Detection Of Microbleeds Research Articles

Toward Automated Detection of Microbleeds with Anatomical Scale Localization Using Deep Learning

Cerebral Microbleeds (CMBs) are chronic deposits of small blood products in the brain tissues, which have explicit relation to various cerebrovascular diseases depending on their anatomical location, including cognitive decline, intracerebral hemorrhage, and cerebral infarction. However, manual detection of CMBs is a time consuming and error-prone process because of their sparse and tiny structural properties. The detection of CMBs is commonly affected by the presence of many CMB mimics that cause a high false-positive rate (FPR), such as calcifications and pial vessels. This paper proposes a novel 3D deep learning framework that not only detects CMBs but also identifies their anatomical location in the brain (i.e., lobar, deep, and infratentorial regions). For the CMBs detection task, we propose a single end-to-end model by leveraging the 3D U-Net as a backbone with Region Proposal Network (RPN). To significantly reduce the false positives within the same single model, we develop a new scheme, containing Feature Fusion Module (FFM) that detects small candidates utilizing contextual information and Hard Sample Prototype Learning (HSPL) that mines CMB mimics and generates additional loss term called concentration loss using Convolutional Prototype Learning (CPL). For the anatomical localization task, we exploit the 3D U-Net segmentation network to segment anatomical structures of the brain. This task not only identifies to which region the CMBs belong but also eliminates some false positives from the detection task by leveraging anatomical information. We utilize Susceptibility-Weighted Imaging (SWI) and phase images as 3D input to efficiently capture 3D information. The results show that the proposed RPN that utilizes the FFM and HSPL outperforms the baseline RPN and achieves a sensitivity of 94.66% vs. 93.33% and an average number of false positives per subject (FPavg) of 0.86 vs. 14.73. Furthermore, the anatomical localization task enhances the detection performance by reducing the FPavg to 0.56 while maintaining the sensitivity of 94.66%.

Comparison study of quantitative susceptibility mapping with GRAPPA and wave-CAIPI: reproducibility, consistency, and microbleeds detection.

We compared quantitative susceptibility mapping (QSM) with wave-CAIPI 9 × (QSM_WC9 ×) with reference standard QSM with GRAPPA 2 × (QSM_G2 ×) in two MR scanners. We also compared detectability of microbleeds in both QSMs to demonstrate clinical feasibility of both QSMs. This prospective study was approved by the institutional review board and written informed consent was obtained from each subject. Healthy subjects were recruited to evaluate intra-scanner reproducibility, inter-scanner consistency, and inter-sequence consistency of QSM_G2 × and QSM_WC9 × at 2 MR scanners. Susceptibility values measured with volume of interests (VOIs) were evaluated. Patients who were requested for susceptibility weighted imaging were also recruited in this study to measure microbleeds on QSM_G2 × and QSM_WC9 × . The number of microbleeds was compared between two QSMs. Total 55 healthy subjects (male 34, female 21, 38.3years [23-79]) were included in this study. We investigated reproducibility and consistency of QSM_WC9 × by comparing reference standard QSM_G2 × in two MR scanners in this study, and high correlation (ρ, 0.93-0.97) and high intraclass correlation coefficient (ICC) (0.97-0.99) were obtained. Sixty patients (male 30, female 30; age, 55.4years [21-85]) were finally enrolled in this prospective study. The ICC of the detected number of microbleeds between QSM_G2 × and QSM_WC9 × was 0.99 (0.98-0.99). QSM_WC9 × and reference standard QSM_G2 × in two MR scanners showed good reproducibility and consistency in estimating magnetic susceptibilities. QSM_WC9 × and QSM_G2 × were also comparable in terms of microbleeds detection with good agreement of raters and high ICC.

Quantifying cerebral microbleeds using quantitative susceptibility mapping from magnetization-prepared rapid gradient-echo.

