Deep Learning Detection of Aneurysm Clips for Magnetic Resonance Imaging Safety.

Flagging the presence of metal devices before a head MRI scan is essential to allow appropriate safety checks. There is an unmet need for an automated system which can flag aneurysm clips prior to MRI appointments. We assess the accuracy with which a machine learning model can classify the presence or absence of an aneurysm clip on CT images. A total of 280 CT head scans were collected, 140 with aneurysm clips visible and 140 without. The data were used to retrain a pre-trained image classification neural network to classify CT localizer images. Models were developed using fivefold cross-validation and then tested on a holdout test set. A mean sensitivity of 100% and a mean accuracy of 82% were achieved. Predictions were explained using SHapley Additive exPlanations (SHAP), which highlighted that appropriate regions of interest were informing the models. Models were also trained from scratch to classify three-dimensional CT head scans. These did not exceed the sensitivity of the localizer models. This work illustrates an application of computer vision image classification to enhance current processes and improve patient safety.

Robust EEG Based Biomarkers to Detect Alzheimer's Disease.

Biomarkers to detect Alzheimer's disease (AD) would enable patients to gain access to appropriate services and may facilitate the development of new therapies. Given the large numbers of people affected by AD, there is a need for a low-cost, easy to use method to detect AD patients. Potentially, the electroencephalogram (EEG) can play a valuable role in this, but at present no single EEG biomarker is robust enough for use in practice. This study aims to provide a methodological framework for the development of robust EEG biomarkers to detect AD with a clinically acceptable performance by exploiting the combined strengths of key biomarkers. A large number of existing and novel EEG biomarkers associated with slowing of EEG, reduction in EEG complexity and decrease in EEG connectivity were investigated. Support vector machine and linear discriminate analysis methods were used to find the best combination of the EEG biomarkers to detect AD with significant performance. A total of 325,567 EEG biomarkers were investigated, and a panel of six biomarkers was identified and used to create a diagnostic model with high performance (≥85% for sensitivity and 100% for specificity).

Early Detection of Alzheimer's Disease with Blood Plasma Proteins Using Support Vector Machines.

The successful development of amyloid-based biomarkers and tests for Alzheimer's disease (AD) represents an important milestone in AD diagnosis. However, two major limitations remain. Amyloid-based diagnostic biomarkers and tests provide limited information about the disease process and they are unable to identify individuals with the disease before significant amyloid-beta accumulation in the brain develops. The objective in this study is to develop a method to identify potential blood-based non-amyloid biomarkers for early AD detection. The use of blood is attractive because it is accessible and relatively inexpensive. Our method is mainly based on machine learning (ML) techniques (support vector machines in particular) because of their ability to create multivariable models by learning patterns from complex data. Using novel feature selection and evaluation modalities, we identified 5 novel panels of non-amyloid proteins with the potential to serve as biomarkers of early AD. In particular, we found that the combination of A2M, ApoE, BNP, Eot3, RAGE and SGOT may be a key biomarker profile of early disease. Disease detection models based on the identified panels achieved sensitivity (SN) > 80%, specificity (SP) > 70%, and area under receiver operating curve (AUC) of at least 0.80 at prodromal stage (with higher performance at later stages) of the disease. Existing ML models performed poorly in comparison at this stage of the disease, suggesting that the underlying protein panels may not be suitable for early disease detection. Our results demonstrate the feasibility of early detection of AD using non-amyloid based biomarkers.

Breath-based non-invasive diagnosis of Alzheimer's disease: a pilot study.

Early detection of Alzheimer's disease (AD) will help researchers to better understand the disease and develop improved treatments. Recent developments have thus focused on identifying biomarkers for mild cognitive impairment due to AD (MCI) and AD during the preclinical phase. The aim of this pilot study is to determine whether exhaled volatile organic compounds (VOCs) can be used as a non-invasive method to distinguish controls from MCI, controls from AD and to determine whether there are differences between MCI and AD. The study used gas chromatography-ion mobility spectrometry (GC-IMS) techniques. Confounding factors, such as age, smoking habits, gender and alcohol consumption are investigated to demonstrate the efficacy of results. One hundred subjects were recruited including 50 controls, 25 AD and 25 MCI patients. The subject cohort was age- and gender-matched to minimise bias. Breath samples were analysed using a commercial GC-IMS instrument (G.A.S. BreathSpec, Dortmund, Germany). Data analysis indicates that the GC-IMS signal was consistently able to separate between diagnostic groups [AUC ± 95%, sensitivity, specificity], controls versus MCI: [0.77 (0.64-0.90), 0.68, 0.80], controls versus AD: [0.83 (0.72-0.94), 0.60, 0.96], and MCI versus AD: [0.70 (0.55-0.85), 0.60, 0.84]. VOC analysis indicates that six compounds play a crucial role in distinguishing between diagnostic groups. Analysis of possible confounding factors indicate that gender, age, smoking habits and alcohol consumption have insignificant influence on breath content. This pilot study confirms the utility of exhaled breath analysis to distinguish between AD, MCI and control subjects. Thus, GC-IMS offers great potential as a non-invasive, high-throughput, diagnostic technique for diagnosing and potentially monitoring AD in a clinical setting.

