Quantifiable brain atrophy synthesis for benchmarking of cortical thickness estimation methods.

Cortical thickness (CTh) is routinely used to quantify grey matter atrophy as it is a significant biomarker in studying neurodegenerative and neurological conditions. Clinical studies commonly employ one of several available CTh estimation software tools to estimate CTh from brain MRI scans. In recent years, machine learning-based methods emerged as a faster alternative to the main-stream CTh estimation methods (e.g. FreeSurfer). Evaluation and comparison of CTh estimation methods often include various metrics and downstream tasks, but none fully covers the sensitivity to sub-voxel atrophy characteristic of neurodegeneration. In addition, current evaluation methods do not provide a framework for the intra-method region-wise evaluation of CTh estimation methods. Therefore, we propose a method for brain MRI synthesis capable of generating a range of sub-voxel atrophy levels (global and local) with quantifiable changes from the baseline scan. We further create a synthetic test set and evaluate four different CTh estimation methods: FreeSurfer (cross-sectional), FreeSurfer (longitudinal), DL+DiReCT and HerstonNet. DL+DiReCT showed superior sensitivity to sub-voxel atrophy over other methods in our testing framework. The obtained results indicate that our synthetic test set is suitable for benchmarking CTh estimation methods on both global and local scales as well as regional inter-and intra-method performance comparison.

Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study.

BACKGROUND: Proximal femoral fractures are an important clinical and public health issue associated with substantial morbidity and early mortality. Artificial intelligence might offer improved diagnostic accuracy for these fractures, but typical approaches to testing of artificial intelligence models can underestimate the risks of artificial intelligence-based diagnostic systems. METHODS: We present a preclinical evaluation of a deep learning model intended to detect proximal femoral fractures in frontal x-ray films in emergency department patients, trained on films from the Royal Adelaide Hospital (Adelaide, SA, Australia). This evaluation included a reader study comparing the performance of the model against five radiologists (three musculoskeletal specialists and two general radiologists) on a dataset of 200 fracture cases and 200 non-fractures (also from the Royal Adelaide Hospital), an external validation study using a dataset obtained from Stanford University Medical Center, CA, USA, and an algorithmic audit to detect any unusual or unexpected model behaviour. FINDINGS: In the reader study, the area under the receiver operating characteristic curve (AUC) for the performance of the deep learning model was 0·994 (95% CI 0·988-0·999) compared with an AUC of 0·969 (0·960-0·978) for the five radiologists. This strong model performance was maintained on external validation, with an AUC of 0·980 (0·931-1·000). However, the preclinical evaluation identified barriers to safe deployment, including a substantial shift in the model operating point on external validation and an increased error rate on cases with abnormal bones (eg, Paget's disease). INTERPRETATION: The model outperformed the radiologists tested and maintained performance on external validation, but showed several unexpected limitations during further testing. Thorough preclinical evaluation of artificial intelligence models, including algorithmic auditing, can reveal unexpected and potentially harmful behaviour even in high-performance artificial intelligence systems, which can inform future clinical testing and deployment decisions. FUNDING: None.

Towards data-driven quantification of skin ageing using reflectance confocal microscopy.

INTRODUCTION: Evaluation of skin ageing is a non-standardized, subjective process, with typical measures relying coarse, qualitatively defined features. Reflectance confocal microscopy depth stacks contain indicators of both chrono-ageing and photo-ageing. We hypothesize that an ageing scale could be constructed using machine learning and image analysis, creating a data-driven quantification of skin ageing without human assessment. METHODS: En-face sections of reflectance confocal microscopy depth stacks from the dorsal and volar forearm of 74 participants (36/18/20 training/testing/validation) were represented using a histogram of visual features learned using unsupervised clustering of small image patches. A logistic regression classifier was trained on these histograms to differentiate between stacks from 20- to 30-year-old and 50- to 70-year-old volunteers. The probabilistic output of the logistic regression was used as the fine-grained ageing score for that stack in the testing set ranging from 0 to 1. Evaluation was performed in two ways: on the test set, the AUC was collected for the binary classification problem as well as by statistical comparison of the scores for age and body site groups. Final validation was performed by assessing the accuracy of the ageing score measurement on 20 depth stacks not used for training or evaluating the classifier. RESULTS: The classifier effectively differentiated stacks from age groups with a test set AUC of 0.908. Mean scores were significantly different when comparing age groups (mean 0.70 vs. 0.44; t = -6.62, p = 0.0000) and also when comparing stacks from dorsal and volar body sites (mean 0.64 vs. 0.53; t = 3.12, p = 0.0062). On the final validation set, 17 out of 20 depth stacks were correctly labelled. DISCUSSION: Despite being limited to only coarse training information in the form of example stacks from two age groups, the trained classifier was still able to effectively discriminate between younger skin and older skin. Curiously, despite being only trained with chronological age, there was still evidence for measurable differences in age scores due to sun exposure-with marked differences in scores on sun-exposed dorsal sites of some volunteers compared with less sun-exposed volar sites. These results suggest that fine-grained data-driven quantification of skin ageing is achievable.

