Review of automatic segmentation methods of multiple sclerosis white matter lesions on conventional magnetic resonance imaging

TAPAS: A Thresholding Approach for Probability Map Automatic Segmentation in Multiple Sclerosis

Total brain white matter lesion (WML) volume is the most widely established magnetic resonance imaging (MRI) outcome measure in studies of multiple sclerosis (MS). To estimate WML volume, there are a number of automatic segmentation methods available, yet manual delineation remains the gold standard approach. Automatic approaches often yield a probability map to which a threshold is applied to create lesion segmentation masks. Unfortunately, few approaches systematically determine the threshold employed; many methods use a manually selected threshold, thus introducing human error and bias into the automated procedure. In this study, we propose and validate an automatic thresholding algorithm, Thresholding Approach for Probability Map Automatic Segmentation in Multiple Sclerosis (TAPAS), to obtain subject-specific threshold estimates for probability map automatic segmentation of T2-weighted (T2) hyperintense WMLs. Using multimodal MRI, the proposed method applies an automatic segmentation algorithm to obtain probability maps. We obtain the true subject-specific threshold that maximizes the Sørensen-Dice similarity coefficient (DSC). Then the subject-specific thresholds are modeled on a naive estimate of volume using a generalized additive model. Applying this model, we predict a subject-specific threshold in data not used for training. We ran a Monte Carlo-resampled split-sample cross-validation (100 validation sets) using two data sets: the first obtained from the Johns Hopkins Hospital (JHH) on a Philips 3 Tesla (3T) scanner (n = 94) and a second collected at the Brigham and Women's Hospital (BWH) using a Siemens 3T scanner (n = 40). By means of the proposed automated technique, in the JHH data we found an average reduction in subject-level absolute error of 0.1 mL per one mL increase in manual volume. Using Bland-Altman analysis, we found that volumetric bias associated with group-level thresholding was mitigated when applying TAPAS. The BWH data showed similar absolute error estimates using group-level thresholding or TAPAS likely since Bland-Altman analyses indicated no systematic biases associated with group or TAPAS volume estimates. The current study presents the first validated fully automated method for subject-specific threshold prediction to segment brain lesions.

Deep-Learning-Based Neural Tissue Segmentation of MRI in Multiple Sclerosis: Effect of Training Set Size

BACKGROUND

The dependence of deep-learning (DL)-based segmentation accuracy of brain MRI on the training size is not known.

PURPOSE

To determine the required training size for a desired accuracy in brain MRI segmentation in multiple sclerosis (MS) using DL.

STUDY TYPE

Retrospective analysis of MRI data acquired as part of a multicenter clinical trial.

STUDY POPULATION

In all, 1008 patients with clinically definite MS.

FIELD STRENGTH/SEQUENCE

MRIs were acquired at 1.5T and 3T scanners manufactured by GE, Philips, and Siemens with dual turbo spin echo, FLAIR, and T1 -weighted turbo spin echo sequences.

ASSESSMENT

Segmentation results using an automated analysis pipeline and validated by two neuroimaging experts served as the ground truth. A DL model, based on a fully convolutional neural network, was trained separately using 16 different training sizes. The segmentation accuracy as a function of the training size was determined. These data were fitted to the learning curve for estimating the required training size for desired accuracy.

STATISTICAL TESTS

The performance of the network was evaluated by calculating the Dice similarity coefficient (DSC), and lesion true-positive and false-positive rates.

RESULTS

The DSC for lesions showed much stronger dependency on the sample size than gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF). When the training size was increased from 10 to 800 the DSC values varied from 0.00 to 0.86 ± 0.016 for T2 lesions, 0.87 ± 009 to 0.94 ± 0.004 for GM, 0.86 ± 0.08 to 0.94 ± 0.005 for WM, and 0.91 ± 0.009 to 0.96 ± 0.003 for CSF.

DATA CONCLUSION

Excellent segmentation was achieved with a training size as small as 10 image volumes for GM, WM, and CSF. In contrast, a training size of at least 50 image volumes was necessary for adequate lesion segmentation.

LEVEL OF EVIDENCE

1 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2020;51:1487-1496.

Unsupervised Domain Adaptation With Optimal Transport in Multi-Site Segmentation of Multiple Sclerosis Lesions From MRI Data

Automatic segmentation of Multiple Sclerosis (MS) lesions from Magnetic Resonance Imaging (MRI) images is essential for clinical assessment and treatment planning of MS. Recent years have seen an increasing use of Convolutional Neural Networks (CNNs) for this task. Although these methods provide accurate segmentation, their applicability in clinical settings remains limited due to a reproducibility issue across different image domains. MS images can have highly variable characteristics across patients, MRI scanners and imaging protocols; retraining a supervised model with data from each new domain is not a feasible solution because it requires manual annotation from expert radiologists. In this work, we explore an unsupervised solution to the problem of domain shift. We present a framework, Seg-JDOT, which adapts a deep model so that samples from a source domain and samples from a target domain sharing similar representations will be similarly segmented. We evaluated the framework on a multi-site dataset, MICCAI 2016, and showed that the adaptation toward a target site can bring remarkable improvements in a model performance over standard training.

Automated separation of diffusely abnormal white matter from focal white matter lesions on MRI in multiple sclerosis

BACKGROUND

Previous histopathology and MRI studies have addressed the differences between focal white matter lesions (FWML) and diffusely abnormal white matter (DAWM) in multiple sclerosis (MS). These two categories of white matter T2-weighted (T2w) hyperintensity show different degrees of demyelination, axonal loss and immune cell density on histopathology, potentially offering distinct correlations with symptoms.

OBJECTIVES

1) To automate the separation of FWML and DAWM using T2w MRI intensity thresholds and to investigate their differences in magnetization transfer ratios (MTR), which are sensitive to myelin content; 2) to correlate MTR values in FWML and DAWM with normalized signal intensity values on fluid attenuated inversion recovery (FLAIR), T2w, and T1-weighted (T1w) contrasts, as well as with the ratio of T2w/T1w normalized values, in order to determine whether these normalized intensities can be used when MTR is not available.

METHODS

We used three MRI datasets: datasets 1 and 2 had 20 MS participants each, scanned with similar 3T MRI protocols in 2 centers, including: 3D T1w (MP2RAGE), 3D FLAIR, 2D T2w, and 3D magnetization-transfer (MT) contrasts. Dataset 3 consisted of 67 scans of participants enrolled in a multisite study and had T1w and T2w contrasts. We used the first dataset to develop an automated technique to separate FWML from DAWM and the second and third to validate the automation of the technique. We applied the automatic thresholds to all datasets to assess the overlap of the manual and the automated masks using Dice kappa. We also assessed differences in mean MTR values between NAWM, DAWM and FWML, using manually and automatically derived masks in datasets 1 and 2. Finally, we used the mean intensity of manually-traced areas of NAWM on T2w images as the normalization factor for each MRI contrast, and compared these with the normalized-intensity values obtained using automated NAWM (A-NAWM) masks as the normalization factor. ANOVA assessed the MTR differences across tissue types. Paired t-test or Wilcoxon signed-ranked test assessed FWML and DAWM differences between manual and automatically derived volumes. Pearson correlations assessed the relationship between MTR and normalized intensity values in the manual and automatically derived masks.

RESULTS

The mean Dice-kappa values for dataset 1 were: 0.79 for DAWM masks and 0.90 for FWML masks. In dataset 2, mean Dice-kappa values were: 0.78 for DAWM and 0.87 for FWML. In dataset 3, mean Dice-kappa values were 0.72 for DAWM, and 0.87 for FWML. Manual and automated DAWM and FWML volumes were not significantly different in all datasets. MTR values were significantly lower in manually and automatically derived FWML compared with DAWM in both datasets (dataset 1 manual: F ​= ​111,08, p ​< ​0.0001; automated: F ​= ​153.90, p ​< ​0.0001; dataset 2 manual: F ​= ​31.25, p ​< ​0.0001; automated: F ​= ​74.04, p ​< ​0.0001). In both datasets, manually derived FWML and DAWM MTR values showed significant correlations with normalized T1w (r ​= ​0.77 to 0.94) intensities.

