Search

Sleep has been hypothesised to facilitate waste clearance from the brain. We aimed to determine whether sleep is associated with perivascular spaces on brain magnetic resonance imaging (MRI), a potential marker of impaired brain waste clearance, in a population-based cohort of middle-aged and elderly people. In 559 participants (mean [SD] age 62 [6] years, 52% women) from the population-based Rotterdam Study, we measured total sleep time, sleep onset latency, wake after sleep onset and sleep efficiency with actigraphy and polysomnography. Perivascular space load was determined with brain MRI in four regions (centrum semiovale, basal ganglia, hippocampus, and midbrain) via a validated machine learning algorithm using T2-weighted MR images. Associations between sleep characteristics and perivascular space load were analysed with zero-inflated negative binomial regression models adjusted for various confounders. We found that higher actigraphy-estimated sleep efficiency was associated with a higher perivascular space load in the centrum semiovale (odds ratio 1.10, 95% confidence interval 1.04–1.16, p = 0.0008). No other actigraphic or polysomnographic sleep characteristics were associated with perivascular space load in other brain regions. We conclude that, contrary to our hypothesis, associations of sleep with perivascular space load in this middle-aged and elderly population remained limited to an association of a high actigraphy-estimated sleep efficiency with a higher perivascular space load in the centrum semiovale.

Adversarial attacks are considered a potentially serious security threat for machine learning systems. Medical image analysis (MedIA) systems have recently been argued to be vulnerable to adversarial attacks due to strong financial incentives and the associated technological infrastructure. In this paper, we study previously unexplored factors affecting adversarial attack vulnerability of deep learning MedIA systems in three medical domains: ophthalmology, radiology, and pathology. We focus on adversarial black-box settings, in which the attacker does not have full access to the target model and usually uses another model, commonly referred to as surrogate model, to craft adversarial examples. We consider this to be the most realistic scenario for MedIA systems. Firstly, we study the effect of weight initialization (ImageNet vs. random) on the transferability of adversarial attacks from the surrogate model to the target model. Secondly, we study the influence of differences in development data between target and surrogate models. We further study the interaction of weight initialization and data differences with differences in model architecture. All experiments were done with a perturbation degree tuned to ensure maximal transferability at minimal visual perceptibility of the attacks. Our experiments show that pre-training may dramatically increase the transferability of adversarial examples, even when the target and surrogate's architectures are different: the larger the performance gain using pre-training, the larger the transferability. Differences in the development data between target and surrogate models considerably decrease the performance of the attack; this decrease is further amplified by difference in the model architecture. We believe these factors should be considered when developing security-critical MedIA systems planned to be deployed in clinical practice., Comment: First three authors contributed equally

Scoliosis is a common medical condition, which occurs most often during the growth spurt just before puberty. Untreated Scoliosis may cause long-term sequelae. Therefore, accurate automated quantitative estimation of spinal curvature is an important task for the clinical evaluation and treatment planning of Scoliosis. A couple of attempts have been made for automated Cobb angle estimation on single-view x-rays. It is very challenging to achieve a highly accurate automated estimation of Cobb angles because it is difficult to utilize x-rays efficiently. With the idea of developing methods for accurate automated spinal curvature estimation, AASCE2019 challenge provides spinal anterior-posterior x-ray images with manual labels for training and testing the participating methods. We review eight top-ranked methods from 12 teams. Experimental results show that overall the best performing method achieved a symmetric mean absolute percentage (SMAPE) of 21.71%. Limitations and possible future directions are also described in the paper. We hope the dataset in AASCE2019 and this paper could provide insights into quantitative measurement of the spine.

Labeled data is a critical resource for training and evaluating machine learning models. However, many real-life datasets are only partially labeled. We propose a semi-supervised machine learning training strategy to improve event detection performance on sequential data, such as video recordings, when only sparse labels are available, such as event start times without their corresponding end times. Our method uses noisy guesses of the events' end times to train event detection models. Depending on how conservative these guesses are, mislabeled samples may be introduced into the training set. We further propose a mathematical model for explaining and estimating the evolution of the classification performance for increasingly noisier end time estimates. We show that neural networks can improve their detection performance by leveraging more training data with less conservative approximations despite the higher proportion of incorrect labels. We adapt sequential versions of CIFAR-10 and MNIST, and use the Berkeley MHAD and HMBD51 video datasets to empirically evaluate our method, and find that our risk-tolerant strategy outperforms conservative estimates by 3.5 points of mean average precision for CIFAR, 30 points for MNIST, 3 points for MHAD, and 14 points for HMBD51. Then, we leverage the proposed training strategy to tackle a real-life application: processing continuous video recordings of epilepsy patients, and show that our method outperforms baseline labeling methods by 17 points of average precision, and reaches a classification performance similar to that of fully supervised models. We share part of the code for this article.