T1-weighted magnetization-prepared rapid gradient-echo (MPRAGE) is commonly included in brain studies for structural imaging using magnitude images; however, its phase images can provide an opportunity to assess microbleed burden using quantitative susceptibility mapping (QSM). This potential application for MPRAGE-based QSM was evaluated using in vivo and simulated measurements. Possible factors affecting image quality were also explored. Detection sensitivity was evaluated against standard multiecho gradient echo (MEGE) QSM using 3-T in vivo data of 15 subjects with a combined total of 108 confirmed microbleeds. The two methods were compared based on the microbleed size and susceptibility measurements. In addition, simulations explored the detection sensitivity of MPRAGE-QSM at different representative magnetic field strengths and echo times using microbleeds of different size, susceptibility, and location. Results showed that in vivo microbleeds appeared to be smaller (× 0.54) and of higher mean susceptibility (× 1.9) on MPRAGE-QSM than on MEGE-QSM, but total susceptibility estimates were in closer agreement (slope: 0.97, r2: 0.94), and detection sensitivity was comparable. In simulations, QSM at 1.5 T had a low contrast-to-noise ratio that obscured the detection of many microbleeds. Signal-to-noise ratio (SNR) levels at 3 T and above resulted in better contrast and increased detection. The detection rates for microbleeds of minimum one-voxel diameter and 0.4-ppm susceptibility were 0.55, 0.80, and 0.88 at SNR levels of 1.5, 3, and 7 T, respectively. Size and total susceptibility estimates were more consistent than mean susceptibility estimates, which showed size-dependent underestimation. MPRAGE-QSM provides an opportunity to detect and quantify the size and susceptibility of microbleeds of at least one-voxel diameter at B0of 3 T or higher with no additional time cost, when standard T2*-weighted images are not available or have inadequate spatial resolution. The total susceptibility measure is more robust against sequence variations and might allow combining data from different protocols.

Retraction: Cerebral microbleed detection via convolutional neural network and extreme learning machine.

[This retracts the article DOI: 10.3389/fncom.2021.738885.].

Deep-Learning-Based MRI Microbleeds Detection for Cerebral Small Vessel Disease on Quantitative Susceptibility Mapping.

Cerebral microbleeds (CMB) are indicators of severe cerebral small vessel disease (CSVD) that can be identified through hemosiderin-sensitive sequences in MRI. Specifically, quantitative susceptibility mapping (QSM) and deep learning were applied to detect CMBs in MRI. To automatically detect CMB on QSM, we proposed a two-stage deep learning pipeline. Retrospective. A total number of 1843 CMBs from 393 patients (69 ± 12) with cerebral small vessel disease were included in this study. Seventy-eight subjects (70 ± 13) were used as external testing. 3 T/QSM. The proposed pipeline consisted of two stages. In stage I, 2.5D fast radial symmetry transform (FRST) algorithm along with a one-layer convolutional network was used to identify CMB candidate regions in QSM images. In stage II, the V-Net was utilized to reduce false positives. The V-Net was trained using CMB and non CMB labels, which allowed for high-level feature extraction and differentiation between CMBs and CMB mimics like vessels. The location of CMB was assessed according to the microbleeds anatomical rating scale (MARS) system. The sensitivity and positive predicative value (PPV) were reported to evaluate the performance of the model. The number of false positive per subject was presented. Our pipeline demonstrated high sensitivities of up to 94.9% at stage I and 93.5% at stage II. The overall sensitivity was 88.9%, and the false positive rate per subject was 2.87. With respect to MARS, sensitivities of above 85% were observed for nine different brain regions. We have presented a deep learning pipeline for detecting CMB in the CSVD cohort, along with a semi-automated MARS scoring system using the proposed method. Our results demonstrated the successful application of deep learning for CMB detection on QSM and outperformed previous handcrafted methods. 2 TECHNICAL EFFICACY: Stage 2.