CONCLUSIONS

The separation of FWML and DAWM on MRI scans of MS patients using automated intensity thresholds on T2w images is feasible. MTR values are significantly lower in FWML than DAWM, and DAWM values are significantly lower than NAWM, reflecting potentially greater demyelination within focal lesions. T1w normalized intensity values exhibit a significant correlation with MTR values in both tissues of interest and could be used as a proxy to assess demyelination when MTR or other myelin-sensitive images are not available.

Automated Detection and Segmentation of Multiple Sclerosis Lesions Using Ultra-High-Field MP2RAGE

OBJECTIVES

The aim of this study was to develop a new automated segmentation method of white matter (WM) and cortical multiple sclerosis (MS) lesions visible on magnetization-prepared 2 inversion-contrast rapid gradient echo (MP2RAGE) images acquired at 7 T MRI.

MATERIALS AND METHODS

The proposed prototype (MSLAST [Multiple Sclerosis Lesion Analysis at Seven Tesla]) takes as input a single image contrast derived from the 7T MP2RAGE prototype sequence and is based on partial volume estimation and topological constraints. First, MSLAST performs a skull-strip of MP2RAGE images and computes tissue concentration maps for WM, gray matter (GM), and cerebrospinal fluid (CSF) using a partial volume model of tissues within each voxel. Second, MSLAST performs (1) connected-component analysis to GM and CSF concentration maps to classify small isolated components as MS lesions; (2) hole-filling in the WM concentration map to classify areas with low WM concentration surrounded by WM (ie, MS lesions); and (3) outlier rejection to the WM mask to improve the classification of small WM lesions. Third, MSLAST unifies the 3 maps obtained from 1, 2, and 3 processing steps to generate a global lesion mask.

RESULTS

Quantitative and qualitative assessments were performed using MSLAST in 25 MS patients from 2 research centers. Overall, MSLAST detected a median of 71% of MS lesions, specifically 74% of WM and 58% of cortical lesions, when a minimum lesion size of 6 μL was considered. The median false-positive rate was 40%. When a 15 μL minimal lesions size was applied, which is the approximation of the minimal size recommended for 1.5/3 T images, the median detection rate was 80% for WM and 63% for cortical lesions, respectively, and the median false-positive rate was 33%. We observed high correlation between MSLAST and manual segmentations (Spearman rank correlation coefficient, ρ = 0.91), although MSLAST underestimated the total lesion volume (average difference of 1.1 mL), especially in patients with high lesion loads. MSLAST also showed good scan-rescan repeatability within the same session with an average absolute volume difference and F1 score of 0.38 ± 0.32 mL and 84%, respectively.

CONCLUSIONS

We propose a new methodology to facilitate the segmentation of WM and cortical MS lesions at 7 T MRI, our approach uses a single MP2RAGE scan and may be of special interest to clinicians and researchers.

Cornea: A Window to White Matter Changes in Stroke; Corneal Confocal Microscopy a Surrogate Marker for the Presence and Severity of White Matter Hyperintensities in Ischemic Stroke

PURPOSE

The presence of white matter hyperintensities (WMH) on MRI imaging confers an increased risk of stroke, dementia, and death. Corneal confocal microscopy (CCM) can detect nerve injury non-invasively and may be a useful surrogate marker for WMH. The objective is to determine whether corneal nerve pathology identified using CCM is associated with the presence of WMH in patients with acute ischemic stroke.

METHODS

This is a cross-sectional study where 196 consecutive individuals with acute ischemic stroke were enrolled and underwent neurological examination, MRI brain imaging and CCM. Participants underwent blinded quantification of WMH and corneal nerve fiber density (CNFD), corneal nerve branch density (CNBD) and corneal nerve fiber length (CNFL).

RESULTS

The prevalence of hypertension [P = .013] was significantly higher and CNFD [P = .031] was significantly lower in patients with WMH compared to those without WMH. CNFD and CNFL were significantly lower in patients with DM without WMH [P = .008, P = .019] and in patients with DM and WMH [P = .042, P = .024] compared to patients without DM or WMH, respectively. In a multivariate model, a 1-unit decrease in the CNFD increased the risk of WMH by 6%, after adjusting for age, DM, gender, dyslipidemia, metabolic syndrome, smoking, and HbA1c. DM was associated with a decrease in all CCM parameters but was not a significant independent factor associated with WMH.

CONCLUSIONS

CCM demonstrates corneal nerve pathology, which is associated with the presence of WMH in participants with acute ischemic stroke. CCM may be a useful surrogate imaging marker for the presence and severity of WMHs.

Are multi-contrast magnetic resonance images necessary for segmenting multiple sclerosis brains? A large cohort study based on deep learning

BACKGROUND

Magnetic resonance images with multiple contrasts or sequences are commonly used for segmenting brain tissues, including lesions, in multiple sclerosis (MS). However, acquisition of images with multiple contrasts increases the scan time and complexity of the analysis, possibly introducing factors that could compromise segmentation quality.

OBJECTIVE

To investigate the effect of various combinations of multi-contrast images as input on the segmented volumes of gray (GM) and white matter (WM), cerebrospinal fluid (CSF), and lesions using a deep neural network.

METHODS

U-net, a fully convolutional neural network was used to automatically segment GM, WM, CSF, and lesions in 1000 MS patients. The input to the network consisted of 15 combinations of FLAIR, T1-, T2-, and proton density-weighted images. The Dice similarity coefficient (DSC) was evaluated to assess the segmentation performance. For lesions, true positive rate (TPR) and false positive rate (FPR) were also evaluated. In addition, the effect of lesion size on lesion segmentation was investigated.

RESULTS

Highest DSC was observed for all the tissue volumes, including lesions, when the input was combination of all four image contrasts. All other input combinations that included FLAIR also provided high DSC for all tissue classes. However, the quality of lesion segmentation showed strong dependence on the input images. The DSC and TPR values for inputs with the four contrast combination and FLAIR alone were very similar, but FLAIR showed a moderately higher FPR for lesion size <100 μl. For lesions smaller than 20 μl all image combinations resulted in poor performance. The segmentation quality improved with lesion size.

CONCLUSIONS

Best performance for segmented tissue volumes was obtained with all four image contrasts as the input, and comparable performance was attainable with FLAIR only as the input, albeit with a moderate increase in FPR for small lesions. This implies that acquisition of only FLAIR images provides satisfactory tissue segmentation. Lesion segmentation was poor for very small lesions and improved rapidly with lesion size.

Automated Segmentation of Tissues Using CT and MRI: A Systematic Review

RATIONALE AND OBJECTIVES

The automated segmentation of organs and tissues throughout the body using computed tomography and magnetic resonance imaging has been rapidly increasing. Research into many medical conditions has benefited greatly from these approaches by allowing the development of more rapid and reproducible quantitative imaging markers. These markers have been used to help diagnose disease, determine prognosis, select patients for therapy, and follow responses to therapy. Because some of these tools are now transitioning from research environments to clinical practice, it is important for radiologists to become familiar with various methods used for automated segmentation.

MATERIALS AND METHODS

The Radiology Research Alliance of the Association of University Radiologists convened an Automated Segmentation Task Force to conduct a systematic review of the peer-reviewed literature on this topic.

RESULTS

The systematic review presented here includes 408 studies and discusses various approaches to automated segmentation using computed tomography and magnetic resonance imaging for neurologic, thoracic, abdominal, musculoskeletal, and breast imaging applications.

CONCLUSION

These insights should help prepare radiologists to better evaluate automated segmentation tools and apply them not only to research, but eventually to clinical practice.

Comparing lesion segmentation methods in multiple sclerosis: Input from one manually delineated subject is sufficient for accurate lesion segmentation

PURPOSE

Accurate lesion segmentation is important for measurements of lesion load and atrophy in subjects with multiple sclerosis (MS). International MS lesion challenges show a preference of convolutional neural networks (CNN) strategies, such as nicMSlesions. However, since the software is trained on fairly homogenous training data, we aimed to test the performance of nicMSlesions in an independent dataset with manual and other automatic lesion segmentations to determine whether this method is suitable for larger, multi-center studies.

