Segmentation of white matter hyperintensities using convolutional neural networks with global spatial information in routine clinical brain MRI with none or mild vascular pathology

Background and purpose: White matter hyperintensities (WMH) are common in elderly individuals and contribute to age-related cognitive dysfunction. Converging evidence indicates that WMH affect white matter (WM) microstructural integrity in WMH and their penumbra. We aimed to investigate whether this effect extends to the distal WM tracts, and to examine the association between distal WM microstructural integrity and cognitive dysfunction in community-dwelling elderly people. Methods: Brain MRI data including FLAIR and DTI sequences of 174 participants (74 ± 5 years) of the Shanghai Aging Study (SAS) were collected and analyzed. For each participant, WMH lesions were segmented automatically. Eighteen major WM tracts were reconstructed using automated quantitative tractography, and the mean diffusivity (MD) of distal WM tracts (excluding an area of 12 mm around the WMH) was calculated. Multivariable linear regression was performed. Results: A high burden of tract-specific WMH was related to a high MD of distal WM tracts in the forceps major (FMA), anterior thalamic radiations (ATR), cingulum cingulate gyrus (CCG), corticospinal tract (CST), superior longitudinal fasciculus-parietal (SLFP), superior longitudinal fasciculus-temporal (SLFT), and uncinate fasciculus (UNC). Furthermore, a high MD of distal tracts was linked to worse attention and executive function in the forceps minor (FMI), right CCG, left inferior longitudinal fasciculus (ILF), SLFP, SLFT and UNC. Conclusions: The effect of WMH on the microstructural integrity of WM tracts may propagate along tracts to distal regions farther than the penumbra and eventually might affect attention and executive function.

Download Full-text

Hierarchical deep convolutional neural networks combine spectral and spatial information for highly accurate Raman-microscopy-based cytopathology

Journal of Biophotonics ◽

10.1002/jbio.201800022 ◽

2018 ◽

Vol 11 (10) ◽

pp. e201800022 ◽

Cited By ~ 10

Author(s):

Sascha D. Krauß ◽

Raphael Roy ◽

Hesham K. Yosef ◽

Tatjana Lechtonen ◽

Samir F. El-Mashtoly ◽

...

Keyword(s):

Neural Networks ◽

Convolutional Neural Networks ◽

Spatial Information ◽

Raman Microscopy ◽

Deep Convolutional Neural Networks

Download Full-text

Integration of Convolutional Neural Networks and Object-Based Post-Classification Refinement for Land Use and Land Cover Mapping with Optical and SAR Data

Remote Sensing ◽

10.3390/rs11060690 ◽

2019 ◽

Vol 11 (6) ◽

pp. 690 ◽

Cited By ~ 25

Author(s):

Shengjie Liu ◽

Zhixin Qi ◽

Xia Li ◽

Anthony Yeh

Keyword(s):

Neural Networks ◽

Land Use ◽

Land Cover ◽

Convolutional Neural Networks ◽

Spatial Resolution ◽

Spatial Information ◽

Object Based ◽

Sar Data ◽

The University ◽

Image Object

Object-based image analysis (OBIA) has been widely used for land use and land cover (LULC) mapping using optical and synthetic aperture radar (SAR) images because it can utilize spatial information, reduce the effect of salt and pepper, and delineate LULC boundaries. With recent advances in machine learning, convolutional neural networks (CNNs) have become state-of-the-art algorithms. However, CNNs cannot be easily integrated with OBIA because the processing unit of CNNs is a rectangular image, whereas that of OBIA is an irregular image object. To obtain object-based thematic maps, this study developed a new method that integrates object-based post-classification refinement (OBPR) and CNNs for LULC mapping using Sentinel optical and SAR data. After producing the classification map by CNN, each image object was labeled with the most frequent land cover category of its pixels. The proposed method was tested on the optical-SAR Sentinel Guangzhou dataset with 10 m spatial resolution, the optical-SAR Zhuhai-Macau local climate zones (LCZ) dataset with 100 m spatial resolution, and a hyperspectral benchmark the University of Pavia with 1.3 m spatial resolution. It outperformed OBIA support vector machine (SVM) and random forest (RF). SVM and RF could benefit more from the combined use of optical and SAR data compared with CNN, whereas spatial information learned by CNN was very effective for classification. With the ability to extract spatial features and maintain object boundaries, the proposed method considerably improved the classification accuracy of urban ground targets. It achieved overall accuracy (OA) of 95.33% for the Sentinel Guangzhou dataset, OA of 77.64% for the Zhuhai-Macau LCZ dataset, and OA of 95.70% for the University of Pavia dataset with only 10 labeled samples per class.