Data augmentation is of paramount importance in biomedical image processing tasks, characterized by inadequate amounts of labelled data, to best use all of the data that is present. In-use techniques range from intensity transformations and elastic deformations, to linearly combining existing data points to make new ones. In this work, we propose the use of spectral techniques for data augmentation, using the discrete cosine and wavelet transforms. We empirically evaluate our approaches on a CT texture analysis task to detect abnormal lung-tissue in patients with cystic fibrosis. Empirical experiments show that the proposed spectral methods perform favourably as compared to the existing methods. When used in combination with existing methods, our proposed approach can increase the relative minor class segmentation performance by 44.1% over a simple replication baseline., 5 pages including references, accepted as Oral presentation at IEEE ISBI 2020

Scoliosis is a condition defined by an abnormal spinal curvature. For diagnosis and treatment planning of scoliosis, spinal curvature can be estimated using Cobb angles. We propose an automated method for the estimation of Cobb angles from X-ray scans. First, the centerline of the spine was segmented using a cascade of two convolutional neural networks. After smoothing the centerline, Cobb angles were automatically estimated using the derivative of the centerline. We evaluated the results using the mean absolute error and the average symmetric mean absolute percentage error between the manual assessment by experts and the automated predictions. For optimization, we used 609 X-ray scans from the London Health Sciences Center, and for evaluation, we participated in the international challenge “Accurate Automated Spinal Curvature Estimation, MICCAI 2019” (100 scans). On the challenge’s test set, we obtained an average symmetric mean absolute percentage error of 22.96.

Registration is a core component of many imaging pipelines. In case of clinical scans, with lower resolution and sometimes substantial motion artifacts, registration can produce poor results. Visual assessment of registration quality in large clinical datasets is inefficient. In this work, we propose to automatically assess the quality of registration to an atlas in clinical FLAIR MRI scans of the brain. The method consists of automatically segmenting the ventricles of a given scan using a neural network, and comparing the segmentation to the atlas ventricles propagated to image space. We used the proposed method to improve clinical image registration to a general atlas by computing multiple registrations - one directly to the general atlas and others via different age-specific atlases - and then selecting the registration that yielded the highest ventricle overlap. Finally, as an example application of the complete pipeline, a voxelwise map of white matter hyperintensity burden was computed using only the scans with registration quality above a predefined threshold. Methods were evaluated in a single-site dataset of more than 1000 scans, as well as a multi-center dataset comprising 142 clinical scans from 12 sites. The automated ventricle segmentation reached a Dice coefficient with manual annotations of 0.89 in the single-site dataset, and 0.83 in the multi-center dataset. Registration via age-specific atlases could improve ventricle overlap compared to a direct registration to the general atlas (Dice similarity coefficient increase up to 0.15). Experiments also showed that selecting scans with the registration quality assessment method could improve the quality of average maps of white matter hyperintensity burden, instead of using all scans for the computation of the white matter hyperintensity map. In this work, we demonstrated the utility of an automated tool for assessing image registration quality in clinical scans. This image quality assessment step could ultimately assist in the translation of automated neuroimaging pipelines to the clinic.