METHODS

Manual lesion segmentation was performed in fourteen subjects with MS on sagittal 3D FLAIR images from a 3T GE whole-body scanner with 8-channel head coil. We compared five different categories of automated lesion segmentation methods for their volumetric and spatial agreement with manual segmentation: (i) unsupervised, untrained (LesionTOADS); (ii) supervised, untrained (LST-LPA and nicMSlesions with default settings); (iii) supervised, untrained with threshold adjustment (LST-LPA optimized for current data); (iv) supervised, trained with leave-one-out cross-validation on fourteen subjects with MS (nicMSlesions and BIANCA); and (v) supervised, trained on a single subject with MS (nicMSlesions). Volumetric accuracy was determined by the intra-class correlation coefficient (ICC) and spatial accuracy by Dice's similarity index (SI). Volumes and SI were compared between methods using repeated measures ANOVA or Friedman tests with post-hoc pairwise comparison.

RESULTS

The best volumetric and spatial agreement with manual was obtained with the supervised and trained methods nicMSlesions and BIANCA (ICC absolute agreement > 0.968 and median SI > 0.643) and the worst with the unsupervised, untrained method LesionTOADS (ICC absolute agreement = 0.140 and median SI = 0.444). Agreement with manual in the single-subject network training of nicMSlesions was poor for input with low lesion volumes (i.e. two subjects with lesion volumes ≤ 3.0 ml). For the other twelve subjects, ICC varied from 0.593 to 0.973 and median SI varied from 0.535 to 0.606. In all cases, the single-subject trained nicMSlesions segmentations outperformed LesionTOADS, and in almost all cases it also outperformed LST-LPA.

CONCLUSION

Input from only one subject to re-train the deep learning CNN nicMSlesions is sufficient for adequate lesion segmentation, with on average higher volumetric and spatial agreement with manual than obtained with the untrained methods LesionTOADS and LST-LPA.

Whole volume brain extraction for multi-centre, multi-disease FLAIR MRI datasets

Automatic segmentation of the brain from magnetic resonance images (MRI) is a fundamental step in many neuroimaging processing frameworks. There are mature technologies for this task for T1- and T2-weighted MRI; however, a widely-accepted brain extraction method for Fluid-Attenuated Inversion Recovery (FLAIR) MRI has yet to be established. FLAIR MRI are becoming increasingly important for the analysis of neurodegenerative diseases and tools developed for this sequence would have clinical value. To maximize translation opportunities and for large scale research studies, algorithms for brain extraction in FLAIR MRI should generalize to multi-centre (MC) data. To this end, this work proposes a fully automated, whole volume brain extraction methodology for MC FLAIR MRI datasets. The framework is built using a novel standardization framework which reduces acquisition artifacts, standardizes the intensities of tissues and normalizes the spatial coordinates of brain tissue across MC datasets. Using the standardized datasets, an intuitive set of features based on intensity, spatial location and gradients are extracted and classified using a random forest (RF) classifier to segment the brain tissue class. A series of experiments were conducted to optimize classifier parameters, and to determine segmentation accuracy for standardized and unstandardized (original) data, as a function of scanner vendor, feature type and disease type. The models are trained, tested and validated on 156 image volumes (∼8000 image slices) from two multi-centre, multi-disease datasets, acquired with varying imaging parameters from 30 centres and three scanner vendors. The image datasets, denoted as CAIN and ADNI for vascular and dementia disease, respectively, represent a diverse collection of MC data to test the generalization capabilities of the proposed design. Results demonstrate the importance of standardization for segmentation of MC data, as models trained on standardized data yielded a drastic improvement in brain extraction accuracy compared to the original, unstandardized data (CAIN: DSC = 91% and ADNI: DSC = 86% vs. CAIN: 78% and ADNI: 65%). It was also found that models created from one scanner vendor based on unstandardized data yielded poor segmentation results in data acquired from other scanner vendors, which was improved through standardization. These results demonstrate that to create consistency in segmentations from multi-institutional datasets it is paramount that MC variability be mitigated to improve stability and to ensure generalization of machine learning algorithms for MRI.

Stacked in-plane histology for quantitative validation of non-invasive imaging biomarkers: Application to an infiltrative brain tumour model

BACKGROUND

While it is generally agreed that histopathology is the gold standard for assessing non-invasive imaging biomarkers, most validation has been by qualitative visual comparison. To date, the difficulties involved in accurately co-registering histology sections with imaging slices have prevented a voxel-by-voxel assessment of imaging modalities. By contrast with previous studies, which focus on improving the registration algorithms, we have taken the approach of improving the quality of the histological processing and analysis.

NEW METHOD

To account for imaging slice orientation and thickness, multiple histology sections were cut in the MR imaging plane and averaged to produce stacked in-plane histology (SIH) maps. When combined with intensity sensitive staining this approach gives histopathology maps, which can be used as the gold standard to validate imaging biomarkers.

RESULTS

We applied this pipeline to a patient-derived mouse model of glioblastoma multiforme (GBM). Increasing the number of stacked histology sections significantly increased SIH measured tumour volume. The SIH technique proposed here resulted in reduced variability of volume measurements and this allowed significant improvements in the quantitative volumetric assessment of multiple MRI modalities. Further, high quality registration enabled a voxel-wise comparison between MRI and histopathology maps. Previous approaches to the validation of imaging biomarkers with histology, have been either qualitative or of limited accuracy. Here we propose a pipeline that allows for a more accurate validation via co-registration with SIH maps, potentially allowing validation in a voxel-wise mode.

CONCLUSION

This work demonstrates that methodically produced SIH maps facilitate the quantitative histopathologic assessment of imaging biomarkers.

Predicting Posttraumatic Stress Disorder Risk: A Machine Learning Approach

BACKGROUND

A majority of adults in the United States are exposed to a potentially traumatic event but only a handful go on to develop impairing mental health conditions such as posttraumatic stress disorder (PTSD).

OBJECTIVE

Identifying those at elevated risk shortly after trauma exposure is a clinical challenge. The aim of this study was to develop computational methods to more effectively identify at-risk patients and, thereby, support better early interventions.

METHODS

We proposed machine learning (ML) induction of models to automatically predict elevated PTSD symptoms in patients 1 month after a trauma, using self-reported symptoms from data collected via smartphones.

RESULTS

We show that an ensemble model accurately predicts elevated PTSD symptoms, with an area under the curve (AUC) of .85, using a bag of support vector machines, naive Bayes, logistic regression, and random forest algorithms. Furthermore, we show that only 7 self-reported items (features) are needed to obtain this AUC. Most importantly, we show that accurate predictions can be made 10 to 20 days posttrauma.

CONCLUSIONS

These results suggest that simple smartphone-based patient surveys, coupled with automated analysis using ML-trained models, can identify those at risk for developing elevated PTSD symptoms and thus target them for early intervention.

Longitudinal analysis of white matter and cortical lesions in multiple sclerosis

PURPOSE

The goals of this study were to assess the performance of a novel lesion segmentation tool for longitudinal analyses, as well as to validate the generated lesion progression map between two time points using conventional and non-conventional MR sequences.

MATERIAL AND METHODS

The lesion segmentation approach was evaluated with (LeMan-PV) and without (LeMan) the partial volume framework using "conventional" and "non-conventional" MR imaging in a two-year follow-up prospective study of 32 early RRMS patients. Manual segmentations of new, enlarged, shrunken, and stable lesions were used to evaluate the performance of the method variants. The true positive rate was estimated for those lesion evolutions in both white matter and cortex. The number of false positives was compared with two strategies for longitudinal analyses. New lesion tissue volume estimation was evaluated using Bland-Altman plots. Wilcoxon signed-rank test was used to evaluate the different setups.