INTRODUCTION: L'évaluation du vieillissement de la peau est un processus subjectif et non standardisé, dont les mesures typiques reposent sur des caractéristiques grossières et définies qualitativement. Les strates de profondeur observées grâce à la microscopie confocale par réflectance contiennent des indicateurs de chrono-vieillissement et de photo-vieillissement. Nous émettons l'hypothèse selon laquelle il serait possible d'établir une échelle de vieillissement à l'aide de l'apprentissage automatique et de l'analyse d'images, permettant la mise en place d'une quantification du vieillissement cutané fondée sur les données et sans évaluation humaine. MÉTHODES: À l'aide d'un histogramme des caractéristiques visuelles apprises à partir de petits ensembles d'images regroupées sans supervision, on a représenté des coupes faciales de strates de profondeur observées grâce à la microscopie confocale par réflectance et issues des faces dorsale et palmaire de l'avant-bras de 74 participants (36/18/20 entraînement/analyse/validation). Après un processus d'entraînement portant sur ces histogrammes, un classificateur de régression logistique a appris à différencier les strates prélevées sur des volontaires âgés de 20 à 30 ans et celles prélevées sur des volontaires âgés de 50 à 70 ans. Le résultat probabiliste de la régression logistique a été utilisé comme score du vieillissement de haute précision, allant de 0 à 1, pour cette strate dans l'ensemble d'analyse. L'évaluation a été effectuée de deux manières : dans l'ensemble d'analyse, l'aire sous la courbe (ASC) a été identifiée pour le problème de classification binaire ainsi que par comparaison statistique des scores selon les tranches d'âge et les catégories de site corporel. La validation finale est passée par une évaluation de l'exactitude de la mesure du score de vieillissement sur 20 strates de profondeur non utilisées dans le cadre du processus d'entraînement ou d'évaluation du classificateur. RÉSULTATS: Le classificateur différenciait efficacement les strates des tranches d'âge, avec une ASC dans l'ensemble d'analyse de 0,908. Les scores moyens affichaient des différences significatives lors de la comparaison entre les tranches d'âge (moyenne de 0,70 contre 0,44 ; t = 6,62 ; p = 0,0000) et lors de la comparaison entre les strates issues des faces dorsale et palmaire des sites corporels (moyenne de 0,64 contre 0,53 ; t = 3,12 ; p = 0,0062). Dans l'ensemble de validation finale, 17 strates sur 20 ont été correctement classées.

Investigation of the inherent left-right flow balancing of rotary total artificial hearts by means of a resistance box.

