Automated Detection of Lupus White Matter Lesions in MRI

Simulation of acquisition shifts in T2 weighted fluid-attenuated inversion recovery magnetic resonance images to stress test artificial intelligence segmentation networks

Purpose

To provide a simulation framework for routine neuroimaging test data, which allows for "stress testing" of deep segmentation networks against acquisition shifts that commonly occur in clinical practice for T2 weighted (T2w) fluid-attenuated inversion recovery magnetic resonance imaging protocols.

Approach

The approach simulates "acquisition shift derivatives" of MR images based on MR signal equations. Experiments comprise the validation of the simulated images by real MR scans and example stress tests on state-of-the-art multiple sclerosis lesion segmentation networks to explore a generic model function to describe the F1 score in dependence of the contrast-affecting sequence parameters echo time (TE) and inversion time (TI).

Results

The differences between real and simulated images range up to 19% in gray and white matter for extreme parameter settings. For the segmentation networks under test, the F1 score dependency on TE and TI can be well described by quadratic model functions (R 2 > 0.9 ). The coefficients of the model functions indicate that changes of TE have more influence on the model performance than TI.

Conclusions

We show that these deviations are in the range of values as may be caused by erroneous or individual differences in relaxation times as described by literature. The coefficients of the F1 model function allow for a quantitative comparison of the influences of TE and TI. Limitations arise mainly from tissues with a low baseline signal (like cerebrospinal fluid) and when the protocol contains contrast-affecting measures that cannot be modeled due to missing information in the DICOM header.

Associations of intracranial artery length and branch number on non-contrast enhanced MRA with cognitive impairment in individuals with carotid atherosclerosis

Developing novel risk markers for vascular contributions to cognitive impairment and dementia is important. This study aimed to extract total length, branch number and average tortuosity of intracranial distal arteries (A2, M2, P2 and more distal) from non-contrast enhanced magnetic resonance angiography (NCE-MRA) images, and explore their associations with global cognition. In 29 subjects (aged 40-90 years) with carotid atherosclerotic disease, the 3 intracranial vascular features on two NCE-MRA techniques (i.e. time of flight, TOF and simultaneous non-contrast angiography and intraplaque hemorrhage, SNAP) were extracted using a custom-developed software named iCafe. Arterial spin labeling (ASL) and phase contrast (PC) cerebral blood flow (CBF) were measured as references. Linear regression was performed to study their associations with global cognition, measured with the Montreal Cognitive Assessment (MoCA). Intracranial artery length and number of branches on NCE-MRA, ASL CBF and PC CBF were found to be positively associated with MoCA scores (P < 0.01). The associations remained significant for artery length and number of branches on NCE-MRA after adjusting for clinical covariates and white matter hyperintensity volume. Further adjustment of confounding factors of ASL CBF or PC CBF did not abolish the significant association for artery length and number of branches on TOF. Our findings suggest that intracranial vascular features, including artery length and number of branches, on NCE-MRA may be useful markers of cerebrovascular health and provide added information over conventional brain blood flow measurements in individuals with cognitive impairment.

Quantitative MRI adds to neuropsychiatric lupus diagnostics

OBJECTIVE

Attributing neuropsychiatric manifestations to SLE is often challenging. Brain white matter lesions are frequent in SLE at MRI, but their diagnostic role is unclear. Here, we assessed whether white matter lesions count, volume and distribution measurement can help in the diagnosis of neuropsychiatric systemic lupus erythematosus (NPSLE).

METHODS

Brain dual-echo and 3D T1-weighted sequences were acquired from 32 patients with SLE and 32 healthy controls with a 3 T-scanner and employed to derive T2-hyperintense lesion volume (T2LV), number (T2LN) and probability maps (LPM) using a semi-automatic local thresholding segmentation technique. NPSLE was classified as per the ACR nomenclature, the Italian Society for Rheumatology algorithm and by clinical impression. Clinical descriptors including the SLE International Collaborating Clinics/ACR damage index (SDI) were also recorded.

RESULTS

Higher T2LV were observed in SLE vs healthy controls (P < 0.001) and in NPSLE vs other SLE (P =0.006). Patients with NPSLE also had higher T2LN (P =0.003) compared with other SLE. In SLE, T2LPM revealed a high prevalence of lesions in the splenium of the corpus callosum, right superior longitudinal fasciculus and right corona radiata. T2LV and T2LN correlated with SLE duration (rho = 0.606; P <0.001 and rho = 0.483; P =0.005, respectively) and age (rho = 0.478; P =0.006 and rho = 0.362; P = 0.042, respectively). T2LV also correlated with SDI (rho = 0.352; P =0.048). SLE patients with fatigue had lower T2LN (P =0.038) compared with patients without fatigue. Thresholds of T2LV ≥ 0.423 cm3 or of T2LN ≥ 12 were associated with definite NPSLE and improved the classification of patients with possible NPSLE per clinical impression.

CONCLUSION

Brain white matter lesions (WML) quantitation adds to NPSLE diagnostics.

Automatic brain lesion segmentation on standard magnetic resonance images: a scoping review

OBJECTIVES

Medical image analysis practices face challenges that can potentially be addressed with algorithm-based segmentation tools. In this study, we map the field of automatic MR brain lesion segmentation to understand the clinical applicability of prevalent methods and study designs, as well as challenges and limitations in the field.

DESIGN

Scoping review.

SETTING

Three databases (PubMed, IEEE Xplore and Scopus) were searched with tailored queries. Studies were included based on predefined criteria. Emerging themes during consecutive title, abstract, methods and whole-text screening were identified. The full-text analysis focused on materials, preprocessing, performance evaluation and comparison.