RESULTS

The best median of the true positive rate was obtained using LeMan-PV with non-conventional sequences (P < .05): 87%, 87%, 100%, 83%, for new, enlarged, shrunken, and stable WM lesions, and 50%, 60%, 50%, 80%, for new, enlarged, shrunken, and stable cortical lesions, respectively. Most of the missed lesions were below the mean lesion size in each category. Lesion progression maps presented a median of 0 false positives (range:0-9) and the partial volume framework improved the volume estimation of new lesion tissue.

CONCLUSION

LeMan-PV exhibited the best performance in the detection of new, enlarged, shrunken and stable WM lesions. The method showed lower performance in the detection of cortical lesions, likely due to their low occurrence, small size and low contrast with respect to surrounding tissues. The proposed lesion progression map might be useful in clinical trials or clinical routine.

FLAIR2 improves LesionTOADS automatic segmentation of multiple sclerosis lesions in non-homogenized, multi-center, 2D clinical magnetic resonance images

Background

Accurate segmentation of MS lesions on MRI is difficult and, if performed manually, time consuming. Automatic segmentations rely strongly on the image contrast and signal-to-noise ratio. Literature examining segmentation tool performances in real-world multi-site data acquisition settings is scarce.

Objective

FLAIR², a combination of T₂-weighted and fluid attenuated inversion recovery (FLAIR) images, improves tissue contrast while suppressing CSF. We compared the use of FLAIR and FLAIR² in LesionTOADS, OASIS and the lesion segmentation toolbox (LST) when applied to non-homogenized, multi-center 2D-imaging data.

Methods

Lesions were segmented on 47 MS patient data sets obtained from 34 sites using LesionTOADS, OASIS and LST, and compared to a semi-automatically generated reference. The performance of FLAIR and FLAIR² was assessed using the relative lesion volume difference (LVD), Dice coefficient (DSC), sensitivity (SEN) and symmetric surface distance (SSD). Performance improvements related to lesion volumes (LVs) were evaluated for all tools. For comparison, LesionTOADS was also used to segment lesions from 3 T single-center MR data of 40 clinically isolated syndrome (CIS) patients.

Results

Compared to FLAIR, the use of FLAIR² in LesionTOADS led to improvements of 31.6% (LVD), 14.0% (DSC), 25.1% (SEN), and 47.0% (SSD) in the multi-center study. DSC and SSD significantly improved for larger LVs, while LVD and SEN were enhanced independent of LV. OASIS showed little difference between FLAIR and FLAIR², likely due to its inherent use of T₂w and FLAIR. LST replicated the benefits of FLAIR² only in part, indicating that further optimization, particularly at low LVs is needed. In the CIS study, LesionTOADS did not benefit from the use of FLAIR² as the segmentation performance for both FLAIR and FLAIR² was heterogeneous.

Conclusions

In this real-world, multi-center experiment, FLAIR² outperformed FLAIR in its ability to segment MS lesions with LesionTOADS. The computation of FLAIR² enhanced lesion detection, at minimally increased computational time or cost, even retrospectively. Further work is needed to determine how LesionTOADS and other tools, such as LST, can optimally benefit from the improved FLAIR² contrast.

•

FLAIR² improves automatic MS lesion segmentation with LesionTOADS compared to FLAIR.

•

Segmentation similarity improves for higher lesion volumes, particularly for FLAIR².

•

FLAIR² provides greater sensitivity independent of lesion volume than FLAIR alone.

•

Other segmentation tools need further optimization to fully benefit from FLAIR².

•

FLAIR² provides immediate benefits at 1.5 T and visually improves segmentation at 3 T.

Automatic Labeling of MR Brain Images Through the Hashing Retrieval Based Atlas Forest

The multi-atlas method is one of the efficient and common automatic labeling method, which uses the prior information provided by expert-labeled images to guide the labeling of the target. However, most multi-atlas-based methods depend on the registration that may not give the correct information during the label propagation. To address the issue, we designed a new automatic labeling method through the hashing retrieval based atlas forest. The proposed method propagates labels without registration to reduce the errors, and constructs a target-oriented learning model to integrate information among the atlases. This method innovates a coarse classification strategy to preprocess the dataset, which retains the integrity of dataset and reduces computing time. Furthermore, the method considers each voxel in the atlas as a sample and encodes these samples with hashing for the fast sample retrieval. In the stage of labeling, the method selects suitable samples through hashing learning and trains atlas forests by integrating the information from the dataset. Then, the trained model is used to predict the labels of the target. Experimental results on two datasets illustrated that the proposed method is promising in the automatic labeling of MR brain images.

Follow-up MRI in multiple sclerosis patients: automated co-registration and lesion color-coding improves diagnostic accuracy and reduces reading time

OBJECTIVES

In multiple sclerosis (MS), the heterogeneous and numerous appearances of lesions may impair diagnostic accuracy. This study investigates if a combined automated co-registration and lesion color-coding method (AC) improves assessment of MS follow-up MRI compared with conventional reading (CR).

METHODS

We retrospectively assessed 70 follow-up MRI of 53 patients. Heterogeneous datasets of diverse scanners and institutions were used. Two readers determined presence of (a) progression, (b) regression, (c) mixed change, or (d) stable disease between the two examinations using corresponding FLAIR sequences in CR and AC-assisted reading. Consensus reference reading was provided by two blinded radiologists. Kappa statistics tested interrater agreement, McNemar's test dichotomous variables, and Wilcoxon's test continuous variables (statistical significance p ≤ 0.05).

RESULTS

The cohort comprised 41 female and 12 male patients with a mean age of 40 (± 14) years. Average rating time was reduced from 78 (± 36) to 44 (±22) s with the AC approach (p < 0.001). The time needed to start and match datasets with AC was 14 (± 1) s. Compared with CR, AC improved interrater agreement, both between raters (0.52 vs. 0.67) and between raters and consensus reference reading (0.47/0.5 vs. 0.83/0.78). Compared with CR, the diagnostic accuracy increased from 67 to 90% (reader 1, p < 0.01) and from 70 to 87% (reader 2, p < 0.05) in the AC-assisted reading.

CONCLUSIONS

Compared with CR, automated co-registration and lesion color-coding of MS-associated FLAIR-lesions in follow-up MRI increased diagnostic accuracy and reduced the time required for follow-up evaluation significantly. The AC algorithm therefore appears to be helpful to improve MS follow-up assessments in clinical routine.

KEY POINTS

• Automated co-registration and lesion color-coding increases diagnostic accuracy in the assessment of MRI follow-up examinations in patients with multiple sclerosis. • Automated co-registration and lesion color-coding reduces reading time of MRI follow-up examinations in patients with multiple sclerosis. • Automated co-registration and lesion color-coding improved interrater agreement in the assessment of MRI follow-up examinations in patients with multiple sclerosis.

Automatic Quantification of Computed Tomography Features in Acute Traumatic Brain Injury

Traumatic brain injury is a complex and diverse medical condition with a high frequency of intracranial abnormalities. These can typically be visualized on a computed tomography (CT) scan, which provides important information for further patient management, such as the need for operative intervention. In order to quantify the extent of acute intracranial lesions and associated secondary injuries, such as midline shift and cisternal compression, visual assessment of CT images has limitations, including observer variability and lack of quantitative interpretation. Automated image analysis can quantify the extent of intracranial abnormalities and provide added value in routine clinical practice. In this article, we present icobrain, a fully automated method that reliably computes acute intracranial lesions volume based on deep learning, cistern volume, and midline shift on the noncontrast CT image of a patient. The accuracy of our method is evaluated on a subset of the multi-center data set from the CENTER-TBI (Collaborative European Neurotrauma Effectiveness Research in Traumatic Brain Injury) study for which expert annotations were used as a reference. Median volume differences between expert assessments and icobrain are 0.07 mL for acute intracranial lesions and -0.01 mL for cistern segmentation. Correlation between expert assessments and icobrain is 0.91 for volume of acute intracranial lesions and 0.94 for volume of the cisterns. For midline shift computations, median error is -0.22 mm, with a correlation of 0.93 with expert assessments.