With the incidence of end-stage heart failure steadily increasing, the need for a practical total artificial heart (TAH) has never been greater. Continuous flow TAHs (CFTAH) are being developed using rotary blood pumps (RBPs), leveraging their small size, mechanical simplicity, and excellent durability. To completely replace the heart with currently available RBPs, two are required; one for providing pulmonary flow and one for providing systemic flow. To prevent hazardous states, it is essential to maintain balance between the pulmonary and systemic circulation at a wide variety of physiologic states. In this study, we investigated factors determining a CFTAH's inherent ability to balance systemic and pulmonary flow passively, without active management of pump rotational speed. Four different RBPs (ReliantHeart HA5, Thoratec HMII, HeartWare HVAD, and Ventracor VentrAssist) were used in various combinations to construct CFTAHs. Each CFTAH's ability to autonomously maintain pressures and flows within defined ranges was evaluated in a hybrid mock loop as systemic and pulmonary vascular resistance (PVR) were changed. The resistance box, a method to quantify the range of vascular resistances that can be safely supported by a CFTAH, was used to compare different CFTAH configurations in an efficient and predictive way. To reduce the need for future in vitro tests and to aid in their analysis, a novel analytical evaluation to predict the resistance box of various CFTAH configurations was also performed. None of the investigated CFTAH configurations fully satisfied the predefined benchmarks for inherent flow balancing, with the VentrAssist (left) and HeartAssist 5 (right) offering the best combination. The extent to which each CFTAH was able to autonomously maintain balance was determined by the pressure sensitivity of each RPB: the sensitivity of outflow to changes in the pressure head. The analytical model showed that by matching left and right pressure sensitivity the inherent balancing performance can be improved. These findings may ultimately lead to a reduced need for manual speed changes or active control systems.

Automatic lesion detection, segmentation and characterization via 3D multiscale morphological sifting in breast MRI.

Previous studies on computer aided detection/diagnosis (CAD) in 4D breast magnetic resonance imaging (MRI) usually regard lesion detection, segmentation and characterization as separate tasks, and typically require users to manually select 2D MRI slices or regions of interest as the input. In this work, we present a breast MRI CAD system that can handle 4D multimodal breast MRI data, and integrate lesion detection, segmentation and characterization with no user intervention. The proposed CAD system consists of three major stages: region candidate generation, feature extraction and region candidate classification. Breast lesions are firstly extracted as region candidates using the novel 3D multiscale morphological sifting (MMS). The 3D MMS, which uses linear structuring elements to extract lesion-like patterns, can segment lesions from breast images accurately and efficiently. Analytical features are then extracted from all available 4D multimodal breast MRI sequences, including T1-, T2-weighted and DCE sequences, to represent the signal intensity, texture, morphological and enhancement kinetic characteristics of the region candidates. The region candidates are lastly classified as lesion or normal tissue by the random under-sampling boost (RUSboost), and as malignant or benign lesion by the random forest. Evaluated on a breast MRI dataset which contains a total of 117 cases with 141 biopsy-proven lesions (95 malignant and 46 benign lesions), the proposed system achieves a true positive rate (TPR) of 0.90 at 3.19 false positives per patient (FPP) for lesion detection and a TPR of 0.91 at a FPP of 2.95 for identifying malignant lesions without any user intervention. The average dice similarity index (DSI) is0.72±0.15for lesion segmentation. Compared with previously proposed lesion detection, detection-segmentation and detection-characterization systems evaluated on the same breast MRI dataset, the proposed CAD system achieves a favourable performance in breast lesion detection and characterization.

A Starling-like total work controller for rotary blood pumps: An in vitro evaluation.

Due to improved durability and survival rates, rotary blood pumps (RBPs) are the preferred left ventricular assist device when compared to volume displacement pumps. However, when operated at constant speed, RBPs lack a volume balancing mechanism which may result in left ventricular suction and suboptimal ventricular unloading. Starling-like controllers have previously been developed to balance circulatory volumes; however, they do not consider ventricular workload as a feedback and may have limited sensitivity to adjust RBP workload when ventricular function deteriorates or improves. To address this, we aimed to develop a Starling-like total work controller (SL-TWC) that matched the energy output of a healthy heart by adjusting RBP hydraulic work based on measured left ventricular stroke work and ventricular preload. In a mock circulatory loop, the SL-TWC was evaluated using a HeartWare HVAD in a range of simulated patient conditions. These conditions included changes in systemic hypertension and hypotension, pulmonary hypertension, blood circulatory volume, exercise, and improvement and deterioration of ventricular function by increasing and decreasing ventricular contractility. The SL-TWC was compared to constant speed control where RBP speed was set to restore cardiac output to 5.0 L/min at rest. Left ventricular suction occurred with constant speed control during pulmonary hypertension but was prevented with the SL-TWC. During simulated exercise, the SL-TWC demonstrated reduced LVSW (0.51 J) and greater RBP flow (9.2 L/min) compared to constant speed control (LVSW: 0.74 J and RBP flow: 6.4 L/min). In instances of increased ventricular contractility, the SL-TWC reduced RBP hydraulic work while maintaining cardiac output similar to the rest condition. In comparison, constant speed overworked and increased cardiac output. The SL-TWC balanced circulatory volumes by mimicking the Starling mechanism, while also considering changes in ventricular workload. Compared to constant speed control, the SL-TWC may reduce complications associated with volume imbalances, adapt to changes in ventricular function and improve patient quality of life.