Ex‐vivo microbleed detection in community‐based older adults using confidence‐aware learning

AbstractBackgroundCerebral microbleed (CMB) detection on MRI scans of autopsied brains of community‐based older adults is crucial for MR‐pathology studies of cerebral small vessel disease (SVD). However, training an ex‐vivo CMB detection model is difficult due to the low prevalence of CMBs in the brains of community‐based older adults, and the fact that CMB mimics greatly outnumber CMBs on ex‐vivo MRI. We demonstrate that confidence‐aware deep learning, a technique for enforcing correct ranking of examples during neural network training, significantly improves ex‐vivo CMB detection performance and enhances the interpretability of model predictions.MethodCMBs from 286 participants from two longitudinal cohort studies of aging were included in this work. T2*‐weighted gradient echo scans of autopsied brains with a voxel size of 1×1×1 mm were used. Confidence‐aware learning was used to regularize deep learning model predictions such that they reflect the degree of confidence the detection model has in a prediction. Model confidence was estimated during training by correctness, the number of correct prediction events across epochs divided by the total number of epochs. The objective loss function was modified to include a correctness ranking loss (CRL) term that encourages correspondence between correctness and predicted probabilities. Models are trained with and without the CRL loss term to evaluate the impact of confidence‐aware learning on model performance. An end‐to‐end CMB detection framework that combines data synthesis, candidate selection, false positive reduction, and full scan evaluation was used as the backbone for this work.ResultCMB detection models trained using correctness ranking loss achieved a significantly higher ensemble average precision (0.2915) compared with models trained without correctness ranking loss (0.2379). The CRL trained model achieved a 25% higher sensitivity at 0.5 false positives per subject and 12% higher sensitivity at 3 false positives per subject when compared to models trained without CRL.ConclusionThis work demonstrates that training with confidence‐aware learning can improve the performance and interpretability of ex‐vivo CMB detection algorithms in community‐based cohorts. The use of a confidence‐aware CMB detection algorithm for assisted CMB annotation in ex‐vivo MRI will aid in future MR‐pathology studies into the relationships between CMBs and neuropathology.

The Current State of Susceptibility-Weighted Imaging and Quantitative Susceptibility Mapping in Head Trauma.

Susceptibility-weighted imaging (SWI) is a MR imaging technique suited to detect structural and microstructural abnormalities in traumatic brain injury (TBI). This review article provide an insight in to the physics principles of SWI and its clinical application in unraveling the complex interaction of the biophysical mechanisms of head injury. Literature evidences support SWI as the most ideal sequence in detection of microbleeds, which is the "tip of the iceberg" biomarker of microvascular injuries. The review also detailed the emerging advance techniques of Quantitative susceptibility mapping (QSM) and artificial intelligence offer the ability to detect and follow the evolution of microbleeds in patient with chronic TBI. These new techniques offers a unique insight into the acute and chronic state of TBI.

Implications of quantitative susceptibility mapping at 7 Tesla MRI for microbleeds detection in cerebral small vessel disease.

Cerebral microbleeds (MBs) are a hallmark of cerebral small vessel disease (CSVD) and can be found on T2*-weighted sequences on MRI. Quantitative susceptibility mapping (QSM) is a postprocessing method that also enables MBs identification and furthermore allows to differentiate them from calcifications. We explored the implications of using QSM at submillimeter resolution for MBs detection in CSVD. Both 3 and 7 Tesla (T) MRI were performed in elderly participants without MBs and patients with CSVD. MBs were quantified on T2*-weighted imaging and QSM. Differences in the number of MBs were assessed, and subjects were classified in CSVD subgroups or controls both on 3T T2*-weighted imaging and 7T QSM. 48 participants [mean age (SD) 70.9 (8.8) years, 48% females] were included: 31 were healthy controls, 6 probable cerebral amyloid angiopathy (CAA), 9 mixed CSVD, and 2 were hypertensive arteriopathy [HA] patients. After accounting for the higher number of MBs detected at 7T QSM (Median = Mdn; Mdn7T-QSM = 2.5; Mdn3T-T2 = 0; z = 4.90; p < 0.001) and false positive MBs (6.1% calcifications), most healthy controls (80.6%) demonstrated at least one MB and more MBs were discovered in the CSVD group. Our observations suggest that QSM at submillimeter resolution improves the detection of MBs in the elderly human brain. A higher prevalence of MBs than so far known in healthy elderly was revealed.

Prospective Assessment of Cerebral Microbleeds with Low-Field Magnetic Resonance Imaging (0.55 Tesla MRI).