Automated segmentation of changes in FLAIR-hyperintense white matter lesions in multiple sclerosis on serial magnetic resonance imaging

Longitudinal analysis of white matter lesion changes on serial MRI has become an important parameter to study diseases with white-matter lesions. Here, we build on earlier work on cross-sectional lesion segmentation; we present a fully automatic pipeline for serial analysis of FLAIR-hyperintense white matter lesions. Our algorithm requires three-dimensional gradient echo T1- and FLAIR- weighted images at 3 Tesla as well as available cross-sectional lesion segmentations of both time points. Preprocessing steps include lesion filling and intrasubject registration. For segmentation of lesion changes, initial lesion maps of different time points are fused; herein changes in intensity are analyzed at the voxel level. Significance of lesion change is estimated by comparison with the difference distribution of FLAIR intensities within normal appearing white matter. The method is validated on MRI data of two time points from 40 subjects with multiple sclerosis derived from two different scanners (20 subjects per scanner). Manual segmentation of lesion increases served as gold standard. Across all lesion increases, voxel-wise Dice coefficient (0.7) as well as lesion-wise detection rate (0.8) and false-discovery rate (0.2) indicate good overall performance. Analysis of scans from a repositioning experiment in a single patient with multiple sclerosis did not yield a single false positive lesion. We also introduce the lesion change plot as a descriptive tool for the lesion change of individual patients with regard to both number and volume. An open source implementation of the algorithm is available at http://www.statistical-modeling.de/lst.html.

•

Quantification of white matter lesion changes is important in multiple sclerosis.

•

We developed and validated an algorithm for automated detection of lesion changes.

•

Our algorithm requires T1-weighted and FLAIR images derived at 3 T as well as available cross-sectional lesion segmentations.

•

With data from 2 different scanners, the tool showed good agreement with manual tracing.

•

An open-source application is available.

Radiological images and machine learning: Trends, perspectives, and prospects

The application of machine learning to radiological images is an increasingly active research area that is expected to grow in the next five to ten years. Recent advances in machine learning have the potential to recognize and classify complex patterns from different radiological imaging modalities such as x-rays, computed tomography, magnetic resonance imaging and positron emission tomography imaging. In many applications, machine learning based systems have shown comparable performance to human decision-making. The applications of machine learning are the key ingredients of future clinical decision making and monitoring systems. This review covers the fundamental concepts behind various machine learning techniques and their applications in several radiological imaging areas, such as medical image segmentation, brain function studies and neurological disease diagnosis, as well as computer-aided systems, image registration, and content-based image retrieval systems. Synchronistically, we will briefly discuss current challenges and future directions regarding the application of machine learning in radiological imaging. By giving insight on how take advantage of machine learning powered applications, we expect that clinicians can prevent and diagnose diseases more accurately and efficiently.

Robust, atlas-free, automatic segmentation of brain MRI in health and disease

Background

Brain- and lesion-volumes derived from magnetic resonance images (MRI) serve as important imaging markers of disease progression in neurodegenerative diseases and aging. While manual segmentation of these volumes is both tedious and impractical in large cohorts of subjects, automated segmentation methods often fail in accurate segmentation of brains with severe atrophy or high lesion loads. The purpose of this study was to develop an atlas-free brain Classification using DErivative-based Features (C-DEF), which utilizes all scans that may be acquired during the course of a routine MRI study at any center.

Methods

Proton-density, T₂-weighted, T₁-weighted, brain-free water, 3D FLAIR, 3D T₂-weighted, and 3D T₂*-weighted images, collected routinely on patients with neuroinflammatory diseases at the NIH, were used to optimize the C-DEF algorithm on healthy volunteers and HIV + subjects (cohort 1). First, manually marked lesions and eroded FreeSurfer brain segmentation masks (compiled into gray and white matter, globus pallidus, CSF labels) were used in training. Next, the optimized C-DEF was applied on a separate cohort of HIV + subjects (cohort two), and the results were compared with that of FreeSurfer and Lesion-TOADS. Finally, C-DEF segmentation was evaluated on subjects clinically diagnosed with various other neurological diseases (cohort three).

Results

C-DEF algorithm was optimized using leave-one-out cross validation on five healthy subjects (age 36 ± 11 years), and five subjects infected with HIV (age 57 ± 2.6 years) in cohort one. The optimized C-DEF algorithm outperformed FreeSurfer and Lesion-TOADS segmentation in 49 other subjects infected with HIV (cohort two, age 54 ± 6 years) in qualitative and quantitative comparisons. Although trained only on HIV brains, sensitivity to detect lesions using C-DEF increased by 45% in HTLV-I-associated myelopathy/tropical spastic paraparesis (n = 5; age 58 ± 7 years), 33% in multiple sclerosis (n = 5; 42 ± 9 years old), and 4% in subjects with polymorphism of the cytotoxic T-lymphocyte-associated protein 4 gene (n = 5; age 24 ± 12 years) compared to Lesion-TOADS.

Conclusion

C-DEF outperformed other segmentation algorithms in the various neurological diseases explored herein, especially in lesion segmentation. While the results reported are from routine images acquired at the NIH, the algorithm can be easily trained and optimized for any set of contrasts and protocols for wider application. We are currently exploring various technical aspects of optimal implementation of CDEF in a clinical setting and evaluating a larger cohort of patients with other neurological diseases. Improving the accuracy of brain segmentation methodology will help better understand the relationship of imaging abnormalities to clinical and neuropsychological markers in disease.

Automatic brain lesion segmentation on standard MRIs of the human head: a scoping review protocol

INTRODUCTION

Automatic brain lesion segmentation from medical images has great potential to support clinical decision making. Although numerous methods have been proposed, significant challenges must be addressed before they will become established in clinical and research practice. We aim to elucidate the state of the art, to provide a synopsis of competing approaches and identify contrasts between them.

METHODS AND ANALYSIS

We present the background and study design of a scoping review for automatic brain lesion segmentation methods for conventional MRI according to the framework proposed by Arksey and O'Malley. We aim to identify common image processing steps as well as mathematical and computational theories implemented in these methods. We will aggregate the evidence on the efficacy and identify limitations of the approaches. Methods to be investigated work with standard MRI sequences from human patients examined for brain lesions, and are validated with quantitative measures against a trusted reference. PubMed, IEEE Xplore and Scopus will be searched using search phrases that will ensure an inclusive and unbiased overview. For matching records, titles and abstracts will be screened to ensure eligibility. Studies will be excluded if a full paper is not available or is not written in English, if non-standard MR sequences are used, if there is no quantitative validation, or if the method is not automatic. In the data charting phase, we will extract information about authors, publication details and study cohort. We expect to find information about preprocessing, segmentation and validation procedures. We will develop an analytical framework to collate, summarise and synthesise the data.

ETHICS AND DISSEMINATION

Ethical approval for this study is not required since the information will be extracted from published studies. We will submit the review report to a peer-reviewed scientific journal and explore other venues for presenting the work.