Pre and post-hoc diagnosis and interpretation of malignancy from breast DCE-MRI.

We propose a new method for breast cancer screening from DCE-MRI based on a post-hoc approach that is trained using weakly annotated data (i.e., labels are available only at the image level without any lesion delineation). Our proposed post-hoc method automatically diagnosis the whole volume and, for positive cases, it localizes the malignant lesions that led to such diagnosis. Conversely, traditional approaches follow a pre-hoc approach that initially localises suspicious areas that are subsequently classified to establish the breast malignancy - this approach is trained using strongly annotated data (i.e., it needs a delineation and classification of all lesions in an image). We also aim to establish the advantages and disadvantages of both approaches when applied to breast screening from DCE-MRI. Relying on experiments on a breast DCE-MRI dataset that contains scans of 117 patients, our results show that the post-hoc method is more accurate for diagnosing the whole volume per patient, achieving an AUC of 0.91, while the pre-hoc method achieves an AUC of 0.81. However, the performance for localising the malignant lesions remains challenging for the post-hoc method due to the weakly labelled dataset employed during training.

Author Correction: Why rankings of biomedical image analysis competitions should be interpreted with care.

In the original version of this Article the values in the rightmost column of Table 1 were inadvertently shifted relative to the other columns. This has now been corrected in the PDF and HTML versions of the Article.

Why rankings of biomedical image analysis competitions should be interpreted with care.

International challenges have become the standard for validation of biomedical image analysis methods. Given their scientific impact, it is surprising that a critical analysis of common practices related to the organization of challenges has not yet been performed. In this paper, we present a comprehensive analysis of biomedical image analysis challenges conducted up to now. We demonstrate the importance of challenges and show that the lack of quality control has critical consequences. First, reproducibility and interpretation of the results is often hampered as only a fraction of relevant information is typically provided. Second, the rank of an algorithm is generally not robust to a number of variables such as the test data used for validation, the ranking scheme applied and the observers that make the reference annotations. To overcome these problems, we recommend best practice guidelines and define open research questions to be addressed in the future.

Automated Analysis of Unregistered Multi-View Mammograms With Deep Learning.

We describe an automated methodology for the analysis of unregistered cranio-caudal (CC) and medio-lateral oblique (MLO) mammography views in order to estimate the patient's risk of developing breast cancer. The main innovation behind this methodology lies in the use of deep learning models for the problem of jointly classifying unregistered mammogram views and respective segmentation maps of breast lesions (i.e., masses and micro-calcifications). This is a holistic methodology that can classify a whole mammographic exam, containing the CC and MLO views and the segmentation maps, as opposed to the classification of individual lesions, which is the dominant approach in the field. We also demonstrate that the proposed system is capable of using the segmentation maps generated by automated mass and micro-calcification detection systems, and still producing accurate results. The semi-automated approach (using manually defined mass and micro-calcification segmentation maps) is tested on two publicly available data sets (INbreast and DDSM), and results show that the volume under ROC surface (VUS) for a 3-class problem (normal tissue, benign, and malignant) is over 0.9, the area under ROC curve (AUC) for the 2-class "benign versus malignant" problem is over 0.9, and for the 2-class breast screening problem (malignancy versus normal/benign) is also over 0.9. For the fully automated approach, the VUS results on INbreast is over 0.7, and the AUC for the 2-class "benign versus malignant" problem is over 0.78, and the AUC for the 2-class breast screening is 0.86.

Identifying relevant biomarkers of brain injury from structural MRI: Validation using automated approaches in children with unilateral cerebral palsy.