RESULTS

Out of 2990 unique articles identified through the search, 441 articles met the eligibility criteria, with an estimated growth rate of 10% per year. We present a general overview and trends in the field with regard to publication sources, segmentation principles used and types of lesions. Algorithms are predominantly evaluated by measuring the agreement of segmentation results with a trusted reference. Few articles describe measures of clinical validity.

CONCLUSIONS

The observed reporting practices leave room for improvement with a view to studying replication, method comparison and clinical applicability. To promote this improvement, we propose a list of recommendations for future studies in the field.

Fulminant Brain Atrophy and Vasculitis on Vessel-Wall Imaging in Neuropsychiatric Lupus: Case Report and Literature Review

Brain atrophy, typically slowly progressive, is a hallmark of neuropsychiatric (NP) systemic lupus erythematosus (SLE). In this article, we report a case of a young female patient with previously diagnosed SLE, without known NPSLE, with abrupt neurological deterioration and rapidly progressive diffuse atrophy in a few months. A comprehensive diagnostic work-up and follow-up magnetic resonance imaging (MRI), including high-resolution advanced vessel-wall sequences, revealed underlying cerebral vasculitis. The novelty factors that the present report brings are the rapid progressive atrophy demonstrated on follow-up MRI in a patient with SLE, and the depiction of an underlying vasculitis on specific vessel-wall MRI techniques. We also reviewed the literature and discussed the main current applications of vessel-wall MRI sequences. The aim of the report is to recognize this dramatic form of presentation of NPSLE and the utility of the new MRI techniques for the diagnosis.

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.

Distinct regional brain atrophy pattern in multiple sclerosis and neuropsychiatric systemic lupus erythematosus patients

Differentiation of systemic lupus erythematosus (SLE) from multiple sclerosis (MS) can be challenging, especially when neuropsychiatric (NP) symptoms are accompanied by white matter lesions in the brain. Given the lack of discriminative power of currently applied tools for their differentiation, there is an unmet need for other measures that can aid in distinguishing between the two autoimmune disorders. In this study we aimed at exploring whether brain atrophy measures could serve as markers differentiating MS and SLE. Thirty-seven relapsing-remitting MS and 38 SLE patients with nervous system manifestations, matched according to age and disease duration, underwent 1.5 Tesla magnetic resonance imaging (MRI), including volumetric sequences, and clinical assessment. Voxelwise analysis was performed using ANTS-SyN elastic registration protocol, FSL Randomise and Gamma methods. Cortical and subcortical segmentation was performed with Freesurfer 5.3 pipeline using T1-weighted MPRAGE sequence data. Using MRI volumetric markers of general and subcortical gray matter atrophy and clinical variables, we built a stepwise multivariable logistic diagnostic model to identify MRI parameters that best differentiate MS and SLE patients. We found that the best volumetric predictors to distinguish them were: fourth ventricle volume (sensitivity 0.86, specificity 0.57, area under the curve, AUC 0.77), posterior corpus callosum (sensitivity 0.81, specificity 0.57, AUC 0.68), and third ventricle to thalamus ratio (sensitivity 0.42, specificity 0.84, AUC 0.65). The same classifiers were identified in a subgroup analysis that included patients with a short disease duration. In MS brain atrophy and lesion load correlated with clinical disability, while in SLE age was the main determinant of brain volume. This study proposes new imaging parameters for differential diagnosis of MS and SLE with central nervous system involvement. We show there is a different pattern of atrophy in MS and SLE, and the key structural volumes that are differentially affected include fourth ventricle and posterior section of corpus callosum, followed by third ventricle to thalamus ratio. Different correlation patterns between volumetric and clinical data may suggest that while in MS atrophy is driven mainly by disease activity, in SLE it is mostly associated with age. However, these results need further replication in a larger cohort.

Lesion Segmentation in Automated 3D Breast Ultrasound: Volumetric Analysis

Mammography is the gold standard screening technique in breast cancer, but it has some limitations for women with dense breasts. In such cases, sonography is usually recommended as an additional imaging technique. A traditional sonogram produces a two-dimensional (2D) visualization of the breast and is highly operator dependent. Automated breast ultrasound (ABUS) has also been proposed to produce a full 3D scan of the breast automatically with reduced operator dependency, facilitating double reading and comparison with past exams. When using ABUS, lesion segmentation and tracking changes over time are challenging tasks, as the three-dimensional (3D) nature of the images makes the analysis difficult and tedious for radiologists. The goal of this work is to develop a semi-automatic framework for breast lesion segmentation in ABUS volumes which is based on the Watershed algorithm. The effect of different de-noising methods on segmentation is studied showing a significant impact ([Formula: see text]) on the performance using a dataset of 28 temporal pairs resulting in a total of 56 ABUS volumes. The volumetric analysis is also used to evaluate the performance of the developed framework. A mean Dice Similarity Coefficient of [Formula: see text] with a mean False Positive ratio [Formula: see text] has been obtained. The Pearson correlation coefficient between the segmented volumes and the corresponding ground truth volumes is [Formula: see text] ([Formula: see text]). Similar analysis, performed on 28 temporal (prior and current) pairs, resulted in a good correlation coefficient [Formula: see text] ([Formula: see text]) for prior and [Formula: see text] ([Formula: see text]) for current cases. The developed framework showed prospects to help radiologists to perform an assessment of ABUS lesion volumes, as well as to quantify volumetric changes during lesions diagnosis and follow-up.