Accurate detection of cerebral microbleeds (CMBs) on susceptibility-weighted (SWI) magnetic resonance imaging (MRI) is crucial for the characterization of many neurological diseases. Low-field MRI offers greater access at lower costs and lower infrastructural requirements, but also reduced susceptibility artifacts. We therefore evaluated the diagnostic performance for the detection of CMBs of a whole-body low-field MRI in a prospective cohort of suspected stroke patients compared to an established 1.5 T MRI. A prospective scanner comparison was performed including 27 patients, of whom 3 patients were excluded because the time interval was >1 h between acquisition of the 1.5 T and 0.55 T MRI. All SWI sequences were assessed for the presence, number, and localization of CMBs by two neuroradiologists and additionally underwent a Likert rating with respect to image impression, resolution, noise, contrast, and diagnostic quality. A total of 24 patients with a mean age of 74 years were included (11 female). Both readers detected the same number and localization of microbleeds in all 24 datasets (sensitivity and specificity 100%; interreader reliability ϰ = 1), with CMBs only being observed in 12 patients. Likert ratings of the sequences at both field strengths regarding overall image quality and diagnostic quality did not reveal significant differences between the 0.55 T and 1.5 T sequences (p = 0.942; p = 0.672). For resolution and contrast, the 0.55 T sequences were even significantly superior (p < 0.0001; p < 0.0003), whereas the 1.5 T sequences were significantly superior (p < 0.0001) regarding noise. Low-field MRI at 0.55 T may have similar accuracy as 1.5 T scanners for the detection of microbleeds and thus may have great potential as a resource-efficient alternative in the near future.

Quantitative susceptibility mapping improves cerebral microbleed detection relative to susceptibility-weighted images.

Cerebral microbleed (CMB) detection impacts disease diagnosis and management. Susceptibility-weighted imaging (SWI) MRI depictions of CMBs are used with phase images (SWIP) to distinguish blood from calcification, via qualitative intensity evaluation (bright/dark). However, the intensities depicted for a single lesion can vary within and across consecutive SWIP image planes, impairing the classification of findings as a CMB. We hypothesize that quantitative susceptibility mapping (QSM) MRI, which maps tissue susceptibility, demonstrates less in- and through-plane intensity variation, improving the clinician's ability to categorize a finding as a CMB. Forty-eight patients with acute intracranial hemorrhage who received multi-echo gradient echo MRI used to generate both SWI/SWIP and morphology-enabled dipole inversion QSM images were enrolled. Five hundred and sixty lesions were visually classified as having homogeneous or heterogeneous in-plane and through-plane intensity by a neuroradiologist and two diagnostic radiology residents using published rating criteria. When available, brain CT scans were analyzed for calcification or acute hemorrhage. Relative risk (RR) ratios and confidence intervals (CIs) were calculated using a generalized linear model with log link and binary error. QSM showed unambiguous lesion signal intensity three times more frequently than SWIP (RR = 0.3235, 95% CI 0.2386-0.4386, p<.0001). The probability of QSM depicting homogeneous lesion intensity was three times greater than SWIP for small (RR = 0.3172, 95% CI 0.2382-0.4225, p<.0001), large (RR = 0.3431, 95% CI 0.2045-0.5758, p<.0001), lobar (RR = 0.3215, 95% CI 0.2151-0.4805, p<.0001), cerebellar (RR = 0.3215, 95% CI 0.2151-0.4805, p<.0001), brainstem (RR = 0.3100, 95% CI 0.1192-0.8061, p = .0163), and basal ganglia (RR = 0.3380, 95% CI 0.1980-0.5769, p<.0001) lesions. QSM more consistently demonstrates interpretable lesion intensity compared to SWIP as used for distinguishing CMBs from calcification.

Using transfer learning for automated microbleed segmentation.