Previous studies have proposed that the early elucidation of brain injury from structural Magnetic Resonance Images (sMRI) is critical for the clinical assessment of children with cerebral palsy (CP). Although distinct aetiologies, including cortical maldevelopments, white and grey matter lesions and ventricular enlargement, have been categorised, these injuries are commonly only assessed in a qualitative fashion. As a result, sMRI remains relatively underexploited for clinical assessments, despite its widespread use. In this study, several automated and validated techniques to automatically quantify these three classes of injury were generated in a large cohort of children (n = 139) aged 5-17, including 95 children diagnosed with unilateral CP. Using a feature selection approach on a training data set (n = 97) to find severity of injury biomarkers predictive of clinical function (motor, cognitive, communicative and visual function), cortical shape and regional lesion burden were most often chosen associated with clinical function. Validating the best models on the unseen test data (n = 42), correlation values ranged between 0.545 and 0.795 (p<0.008), indicating significant associations with clinical function. The measured prevalence of injury, including ventricular enlargement (70%), white and grey matter lesions (55%) and cortical malformations (30%), were similar to the prevalence observed in other cohorts of children with unilateral CP. These findings support the early characterisation of injury from sMRI into previously defined aetiologies as part of standard clinical assessment. Furthermore, the strong and significant association between quantifications of injury observed on structural MRI and multiple clinical scores accord with empirically established structure-function relationships.

Precision Radiology: Predicting longevity using feature engineering and deep learning methods in a radiomics framework.

Precision medicine approaches rely on obtaining precise knowledge of the true state of health of an individual patient, which results from a combination of their genetic risks and environmental exposures. This approach is currently limited by the lack of effective and efficient non-invasive medical tests to define the full range of phenotypic variation associated with individual health. Such knowledge is critical for improved early intervention, for better treatment decisions, and for ameliorating the steadily worsening epidemic of chronic disease. We present proof-of-concept experiments to demonstrate how routinely acquired cross-sectional CT imaging may be used to predict patient longevity as a proxy for overall individual health and disease status using computer image analysis techniques. Despite the limitations of a modest dataset and the use of off-the-shelf machine learning methods, our results are comparable to previous 'manual' clinical methods for longevity prediction. This work demonstrates that radiomics techniques can be used to extract biomarkers relevant to one of the most widely used outcomes in epidemiological and clinical research - mortality, and that deep learning with convolutional neural networks can be usefully applied to radiomics research. Computer image analysis applied to routinely collected medical images offers substantial potential to enhance precision medicine initiatives.

Closed-form equation to estimate the dielectric properties of biological tissues as a function of age.

Developing microwave systems for biomedical applications requires accurate dielectric properties of biological tissues for reliable modeling before prototyping and subject testing. Dielectric properties of tissues decrease with age due to the change in their water content, but there are no detailed age-dependent data, especially for young tissue-like newborns, in the literature. In this article, an age-dependent formula to predict the dielectric properties of biological tissues was derived. In the proposed method, the variation of water concentration in each type of tissue as a function of age was used to calculate its relative permittivity and conductivity. The derived formula shows that the concentration of water in each tissue type can be modeled as a negative exponential function of age. The dielectric properties of each tissue type can then be calculated as a function of the dielectric properties of water and dielectric properties of the organ forming the tissue and its water concentration. The derived formula was used to generate the dielectric properties of several types of human tissues at different ages using the dielectric properties of a human adult. Moreover, the formula was validated on pig tissues of different ages. A close agreement was achieved between the calculated and measured data with a maximum difference of only 2%. Bioelectromagnetics. 38:474-481, 2017. © 2017 Wiley Periodicals, Inc.

A deep learning approach for the analysis of masses in mammograms with minimal user intervention.

We present an integrated methodology for detecting, segmenting and classifying breast masses from mammograms with minimal user intervention. This is a long standing problem due to low signal-to-noise ratio in the visualisation of breast masses, combined with their large variability in terms of shape, size, appearance and location. We break the problem down into three stages: mass detection, mass segmentation, and mass classification. For the detection, we propose a cascade of deep learning methods to select hypotheses that are refined based on Bayesian optimisation. For the segmentation, we propose the use of deep structured output learning that is subsequently refined by a level set method. Finally, for the classification, we propose the use of a deep learning classifier, which is pre-trained with a regression to hand-crafted feature values and fine-tuned based on the annotations of the breast mass classification dataset. We test our proposed system on the publicly available INbreast dataset and compare the results with the current state-of-the-art methodologies. This evaluation shows that our system detects 90% of masses at 1 false positive per image, has a segmentation accuracy of around 0.85 (Dice index) on the correctly detected masses, and overall classifies masses as malignant or benign with sensitivity (Se) of 0.98 and specificity (Sp) of 0.7.