Automatic segmentation of the spinal cord and intramedullary multiple sclerosis lesions with convolutional neural networks

The spinal cord is frequently affected by atrophy and/or lesions in multiple sclerosis (MS) patients. Segmentation of the spinal cord and lesions from MRI data provides measures of damage, which are key criteria for the diagnosis, prognosis, and longitudinal monitoring in MS. Automating this operation eliminates inter-rater variability and increases the efficiency of large-throughput analysis pipelines. Robust and reliable segmentation across multi-site spinal cord data is challenging because of the large variability related to acquisition parameters and image artifacts. In particular, a precise delineation of lesions is hindered by a broad heterogeneity of lesion contrast, size, location, and shape. The goal of this study was to develop a fully-automatic framework - robust to variability in both image parameters and clinical condition - for segmentation of the spinal cord and intramedullary MS lesions from conventional MRI data of MS and non-MS cases. Scans of 1042 subjects (459 healthy controls, 471 MS patients, and 112 with other spinal pathologies) were included in this multi-site study (n = 30). Data spanned three contrasts (T1-, T2-, and T2∗-weighted) for a total of 1943 vol and featured large heterogeneity in terms of resolution, orientation, coverage, and clinical conditions. The proposed cord and lesion automatic segmentation approach is based on a sequence of two Convolutional Neural Networks (CNNs). To deal with the very small proportion of spinal cord and/or lesion voxels compared to the rest of the volume, a first CNN with 2D dilated convolutions detects the spinal cord centerline, followed by a second CNN with 3D convolutions that segments the spinal cord and/or lesions. CNNs were trained independently with the Dice loss. When compared against manual segmentation, our CNN-based approach showed a median Dice of 95% vs. 88% for PropSeg (p ≤ 0.05), a state-of-the-art spinal cord segmentation method. Regarding lesion segmentation on MS data, our framework provided a Dice of 60%, a relative volume difference of -15%, and a lesion-wise detection sensitivity and precision of 83% and 77%, respectively. In this study, we introduce a robust method to segment the spinal cord and intramedullary MS lesions on a variety of MRI contrasts. The proposed framework is open-source and readily available in the Spinal Cord Toolbox.

Quantitative histopathologic assessment of perfusion MRI as a marker of glioblastoma cell infiltration in and beyond the peritumoral edema region

BACKGROUND

Conventional MRI fails to detect regions of glioblastoma cell infiltration beyond the contrast-enhanced T₁ solid tumor region, with infiltrating tumor cells often migrating along host blood vessels.

PURPOSE

To quantitatively and qualitatively analyze the correlation between perfusion MRI signal and tumor cell density in order to assess whether local perfusion perturbation could provide a useful biomarker of glioblastoma cell infiltration.

STUDY TYPE

Animal model.

SUBJECTS

Mice bearing orthotopic glioblastoma xenografts generated from a patient-derived glioblastoma cell line.

FIELD STRENGTH/SEQUENCES

7T perfusion images acquired using a high signal-to-noise ratio (SNR) multiple boli arterial spin labeling sequence were compared with conventional MRI (T₁ /T₂ weighted, contrast-enhanced T₁ , diffusion-weighted, and apparent diffusion coefficient).

ASSESSMENT

Immunohistochemistry sections were stained for human leukocyte antigen (probing human-derived tumor cells). To achieve quantitative MRI-tissue comparison, multiple histological slices cut in the MRI plane were stacked to produce tumor cell density maps acting as a "ground truth."

STATISTICAL TESTS

Sensitivity, specificity, accuracy, and Dice similarity indices were calculated and a two-tailed, paired t-test used for statistical analysis.

RESULTS

High comparison test results (Dice 0.62-0.72, Accuracy 0.86-0.88, Sensitivity 0.51-0.7, and Specificity 0.92-0.97) indicate a good segmentation for all imaging modalities and highlight the quality of the MRI tissue assessment protocol. Perfusion imaging exhibits higher sensitivity (0.7) than conventional MRI (0.51-0.61). MRI/histology voxel-to-voxel comparison revealed a negative correlation between tumor cell infiltration and perfusion at the tumor margins (P = 0.0004).

DATA CONCLUSION

These results demonstrate the ability of perfusion imaging to probe regions of low tumor cell infiltration while confirming the sensitivity limitations of conventional imaging modalities. The quantitative relationship between tumor cell density and perfusion identified in and beyond the edematous T₂ hyperintensity region surrounding macroscopic tumor could be used to detect marginal tumor cell infiltration with greater accuracy.

LEVEL OF EVIDENCE

1 Technical stage: 2 J. Magn. Reson. Imaging 2019;50:529-540.

An overview of deep learning in medical imaging focusing on MRI

What has happened in machine learning lately, and what does it mean for the future of medical image analysis? Machine learning has witnessed a tremendous amount of attention over the last few years. The current boom started around 2009 when so-called deep artificial neural networks began outperforming other established models on a number of important benchmarks. Deep neural networks are now the state-of-the-art machine learning models across a variety of areas, from image analysis to natural language processing, and widely deployed in academia and industry. These developments have a huge potential for medical imaging technology, medical data analysis, medical diagnostics and healthcare in general, slowly being realized. We provide a short overview of recent advances and some associated challenges in machine learning applied to medical image processing and image analysis. As this has become a very broad and fast expanding field we will not survey the entire landscape of applications, but put particular focus on deep learning in MRI. Our aim is threefold: (i) give a brief introduction to deep learning with pointers to core references; (ii) indicate how deep learning has been applied to the entire MRI processing chain, from acquisition to image retrieval, from segmentation to disease prediction; (iii) provide a starting point for people interested in experimenting and perhaps contributing to the field of deep learning for medical imaging by pointing out good educational resources, state-of-the-art open-source code, and interesting sources of data and problems related medical imaging.

Asymmetric Loss Functions and Deep Densely Connected Networks for Highly Imbalanced Medical Image Segmentation: Application to Multiple Sclerosis Lesion Detection

Fully convolutional deep neural networks have been asserted to be fast and precise frameworks with great potential in image segmentation. One of the major challenges in training such networks raises when data is unbalanced, which is common in many medical imaging applications such as lesion segmentation where lesion class voxels are often much lower in numbers than non-lesion voxels. A trained network with unbalanced data may make predictions with high precision and low recall, being severely biased towards the non-lesion class which is particularly undesired in most medical applications where false negatives are actually more important than false positives. Various methods have been proposed to address this problem including two step training, sample re-weighting, balanced sampling, and more recently similarity loss functions, and focal loss. In this work we trained fully convolutional deep neural networks using an asymmetric similarity loss function to mitigate the issue of data imbalance and achieve much better trade-off between precision and recall. To this end, we developed a 3D fully convolutional densely connected network (FC-DenseNet) with large overlapping image patches as input and an asymmetric similarity loss layer based on Tversky index (using F β scores). We used large overlapping image patches as inputs for intrinsic and extrinsic data augmentation, a patch selection algorithm, and a patch prediction fusion strategy using B-spline weighted soft voting to account for the uncertainty of prediction in patch borders. We applied this method to multiple sclerosis (MS) lesion segmentation based on two different datasets of MSSEG 2016 and ISBI longitudinal MS lesion segmentation challenge, where we achieved average Dice similarity coefficients of 69.9% and 65.74%, respectively, achieving top performance in both challenges. We compared the performance of our network trained with F β loss, focal loss, and generalized Dice loss (GDL) functions. Through September 2018 our network trained with focal loss ranked first according to the ISBI challenge overall score and resulted in the lowest reported lesion false positive rate among all submitted methods. Our network trained with the asymmetric similarity loss led to the lowest surface distance and the best lesion true positive rate that is arguably the most important performance metric in a clinical decision support system for lesion detection. The asymmetric similarity loss function based on F β scores allows training networks that make a better balance between precision and recall in highly unbalanced image segmentation. We achieved superior performance in MS lesion segmentation using a patchwise 3D FC-DenseNet with a patch prediction fusion strategy, trained with asymmetric similarity loss functions.

Detection of white matter lesion regions in MRI using SLIC0 and convolutional neural network

BACKGROUND AND OBJECTIVE

White matter lesions are non-static brain lesions that have a prevalence rate up to 98% in the elderly population. Because they may be associated with several brain diseases, it is important that they are detected as soon as possible. Magnetic Resonance Imaging (MRI) provides three-dimensional data with the possibility to detect and emphasize contrast differences in soft tissues, providing rich information about the human soft tissue anatomy. However, the amount of data provided for these images is far too much for manual analysis/interpretation, representing a difficult and time-consuming task for specialists. This work presents a computational methodology capable of detecting regions of white matter lesions of the brain in MRI of FLAIR modality. The techniques highlighted in this methodology are SLIC0 clustering for candidate segmentation and convolutional neural networks for candidate classification.

METHODS

The methodology proposed here consists of four steps: (1) images acquisition, (2) images preprocessing, (3) candidates segmentation and (4) candidates classification.