Cerebral microbleeds are small perivascular hemorrhages that can occur in both gray and white matter brain regions. Microbleeds are a marker of cerebrovascular pathology and are associated with an increased risk of cognitive decline and dementia. Microbleeds can be identified and manually segmented by expert radiologists and neurologists, usually from susceptibility-contrast MRI. The latter is hard to harmonize across scanners, while manual segmentation is laborious, time-consuming, and subject to interrater and intrarater variability. Automated techniques so far have shown high accuracy at a neighborhood ("patch") level at the expense of a high number of false positive voxel-wise lesions. We aimed to develop an automated, more precise microbleed segmentation tool that can use standardizable MRI contrasts. We first trained a ResNet50 network on another MRI segmentation task (cerebrospinal fluid vs. background segmentation) using T1-weighted, T2-weighted, and T2* MRIs. We then used transfer learning to train the network for the detection of microbleeds with the same contrasts. As a final step, we employed a combination of morphological operators and rules at the local lesion level to remove false positives. Manual segmentation of microbleeds from 78 participants was used to train and validate the system. We assessed the impact of patch size, freezing weights of the initial layers, mini-batch size, learning rate, and data augmentation on the performance of the Microbleed ResNet50 network. The proposed method achieved high performance, with a patch-level sensitivity, specificity, and accuracy of 99.57, 99.16, and 99.93%, respectively. At a per lesion level, sensitivity, precision, and Dice similarity index values were 89.1, 20.1, and 0.28% for cortical GM; 100, 100, and 1.0% for deep GM; and 91.1, 44.3, and 0.58% for WM, respectively. The proposed microbleed segmentation method is more suitable for the automated detection of microbleeds with high sensitivity.

Effective feature extraction for Cerebral Microbleed detection using Edge Emphasized Weber Maximum Directional Co-occurance Matrix

Effective feature extraction for Cerebral Microbleed detection using Edge Emphasized Weber Maximum Directional Co-occurance Matrix

Automated cerebral microbleed detection using selective 3D gradient co-occurance matrix and convolutional neural network

Cerebral Microbleeds (CMBs) are small blood clots that form in the brains of many elderly and stroke victims. This promotes dementia as well as other physiological challenges in their daily activities. Early detection of CMBs is critical for limiting the long-term impacts that lead to confusing behavior. Because CMBs in the brain are so minute, manual identification by neurologists might be difficult at times, leading to misinterpretations. In this paper, we propound a newfangled feature extraction technique for the detection of CMB, by Selective 3D Gradient Co-occurance Matrix (S3DGCM). Our proposed framework efficiently detects the cerebral microbleeds (CMB) from magnetic resonance images by incorporating the Scharr Gradient operator with the concepts of the Co-occurance Matrix. The selected ROI was classified as CMB or non-CMB using a Convolutional Neural Network (CNN) classifier with Long Short Term Memory (LSTM) based on the extracted features from the S3DGCM. When compared with the previous methods, our method can extract more useful features which increase the accuracy rate in the detection process. The proposed work has been verified on two datasets containing MRI images with CMBs and Non-CMBs. The experimental results specify that our proposed scheme yields a promising result with the best sensitivity of 98.76%, specificity of 97.21%, an accuracy of 98.24%, and an F1-score of 98.76%.

A data-driven deep learning pipeline for quantitative susceptibility mapping (QSM)

PurposeThis study developed a data-driven optimization to improve the accuracy of deep learning QSM quantification. MethodsThe proposed deep learning QSM pipeline consisted of two projections onto convex set (POCS) models designed to decouple trainable network components with the spherical mean value (SMV) filters and dipole kernel in the data-driven optimization. They were a background field removal network (named POCSnet1) and a dipole inversion network (named POCSnet2). Both POCSnet1 and POCSnet2 were the unrolled V-Net with iterative data-driven optimization to enforce the data fidelity. For training POCSnet1, we simulated phantom data with random geometric shapes as the background susceptibility sources. For training POCSnet2, we used geometric shapes to mimic the QSM. The evaluation was performed on synthetic data, a public COSMOS (N = 1), and clinical data from a Parkinson's disease cohort (N = 71) and small-vessel disease cohort (N = 26). For comparison, DLL2, FINE, and autoQSM, were implemented and tested under the same experimental setting. ResultsOn COSMOS, results from POCSnet1 were more similar to that of the V-SHARP method with NRMSE = 23.7% and SSIM = 0.995, compared with the NRMSE = 62.7% and SSIM = 0.975 for SHARQnet, a naïve V-Net model. On COSMOS, the NRMSE and HFEN for POCSnet2 were 58.1% and 56.7%; while for DLL2, FINE, and autoQSM, they were 62.0% and 61.2%, 69.8% and 67.5%, and 87.5% and 85.3%, respectively. On the Parkinson's disease cohort, our results were consistent with those obtained from VSHARP+STAR-QSM with biases <3% and outperformed the SHARQnet+DeepQSM that had biases of 7% to 10%. The sensitivity of cerebral microbleed detection using our pipeline was 100%, compared with 92% by SHARQnet+DeepQSM. ConclusionData-driven optimization improved the accuracy of QSM quantification compared with that of naïve V-Net models.