Download Full-text

GLIA-Deep: Glioblastoma Image Analysis Using Deep Learning Convolutional Neural Networks to Accurately Classify Gene Methylation and Predict Drug Effectiveness

10.20944/preprints202012.0386.v1 ◽

2020 ◽

Author(s):

Viraj Mehta

Keyword(s):

Neural Networks ◽

Dna Repair ◽

Precision Medicine ◽

Survival Time ◽

Convolutional Neural Networks ◽

Alkylating Agent ◽

Brain Mri ◽

Methylation Status ◽

Chemotherapy Drugs ◽

Mri Scans

Glioblastoma multiforme is a deadly brain cancer with a median patient survival time of 18-24 months, despite aggressive treatments. This limited success is due to a combination of aggressive tumor behavior, genetic heterogeneity of the disease within a single patient’s tumor, resistance to therapy, and lack of precision medicine treatments. A single specimen using a biopsy cannot be used for complete assessment of the tumor’s microenvironment, making personalized care limited and challenging. Temozolomide (TMZ) is a commercially approved alkylating agent used to treat glioblastoma, but around 50% of temozolomide-treated patients do not respond to it due to the over-expression of O6-methylguanine methyltransferase (MGMT). MGMT is a DNA repair enzyme that rescues tumor cells from alkylating agent-induced damage, leading to resistance to chemotherapy drugs. Epigenetic silencing of the MGMT gene by promoter methylation results in decreased MGMT protein expression, reduced DNA repair activity, increased sensitivity to TMZ, and longer survival time. Thus, it is paramount that clinicians determine the methylation status of patients to provide personalized chemotherapy drugs. However, current methods for determining this via invasive biopsies or manually curated features from brain MRI (Magnetic Resonance Imaging) scans are time- and cost- intensive, and have a very low accuracy. Authors present a novel approach of using convolutional neural networks to predict methylation status and recommend patient-specific treatments via an analysis of brain MRI scans. The authors have developed an AI platform, GLIA-Deep, using a U-Net architecture and a ResNet-50 architecture trained on genomic data from TCGA (The Cancer Genome Atlas through the National Cancer Institute) and brain MRI scans from TCIA (The Cancer Imaging Archive). GLIA-Deep performs tumor region identification and determines MGMT methylation status with 90% accuracy in less than 5 seconds, a real-time analysis that eliminates huge time and cost investments of invasive biopsies. Using computational modeling, the analysis further recommends microRNAs that modulate MGMT gene expression by translational repression to make glioma cells TMZ sensitive, thereby improving the survival of glioblastoma patients with unmethylated MGMT. GLIA-Deep is a completely integrated, end-to-end, cost-effective and time-efficient platform that advances precision medicine by recommending personalized therapies from an analysis of individual MRI scans to improving glioblastoma treatment options.

Download Full-text

Slow-wave Sleep and MRI Markers of Brain Aging in a Community-based Sample

Neurology ◽

10.1212/wnl.0000000000011377 ◽

2020 ◽

pp. 10.1212/WNL.0000000000011377

Author(s):

Andree-Ann Baril ◽

Alexa S Beiser ◽

Vincent Mysliwiec ◽

Erlan Sanchez ◽

Charles S DeCarli ◽

...

Keyword(s):

White Matter ◽

Slow Wave ◽

White Matter Hyperintensities ◽

Slow Wave Sleep ◽

Brain Volume ◽

Brain Aging ◽

Brain Mri ◽

Community Based ◽

Brain Infarcts ◽

Absolute Duration

Objective:To test the hypothesis that reduced slow-wave sleep, or N3 sleep, which is thought to underlie the restorative functions of sleep, is associated with MRI markers of brain aging, we evaluated this relationship in the community-based Framingham Heart Study Offspring cohort using polysomnography and brain MRI.Methods:We studied 492 participants (58.8 ± 8.8 years, 49.4% male) free of neurological diseases who completed a brain MRI scan and in-home overnight polysomnography to assess slow-wave sleep (absolute duration and percentage of total sleep). Volumes of total brain, total cortical, frontal cortical, subcortical gray matter, hippocampus, and white matter hyperintensities were investigated as a percentage of intracranial volume and the presence of covert brain infarcts was evaluated. Linear and logistic regression models were adjusted for age, age squared, sex, time interval between polysomnography and MRI (3.3 ± 1.0 years), APOE4 carrier status, stroke risk factors, sleeping pill use, body mass index and depression.Results:Less slow-wave sleep was associated with lower cortical brain volume (absolute duration, β[standard error]: 0.20[0.08], p=0.015; percentage, 0.16[0.08], p=0.044), lower subcortical brain volume (percentage, 0.03[0.02], p=0.034), and higher white matter hyperintensities volume (absolute duration, -0.12[0.05], p=0.010; percentage -0.10[0.04], p=0.033). Slow-wave sleep duration was not associated with hippocampal volume or the presence of covert brain infarcts.Conclusion:Loss of slow-wave sleep might facilitate accelerated brain aging, as evidence by its association with MRI markers suggestive of brain atrophy and injury. Alternatively, subtle injuries and accelerated aging might reduce the ability of the brain to produce slow-wave sleep.

Download Full-text