RESULTS

The methodology was applied on 91 magnetic resonance images provided by DASA, and achieved an accuracy of 98.73%, specificity of 98.77% and sensitivity of 78.79% with 0.005 of false positives, without any false positives reduction technique, in detection of white matter lesion regions.

CONCLUSIONS

It is demonstrated the feasibility of the analysis of brain MRI using SLIC0 and convolutional neural network techniques to achieve success in detection of white matter lesions regions.

A Novel Classification Method using Effective Neural Network and Quantitative Magnetization Transfer Imaging of Brain White Matter in Relapsing Remitting Multiple Sclerosis

BACKGROUND

Quantitative Magnetization Transfer Imaging (QMTI) is often used to quantify the myelin content in multiple sclerosis (MS) lesions and normal appearing brain tissues. Also, automated classifiers such as artificial neural networks (ANNs) can significantly improve the identification and classification processes of MS clinical datasets.

OBJECTIVE

We classified patients with relapsing-remitting multiple sclerosis (RRMS) from healthy subjects using QMTI and T1 longitudinal relaxation time data of brain white matter, then the performance of three ANN-based classifiers have been investigated.

MATERIALS AND METHODS

The input features of ANN algorithms, including multilayer perceptron (MLP), radial basis function (RBF) and ensemble neural networks based on Akaike information criterion (ENN-AIC) were extracted in the form of QMTI and T1 mean values from parametric maps. The ANNs quantitative performance is measured by the standard evaluation of confusion matrix criteria.

RESULTS

The results indicate that ENN-AIC-based classification method has achieved 90% accuracy, 92% sensitivity and 86% precision compared to other ANN models. NPV, FPR and FDR values were found to be 0.933, 0.125 and 0.133, respectively, according to the proposed ENN-AIC model. A graphical representation of how to track actual data by the predictive values derived from ANN algorithms, was also presented.

CONCLUSION

It has been demonstrated that ENN-AIC as an effective neural network improves the quality of classification results compared to MLP and RBF.In addition, this research provides a new direction to classify a large amount of quantitative MRI data that can help the physician in a correct MS diagnosis.

A Novel Classification Method using Effective Neural Network and Quantitative Magnetization Transfer Imaging of Brain White Matter in Relapsing Remitting Multiple Sclerosis

Background

Quantitative Magnetization Transfer Imaging (QMTI) is often used to quantify the myelin content in multiple sclerosis (MS) lesions and normal appearing brain tissues. Also, automated classifiers such as artificial neural networks (ANNs) can significantly improve the identification and classification processes of MS clinical datasets.

Objective

We classified patients with relapsing-remitting multiple sclerosis (RRMS) from healthy subjects using QMTI and T1 longitudinal relaxation time data of brain white matter, then the performance of three ANN-based classifiers have been investigated.

Materials and Methods

The input features of ANN algorithms, including multilayer perceptron (MLP), radial basis function (RBF) and ensemble neural networks based on Akaike information criterion (ENN-AIC) were extracted in the form of QMTI and T1 mean values from parametric maps. The ANNs quantitative performance is measured by the standard evaluation of confusion matrix criteria.

Results

The results indicate that ENN-AIC-based classification method has achieved 90% accuracy, 92% sensitivity and 86% precision compared to other ANN models. NPV, FPR and FDR values were found to be 0.933, 0.125 and 0.133, respectively, according to the proposed ENN-AIC model. A graphical representation of how to track actual data by the predictive values derived from ANN algorithms, was also presented.

Conclusion

It has been demonstrated that ENN-AIC as an effective neural network improves the quality of classification results compared to MLP and RBF.In addition, this research provides a new direction to classify a large amount of quantitative MRI data that can help the physician in a correct MS diagnosis.

A dual modeling approach to automatic segmentation of cerebral T2 hyperintensities and T1 black holes in multiple sclerosis

BACKGROUND AND PURPOSE

Magnetic resonance imaging (MRI) is crucial for in vivo detection and characterization of white matter lesions (WML) in multiple sclerosis (MS). The most widely established MRI outcome measure is the volume of hyperintense lesions on T2-weighted images (T2L). Unfortunately, T2L are non-specific for the level of tissue destruction and show a weak relationship to clinical status. Interest in lesions that appear hypointense on T1-weighted images (T1L) ("black holes") has grown because T1L provide more specificity for axonal loss and a closer link to neurologic disability. The technical difficulty of T1L segmentation has led investigators to rely on time-consuming manual assessments prone to inter- and intra-rater variability. This study aims to develop an automatic T1L segmentation approach, adapted from a T2L segmentation algorithm.

MATERIALS AND METHODS

T1, T2, and fluid-attenuated inversion recovery (FLAIR) sequences were acquired from 40 MS subjects at 3 Tesla (3 T). T2L and T1L were manually segmented. A Method for Inter-Modal Segmentation Analysis (MIMoSA) was then employed.

RESULTS

Using cross-validation, MIMoSA proved to be robust for segmenting both T2L and T1L. For T2L, a Sørensen-Dice coefficient (DSC) of 0.66 and partial AUC (pAUC) up to 1% false positive rate of 0.70 were achieved. For T1L, 0.53 DSC and 0.64 pAUC were achieved. Manual and MIMoSA segmented volumes were correlated and resulted in 0.88 for T1L and 0.95 for T2L. The correlation between Expanded Disability Status Scale (EDSS) scores and manual versus automatic volumes were similar for T1L (0.32 manual vs. 0.34 MIMoSA), T2L (0.33 vs. 0.32), and the T1L/T2L ratio (0.33 vs 0.33).

CONCLUSIONS

Though originally designed to segment T2L, MIMoSA performs well for segmenting T1 black holes in patients with MS.

Strong and specific associations between cardiovascular risk factors and white matter micro- and macrostructure in healthy aging

Cardiovascular health declines with age, increasing the risk of hypertension and elevated heart rate in middle and old age. Here, we used multivariate techniques to investigate the associations between cardiovascular health (diastolic blood pressure, systolic blood pressure, and heart rate) and white matter macrostructure (lesion volume and number) and microstructure (as measured by diffusion-weighted imaging) in the cross-sectional, population-based Cam-CAN cohort (N = 667, aged 18–88). We found that cardiovascular health and age made approximately similar contributions to white matter health and explained up to 56% of variance therein. Lower diastolic blood pressure, higher systolic blood pressure, and higher heart rate were each strongly, and independently, associated with white matter abnormalities on all indices. Body mass and exercise were associated with white matter health, both directly and indirectly via cardiovascular health. These results highlight the importance of cardiovascular risk factors for white matter health across the adult lifespan and suggest that systolic blood pressure, diastolic blood pressure, and heart rate affect white matter health via separate mechanisms.

•

Cardiovascular health is related to white matter lesion burden and diffusivity.

•

Low diastolic pressure, high systolic pressure, and higher heart rate contribute independently.

•

Cardiovascular health and age explain up to 56% of variance in white matter health.

•

The uncinate fasciculus, inferior fronto-occipital fasciculus, and forceps minor show most sensitivity.

•

Lower BMI and more exercise may have protective effects.

Objective Evaluation of Multiple Sclerosis Lesion Segmentation using a Data Management and Processing Infrastructure

We present a study of multiple sclerosis segmentation algorithms conducted at the international MICCAI 2016 challenge. This challenge was operated using a new open-science computing infrastructure. This allowed for the automatic and independent evaluation of a large range of algorithms in a fair and completely automatic manner. This computing infrastructure was used to evaluate thirteen methods of MS lesions segmentation, exploring a broad range of state-of-theart algorithms, against a high-quality database of 53 MS cases coming from four centers following a common definition of the acquisition protocol. Each case was annotated manually by an unprecedented number of seven different experts. Results of the challenge highlighted that automatic algorithms, including the recent machine learning methods (random forests, deep learning, …), are still trailing human expertise on both detection and delineation criteria. In addition, we demonstrate that computing a statistically robust consensus of the algorithms performs closer to human expertise on one score (segmentation) although still trailing on detection scores.