Return to Work Within Four Months of Grade 3 Diffuse Axonal Injury.

Neuroprognostication following diffuse axonal injury (DAI) has historically relied on neuroimaging techniques with lower spatial resolution and contrast than techniques currently available in clinical practice. Since the initial studies of DAI classification and prognosis in the 1980s and 1990s, advances in neuroimaging have improved detection of brainstem microbleeds, a hallmark feature of Grade 3 DAI that has traditionally been associated with poor neurologic outcome. Here, we report clinical and radiologic data from two patients with severe traumatic brain injury and grade 3 DAI who recovered functional independence and returned to work within 4months of injury. Importantly, both patients were scanned using 3Tesla MRI protocols that included susceptibility-weighted imaging (SWI), a technique that provides enhanced sensitivity for detecting brainstem microbleeds. These observations highlight the importance of developing approaches to DAI classification and prognosis that better align with contemporary neuroimaging capabilities.

A Transfer Learning-Based Approach to Detect Cerebral Microbleeds

Cerebral microbleeds are small chronic vascular diseases that occur because of irregularities in the cerebrum vessels. Individuals and elderly people with brain injury and dementia can have small microbleeds in their brains. A recent study has shown that cerebral microbleeds could be remarkably risky in terms of life and can be riskier for patients with dementia. In this study, we proposed an efficient approach to automatically identify microbleeds by reducing the false positives in openly available susceptibility-weighted imaging (SWI) data samples. The proposed structure comprises two different pre-trained convolutional models with four stages. These stages include (i) skull removal and augmentation, (ii) making clusters of data samples using the k-mean classifier, (iii) reduction of false positives for efficient performance, and (iv) transfer-learning classification. The proposed technique was assessed using the SWI dataset available for 20 subjects. For our findings, we attained an accuracy of 97.26% with a 1.8% false-positive rate using data augmentation on the AlexNet transfer learning model and a 1.1% false-positive rate with 97.89% accuracy for the ResNet 50 model with data augmentation approaches. The results show that our models outperformed the existing approach for the detection of microbleeds.

Early Cerebrovascular Silent Changes in Long-Standing End-Stage Renal Disease Patients on Hemodialysis Value of Adding Advanced Unenhanced MRI Sequences to Imaging Protocols

Background: End-Stage Renal Disease (ESRD) patients on hemodialysis suffered from many central and peripheral neurological insults. Aim of the work:This study aimed to assess the value of using MRA, MRV, and SWAN sequences in early visualization of the silent cerebrovascular complications in those patients. Patients and Methods: Our study was conducted on forty-five patients with well-documented ESRD on regular hemodialysis for more than 5 years with no neurological manifestation, all undergone unenhanced MRI, DWI with ADC and measuring the ADC value, SWAN, MRA, and MRV. Results: we found that 11% of cases have acute infarction at the basal ganglia region. 36.7% of patients are diagnosed with intracerebral hemorrhage, only 26.7% of them are seen by the conventional MRI, and 16.7% of cases showed microbleeds on SWAN with normal conventional MRI. The sensitivity and specificity of SWAN with ADC value to detect brain hemorrhage are 100% and 88.8% respectively. Conclusion: Visualization of cerebrovascular complications such as infarction, hemorrhage, atherosclerosis, and arterial and venous occlusion using DWI with ADC, MRA, MRV, and SWAN and is very useful in early management and better prognosis of ESRD patients even with silent complications that don’t give up symptoms with high sensitivity and specificity of SWAN in early detection of hemorrhage and microbleeds.