Evaluation of Three Algorithms for the Segmentation of Overlapping Cervical Cells.

In this paper, we introduce and evaluate the systems submitted to the first Overlapping Cervical Cytology Image Segmentation Challenge, held in conjunction with the IEEE International Symposium on Biomedical Imaging 2014. This challenge was organized to encourage the development and benchmarking of techniques capable of segmenting individual cells from overlapping cellular clumps in cervical cytology images, which is a prerequisite for the development of the next generation of computer-aided diagnosis systems for cervical cancer. In particular, these automated systems must detect and accurately segment both the nucleus and cytoplasm of each cell, even when they are clumped together and, hence, partially occluded. However, this is an unsolved problem due to the poor contrast of cytoplasm boundaries, the large variation in size and shape of cells, and the presence of debris and the large degree of cellular overlap. The challenge initially utilized a database of 16 high-resolution ( ×40 magnification) images of complex cellular fields of view, in which the isolated real cells were used to construct a database of 945 cervical cytology images synthesized with a varying number of cells and degree of overlap, in order to provide full access of the segmentation ground truth. These synthetic images were used to provide a reliable and comprehensive framework for quantitative evaluation on this segmentation problem. Results from the submitted methods demonstrate that all the methods are effective in the segmentation of clumps containing at most three cells, with overlap coefficients up to 0.3. This highlights the intrinsic difficulty of this challenge and provides motivation for significant future improvement.

A Parametric Simulation of Neuronal Noise From Microelectrode Recordings.

In this paper we present an efficient model of microelectrode recordings (MER) from the subthalamic nucleus acquired during deep brain stimulation (DBS) surgery. The model shows how changes in the "noise" relate to the neuronal spike time statistics. A top-down approach is used with analysis-by-synthesis of the MER power spectra. The model is built around a sum of filtered point processes consisting of thousands of neurons and including extracellular filtering. The quality of the model is demonstrated through comparisons to recordings from eight individuals (both hemispheres in six) who have undergone DBS implantation for the treatment of Parkinson's disease. The simulated recordings were compared using their voltage amplitude distributions, power spectral density estimates and phase synchrony while varying only one free parameter (the shape of the inter-spike interval distribution). Through this simple model, we show that the noise present in a DBS MER contains properties that match that of patient recordings when a Weibull distribution with shape parameter of 0.8 is used for the inter-spike interval.

Dielectric properties of dog brain tissue measured in vitro across the 0.3-3 GHz band.

Dielectric properties of dead Greyhound female dogs' brain tissues at different ages were measured at room temperature across the frequency range of 0.3-3 GHz. Measurements were made on excised tissues, in vitro in the laboratory, to carry out dielectric tests on sample tissues. Each dataset for a brain tissue was parametrized using the Cole-Cole expression, and the relevant Cole-Cole parameters for four tissue types are provided. A comparison was made with the database available in literature for other animals and human brain tissue. Results of two types of tissues (white matter and skull) showed systematic variation in dielectric properties as a function of animal age, whereas no significant change related to age was noticed for other tissues. Results provide critical information regarding dielectric properties of animal tissues for a realistic animal head model that can be used to verify the validity and reliability of a microwave head scanner for animals prior to testing on live animals. Bioelectromagnetics. 37:549-556, 2016. © 2016 Wiley Periodicals, Inc.

Alterations in regional shape on ipsilateral and contralateral cortex contrast in children with unilateral cerebral palsy and are predictive of multiple outcomes.