Compromised prefrontal structure and function are associated with slower walking in older adults

Our previous work demonstrates that reduced activation of the executive network is associated with slow walking speed in a cohort of older adults from the MOBILIZE Boston Study. However, the influence of underlying white matter integrity on the activation of this network and walking speed is unknown. Thus, we used diffusion-weighted imaging and fMRI during an n-back task to assess associations between executive network structure, function, and walking speed. Whole-brain tract-based spatial statistics (TBSS) were used to identify regions of white matter microstructural integrity that were associated with walking speed. The integrity of these regions was then entered into multiple regression models to predict task performance and executive network activation during the n-back task. Among the significant associations of FA with walking speed, we observed the anterior thalamic radiation and superior longitudinal fasciculus were further associated with both n-back response speed and executive network activation. These findings suggest that subtle damage to frontal white matter may contribute to altered executive network activation and slower walking in older adults.

•

Older adult walking speed was not associated with white matter lesion burden.

•

Walking speed was associated with microstructural white matter integrity.

•

The integrity of prefrontal areas was associated with executive network activation.

•

Low executive network activation also corresponded to slower walking.

•

Interventions targeting the executive network may preserve older adult mobility.

Structural neuroimaging as clinical predictor: A review of machine learning applications

In this paper, we provide an extensive overview of machine learning techniques applied to structural magnetic resonance imaging (MRI) data to obtain clinical classifiers. We specifically address practical problems commonly encountered in the literature, with the aim of helping researchers improve the application of these techniques in future works. Additionally, we survey how these algorithms are applied to a wide range of diseases and disorders (e.g. Alzheimer's disease (AD), Parkinson's disease (PD), autism, multiple sclerosis, traumatic brain injury, etc.) in order to provide a comprehensive view of the state of the art in different fields.

Current Applications and Future Impact of Machine Learning in Radiology

Recent advances and future perspectives of machine learning techniques offer promising applications in medical imaging. Machine learning has the potential to improve different steps of the radiology workflow including order scheduling and triage, clinical decision support systems, detection and interpretation of findings, postprocessing and dose estimation, examination quality control, and radiology reporting. In this article, the authors review examples of current applications of machine learning and artificial intelligence techniques in diagnostic radiology. In addition, the future impact and natural extension of these techniques in radiology practice are discussed.

Voxel-Wise Logistic Regression and Leave-One-Source-Out Cross Validation for white matter hyperintensity segmentation

Many algorithms have been proposed for automated segmentation of white matter hyperintensities (WMH) in brain MRI. Yet, broad uptake of any particular algorithm has not been observed. In this work, we argue that this may be due to variable and suboptimal validation data and frameworks, precluding direct comparison of methods on heterogeneous data. As a solution, we present Leave-One-Source-Out Cross Validation (LOSO-CV), which leverages all available data for performance estimation, and show that this gives more realistic (lower) estimates of segmentation algorithm performance on data from different scanners. We also develop a FLAIR-only WMH segmentation algorithm: Voxel-Wise Logistic Regression (VLR), inspired by the open-source Lesion Prediction Algorithm (LPA). Our variant facilitates more accurate parameter estimation, and permits intuitive interpretation of model parameters. We illustrate the performance of the VLR algorithm using the LOSO-CV framework with a dataset comprising freely available data from several recent competitions (96 images from 7 scanners). The performance of the VLR algorithm (median Similarity Index of 0.69) is compared to its LPA predecessor (0.58), and the results of the VLR algorithm in the 2017 WMH Segmentation Competition are also presented.

Improved Automatic Segmentation of White Matter Hyperintensities in MRI Based on Multilevel Lesion Features

Brain white matter hyperintensities (WMHs) are linked to increased risk of cerebrovascular and neurodegenerative diseases among the elderly. Consequently, detection and characterization of WMHs are of significant clinical importance. We propose a novel approach for WMH segmentation from multi-contrast MRI where both voxel-based and lesion-based information are used to improve overall performance in both volume-oriented and object-oriented metrics. Our segmentation method (AMOS-2D) consists of four stages following a "generate-and-test" approach: pre-processing, Gaussian white matter (WM) modelling, hierarchical multi-threshold WMH segmentation and object-based WMH filtering using support vector machines. Data from 28 subjects was used in this study covering a wide range of lesion loads. Volumetric T1-weighted images and 2D fluid attenuated inversion recovery (FLAIR) images were used as basis for the WM model and lesion masks defined manually in each subject by experts were used for training and evaluating the proposed method. The method obtained an average agreement (in terms of the Dice similarity coefficient, DSC) with experts equivalent to inter-expert agreement both in terms of WMH number (DSC = 0.637 vs. 0.651) and volume (DSC = 0.743 vs. 0.781). It allowed higher accuracy in detecting WMH compared to alternative methods tested and was further found to be insensitive to WMH lesion burden. Good agreement with expert annotations combined with stable performance largely independent of lesion burden suggests that AMOS-2D will be a valuable tool for fully automated WMH segmentation in patients with cerebrovascular and neurodegenerative pathologies.

Supervoxel based method for multi-atlas segmentation of brain MR images

Multi-atlas segmentation has been widely applied to the analysis of brain MR images. However, the state-of-the-art techniques in multi-atlas segmentation, including both patch-based and learning-based methods, are strongly dependent on the pairwise registration or exhibit huge spatial inconsistency. The paper proposes a new segmentation framework based on supervoxels to solve the existing challenges of previous methods. The supervoxel is an aggregation of voxels with similar attributes, which can be used to replace the voxel grid. By formulating the segmentation as a tissue labeling problem associated with a maximum-a-posteriori inference in Markov random field, the problem is solved via a graphical model with supervoxels being considered as the nodes. In addition, a dense labeling scheme is developed to refine the supervoxel labeling results, and the spatial consistency is incorporated in the proposed method. The proposed approach is robust to the pairwise registration errors and of high computational efficiency. Extensive experimental evaluations on three publically available brain MR datasets demonstrate the effectiveness and superior performance of the proposed approach.

Partial volume-aware assessment of multiple sclerosis lesions

White-matter lesion count and volume estimation are key to the diagnosis and monitoring of multiple sclerosis (MS). Automated MS lesion segmentation methods that have been proposed in the past 20 years reach their limits when applied to patients in early disease stages characterized by low lesion load and small lesions. We propose an algorithm to automatically assess MS lesion load (number and volume) while taking into account the mixing of healthy and lesional tissue in the image voxels due to partial volume effects. The proposed method works on 3D MPRAGE and 3D FLAIR images as obtained from current routine MS clinical protocols. The method was evaluated and compared with manual segmentation on a cohort of 39 early-stage MS patients with low disability, and showed higher Dice similarity coefficients (median DSC = 0.55) and higher detection rate (median DR = 61%) than two widely used methods (median DSC = 0.50, median DR < 45%) for automated MS lesion segmentation. We argue that this is due to the higher performance in segmentation of small lesions, which are inherently prone to partial volume effects.

•

Modeling the partial volume improves lesion volumetric measurements.

•

Higher detection of small lesions inherently prone to partial volume effects.

•

Partial volume effects should be taken into account in early stages of MS.

Automated Segmentation of Head Computed Tomography Images Using FSL

OBJECTIVE

The aim of this study was to investigate the use of one magnetic resonance image-processing tool, FSL, in its ability to perform automated segmentation of computed tomographic images of the brain.

METHODS

Head computed tomography (CT) images were brain extracted and segmented using the FSL tools BET and FAST, respectively. The products of segmentation were analyzed by histogram. The impact of image intensity inhomogeneity correction was investigated using simulated bias fields, 14 routine head CT scans, and selected illustrative clinical cases.

RESULTS

FSL FAST performs direct segmentation of head CT images, permitting quantitation of gray and white matter densities and volumes, achieving a more complete segmentation than masking methods. "Bias field correction" reduced the covariance of image signal intensities of the total brain and gray matter images (P < 0.01). Correction is larger when the effects of beam hardening and radiation scatter are larger, resulting in improved segmentation.

CONCLUSIONS

FSL FAST enables direct segmentation of head CT images.