A Comprehensive Analysis of Deep Neural-Based Cerebral Microbleeds Detection System

Machine learning-based systems are gaining interest in the field of medicine, mostly in medical imaging and diagnosis. In this paper, we address the problem of automatic cerebral microbleeds (CMB) detection in magnetic resonance images. It is challenging due to difficulty in distinguishing a true CMB from its mimics, however, if successfully solved, it would streamline the radiologists work. To deal with this complex three-dimensional problem, we propose a machine learning approach based on a 2D Faster RCNN network. We aimed to achieve a reliable system, i.e., with balanced sensitivity and precision. Therefore, we have researched and analysed, among others, impact of the way the training data are provided to the system, their pre-processing, the choice of model and its structure, and also the ways of regularisation. Furthermore, we also carefully analysed the network predictions and proposed an algorithm for its post-processing. The proposed approach enabled for obtaining high precision (89.74%), sensitivity (92.62%), and F1 score (90.84%). The paper presents the main challenges connected with automatic cerebral microbleeds detection, its deep analysis and developed system. The conducted research may significantly contribute to automatic medical diagnosis.

Detecting cerebral microbleeds via deep learning with features enhancement by reusing ground truth

Background and objectivesCerebral microbleeds (CMBs) are cerebral small vascular diseases and are often used to diagnose symptoms such as stroke and dementia. Manual detection of cerebral microbleeds is a time-consuming and error-prone task, so the application of microbleed detection algorithms based on deep learning is of great significance. This study presents the feature enhancement technology applying to improve the performances of detecting CMBs. The primary purpose of the feature enhancement is emphasizing the meaningful features, leading deep learning network easier and correctly to optimize. MethodIn this study, we applied feature enhancement in detecting CMBs from brain MRI images. Feature enhancement enhanced specific intervals and suppressed the useless intervals of the feature map. This method was applied in SSD-512 and SSD-300 algorithm, using VGG architecture pre-trained in the ImageNet dataset. ResultsThe proposed method was applied in SSD-512. Moreover, the model was trained and tested on the sequence of SWAN images of brain MRI images. The results of the experiment demonstrate that our method effectively improves the detection performance of the SSD network in detecting CMBs. We train SSD-512 120000 iterations and test results on the test datasets, by applying the feature enhancement layer, improving the precision with 3.3% and the mAP of 2.3%. In the same way, we trained SSD-300, improving the mAP of 2.0%. 2.8% and 7.4% precision are improved by applying feature enhancement layer In ResNet-34 and MobileNet. ConclusionsThe proposed method achieved more effective performance, demonstrated that feature enhancement can be a helpful algorithm to enhance the deep learning model.

Automated detection of cerebral microbleeds on T2*-weighted MRI

Cerebral microbleeds, observed as small, spherical hypointense regions on gradient echo (GRE) or susceptibility weighted (SWI) magnetic resonance imaging (MRI) sequences, reflect small hemorrhagic infarcts, and are associated with conditions such as vascular dementia, small vessel disease, cerebral amyloid angiopathy, and Alzheimer’s disease. The current gold standard for detecting and rating cerebral microbleeds in a research context is visual inspection by trained raters, a process that is both time consuming and subject to poor reliability. We present here a novel method to automate microbleed detection on GRE and SWI images. We demonstrate in a community-based cohort of older adults that the method is highly sensitive (greater than 92% of all microbleeds accurately detected) across both modalities, with reasonable precision (fewer than 20 and 10 false positives per scan on GRE and SWI, respectively). We also demonstrate that the algorithm can be used to identify microbleeds over longitudinal scans with a higher level of sensitivity than visual ratings (50% of longitudinal microbleeds correctly labeled by the algorithm, while manual ratings was 30% or lower). Further, the algorithm identifies the anatomical localization of microbleeds based on brain atlases, and greatly reduces time spent completing visual ratings (43% reduction in visual rating time). Our automatic microbleed detection instrument is ideal for implementation in large-scale studies that include cross-sectional and longitudinal scanning, as well as being capable of performing well across multiple commonly used MRI modalities.