Key PointsQuestionIs the gap between brain age predicted from MRI and chronological age associated with incident dementia in a general population of Dutch adults?FindingsBrain age was predicted using a deep learning model, using MRI-derived grey matter density maps. In a population based study including 5496 participants, the observed gap was significantly associated with the risk of dementia.MeaningThe gap between MRI-brain predicted and chronological age is potentially a biomarker for dementia risk screening.AbstractImportanceThe gap between predicted brain age using magnetic resonance imaging (MRI) and chronological age may serve as biomarker for early-stage neurodegeneration and potentially as a risk indicator for dementia. However, owing to the lack of large longitudinal studies, it has been challenging to validate this link.ObjectiveWe aimed to investigate the utility of such a gap as a risk biomarker for incident dementia in a general Dutch population, using a deep learning approach for predicting brain age based on MRI-derived grey matter maps.DesignData was collected from participants of the cohort-based Rotterdam Study who underwent brain magnetic resonance imaging between 2006 and 2015. This study was performed in a longitudinal setting and all participant were followed up for incident dementia until 2016.SettingThe Rotterdam Study is a prospective population-based study, initiated in 1990 in the suburb Ommoord of in Rotterdam, the Netherlands.ParticipantsAt baseline, 5496 dementia- and stroke-free participants (mean age 64.67±9.82, 54.73% women) were scanned and screened for incident dementia. During 6.66±2.46 years of follow-up, 159 people developed dementia.Main outcomes and measuresWe built a convolutional neural network (CNN) model to predict brain age based on its MRI. Model prediction performance was measured in mean absolute error (MAE). Reproducibility of prediction was tested using the intraclass correlation coefficient (ICC) computed on a subset of 80 subjects. Logistic regressions and Cox proportional hazards were used to assess the association of the age gap with incident dementia, adjusted for years of education, ApoEε4 allele carriership, grey matter volume and intracranial volume. Additionally, we computed the attention maps of CNN, which shows which brain regions are important for age prediction.ResultsMAE of brain age prediction was 4.45±3.59 years and ICC was 0.97 (95% confidence interval CI=0.96-0.98). Logistic regression and Cox proportional hazards models showed that the age gap was significantly related to incident dementia (odds ratio OR=1.11 and 95% confidence intervals CI=1.05-1.16; hazard ratio HR=1.11 and 95% CI=1.06-1.15, respectively). Attention maps indicated that grey matter density around the amygdalae and hippocampi primarily drive the age estimation.Conclusion and relevanceWe show that the gap between predicted and chronological brain age is a biomarker associated with risk of dementia development. This suggests that it can be used as a biomarker, complimentary to those that are known, for dementia risk screening.

Enlarged perivascular spaces (PVS) are structural brain changes visible in MRI, are common in aging, and are considered a reflection of cerebral small vessel disease. As such, assessing the burden of PVS has promise as a brain imaging marker. Visual and manual scoring of PVS is a tedious and observer-dependent task. Automated methods would advance research into the etiology of PVS, could aid to assess what a “normal” burden is in aging, and could evaluate the potential of PVS as a biomarker of cerebral small vessel disease. In this work, we propose and evaluate an automated method to quantify PVS in the midbrain, hippocampi, basal ganglia and centrum semiovale. We also compare associations between (earlier established) determinants of PVS and visual PVS scores versus the automated PVS scores, to verify whether automated PVS scores could replace visual scoring of PVS in epidemiological and clinical studies. Our approach is a deep learning algorithm based on convolutional neural network regression, and is contingent on successful brain structure segmentation. In our work we used FreeSurfer segmentations. We trained and validated our method on T2-contrast MR images acquired from 2115 subjects participating in a population-based study. These scans were visually scored by an expert rater, who counted the number of PVS in each brain region. Agreement between visual and automated scores was found to be excellent for all four regions, with intraclass correlation coefficients (ICCs) between 0.75 and 0.88. These values were higher than the inter-observer agreement of visual scoring (ICCs between 0.62 and 0.80). Scan-rescan reproducibility was high (ICCs between 0.82 and 0.93). The association between 20 determinants of PVS, including aging, and the automated scores were similar to those between the same 20 determinants of PVS and visual scores. We conclude that this method may replace visual scoring and facilitate large epidemiological and clinical studies of PVS.

Deep learning techniques are often criticized to heavily depend on a large quantity of labeled data. This problem is even more challenging in medical image analysis where the annotator expertise is often scarce. We propose a novel data-augmentation method to regularize neural network regressors that learn from a single global label per image. The principle of the method is to create new samples by recombining existing ones. We demonstrate the performance of our algorithm on two tasks: estimation of the number of enlarged perivascular spaces in the basal ganglia, and estimation of white matter hyperintensities volume. We show that the proposed method improves the performance over more basic data augmentation. The proposed method reached an intraclass correlation coefficient between ground truth and network predictions of 0.73 on the first task and 0.84 on the second task, only using between 25 and 30 scans with a single global label per scan for training. With the same number of training scans, more conventional data augmentation methods could only reach intraclass correlation coefficients of 0.68 on the first task, and 0.79 on the second task., accepted in MICCAI 2019