Congenital brain lesions result in a wide range of cerebral tissue alterations observed in children with cerebral palsy (CP) that are associated with a range of functional impairments. The relationship between injury severity and functional outcomes, however, remains poorly understood. This research investigates the differences in cortical shape between children with congenital brain lesions and typically developing children (TDC) and investigates the correlations between cortical shape and functional outcome in a large cohort of patients diagnosed with unilateral CP. Using 139 structural magnetic resonance images, including 95 patients with clinically diagnosed CP and 44 TDC, cortical segmentations were obtained using a modified expectation maximization algorithm. Three shape characteristics (cortical thickness, curvature, and sulcal depth) were computed within a number of cortical regions. Significant differences in these shape measures compared to the TDC were observed on both the injured hemisphere of children with CP (P < 0.004), as well as on the apparently uninjured hemisphere, illustrating potential compensatory mechanisms in these children. Furthermore, these shape measures were significantly correlated with several functional outcomes, including motor, cognition, vision, and communication (P < 0.012), with three out of these four models performing well on test set validation. This study highlights that cortical neuroplastic effects may be quantified using MR imaging, allowing morphological changes to be studied longitudinally, including any influence of treatment. Ultimately, such approaches could be used for the long term prediction of outcomes and the tailoring of treatment to individuals. Hum Brain Mapp 37:3588-3603, 2016. © 2016 Wiley Periodicals, Inc.

Using ventricular modeling to robustly probe significant deep gray matter pathologies: Application to cerebral palsy.

Understanding the relationships between the structure and function of the brain largely relies on the qualitative assessment of Magnetic Resonance Images (MRIs) by expert clinicians. Automated analysis systems can support these assessments by providing quantitative measures of brain injury. However, the assessment of deep gray matter structures, which are critical to motor and executive function, remains difficult as a result of large anatomical injuries commonly observed in children with Cerebral Palsy (CP). Hence, this article proposes a robust surrogate marker of the extent of deep gray matter injury based on impingement due to local ventricular enlargement on surrounding anatomy. Local enlargement was computed using a statistical shape model of the lateral ventricles constructed from 44 healthy subjects. Measures of injury on 95 age-matched CP patients were used to train a regression model to predict six clinical measures of function. The robustness of identifying ventricular enlargement was demonstrated by an area under the curve of 0.91 when tested against a dichotomised expert clinical assessment. The measures also showed strong and significant relationships for multiple clinical scores, including: motor function (r2 = 0.62, P < 0.005), executive function (r2 = 0.55, P < 0.005), and communication (r2 = 0.50, P < 0.005), especially compared to using volumes obtained from standard anatomical segmentation approaches. The lack of reliance on accurate anatomical segmentations and its resulting robustness to large anatomical variations is a key feature of the proposed automated approach. This coupled with its strong correlation with clinically meaningful scores, signifies the potential utility to repeatedly assess MRIs for clinicians diagnosing children with CP. Hum Brain Mapp 37:3795-3809, 2016. © 2016 Wiley Periodicals, Inc.

Automated, quantitative measures of grey and white matter lesion burden correlates with motor and cognitive function in children with unilateral cerebral palsy.

White and grey matter lesions are the most prevalent type of injury observable in the Magnetic Resonance Images (MRIs) of children with cerebral palsy (CP). Previous studies investigating the impact of lesions in children with CP have been qualitative, limited by the lack of automated segmentation approaches in this setting. As a result, the quantitative relationship between lesion burden has yet to be established. In this study, we perform automatic lesion segmentation on a large cohort of data (107 children with unilateral CP and 18 healthy children) with a new, validated method for segmenting both white matter (WM) and grey matter (GM) lesions. The method has better accuracy (94%) than the best current methods (73%), and only requires standard structural MRI sequences. Anatomical lesion burdens most predictive of clinical scores of motor, cognitive, visual and communicative function were identified using the Least Absolute Shrinkage and Selection operator (LASSO). The improved segmentations enabled identification of significant correlations between regional lesion burden and clinical performance, which conform to known structure-function relationships. Model performance was validated in an independent test set, with significant correlations observed for both WM and GM regional lesion burden with motor function (p < 0.008), and between WM and GM lesions alone with cognitive and visual function respectively (p < 0.008). The significant correlation of GM lesions with functional outcome highlights the serious implications GM lesions, in addition to WM lesions, have for prognosis, and the utility of structural MRI alone for quantifying lesion burden and planning therapy interventions.