from MRI studies investigating brain structure and function in HD.