What electrophysiology tells us about Alzheimer's disease: a window into the synchronization and connectivity of brain neurons.

Electrophysiology provides a real-time readout of neural functions and network capability in different brain states, on temporal (fractions of milliseconds) and spatial (micro, meso, and macro) scales unmet by other methodologies. However, current international guidelines do not endorse the use of electroencephalographic (EEG)/magnetoencephalographic (MEG) biomarkers in clinical trials performed in patients with Alzheimer's disease (AD), despite a surge in recent validated evidence. This position paper of the ISTAART Electrophysiology Professional Interest Area endorses consolidated and translational electrophysiological techniques applied to both experimental animal models of AD and patients, to probe the effects of AD neuropathology (i.e., brain amyloidosis, tauopathy, and neurodegeneration) on neurophysiological mechanisms underpinning neural excitation/inhibition and neurotransmission as well as brain network dynamics, synchronization, and functional connectivity, reflecting thalamocortical and corticocortical residual capacity. Converging evidence shows relationships between abnormalities in EEG/MEG markers and cognitive deficits in groups of AD patients at different disease stages. The supporting evidence for the application of electrophysiology in AD clinical research as well as drug discovery pathways warrants an international initiative to include the use of EEG/MEG biomarkers in the main multicentric projects planned in AD patients, to produce conclusive findings challenging the present regulatory requirements and guidelines for AD studies.

Machine-learning based identification of undiagnosed dementia in primary care: a feasibility study.

BACKGROUND: Up to half of patients with dementia may not receive a formal diagnosis, limiting access to appropriate services. It is hypothesised that it may be possible to identify undiagnosed dementia from a profile of symptoms recorded in routine clinical practice. AIM: The aim of this study is to develop a machine learning-based model that could be used in general practice to detect dementia from routinely collected NHS data. The model would be a useful tool for identifying people who may be living with dementia but have not been formally diagnosed. DESIGN & SETTING: The study involved a case-control design and analysis of primary care data routinely collected over a 2-year period. Dementia diagnosed during the study period was compared to no diagnosis of dementia during the same period using pseudonymised routinely collected primary care clinical data. METHOD: Routinely collected Read-encoded data were obtained from 18 consenting GP surgeries across Devon, for 26 483 patients aged >65 years. The authors determined Read codes assigned to patients that may contribute to dementia risk. These codes were used as features to train a machine-learning classification model to identify patients that may have underlying dementia. RESULTS: The model obtained sensitivity and specificity values of 84.47% and 86.67%, respectively. CONCLUSION: The results show that routinely collected primary care data may be used to identify undiagnosed dementia. The methodology is promising and, if successfully developed and deployed, may help to increase dementia diagnosis in primary care.

Label-Free Sensors Based on Graphene Field-Effect Transistors for the Detection of Human Chorionic Gonadotropin Cancer Risk Biomarker.

We report on the development of label-free chemical vapour deposition (CVD) graphene field effect transistor (GFET) immunosensors for the sensitive detection of Human Chorionic Gonadotropin (hCG), a glycoprotein risk biomarker of certain cancers. The GFET sensors were fabricated on Si/SiO2 substrate using photolithography with evaporated chromium and sputtered gold contacts. GFET channels were functionalised with a linker molecule to an immobile anti-hCG antibody on the surface of graphene. The binding reaction of the antibody with varying concentration levels of hCG antigen demonstrated the limit of detection of the GFET sensors to be below 1 pg/mL using four-probe electrical measurements. We also show that annealing can significantly improve the carrier transport properties of GFETs and shift the Dirac point (Fermi level) with reduced p-doping in back-gated measurements. The developed GFET biosensors are generic and could find applications in a broad range of medical diagnostics in addition to cancer, such as neurodegenerative (Alzheimer's and Parkinson's ) and cardiovascular disorders.

Higuchi fractal dimension of the electroencephalogram as a biomarker for early detection of Alzheimer's disease.

It is widely accepted that early diagnosis of Alzheimer's disease (AD) makes it possible for patients to gain access to appropriate health care services and would facilitate the development of new therapies. AD starts many years before its clinical manifestations and a biomarker that provides a measure of changes in the brain in this period would be useful for early diagnosis of AD. Given the rapid increase in the number of older people suffering from AD, there is a need for an accurate, low-cost and easy to use biomarkers that could be used to detect AD in its early stages. Potentially, the electroencephalogram (EEG) can play a vital role in this but at present, no reliable EEG biomarker exists for early diagnosis of AD. The gradual slowing of brain activity caused by AD starts from the back of the brain and spreads out towards other parts. Consequently, determining the brain regions that are first affected by AD may be useful in its early diagnosis. Higuchi fractal dimension (HFD) has characteristics which make it suited to capturing region-specific neural changes due to AD. The aim of this study is to investigate the potential of HFD of the EEG as a biomarker which is associated with the brain region first affected by AD. Mean HFD value was calculated for all channels of EEG signals recorded from 52 subjects (20-AD and 32-normal). Then, p-values were calculated between the two groups (AD and normal) to detect EEG channels that have a significant association with AD. k-nearest neighbor (KNN) algorithm was used to compute the distance between AD patients and normal subjects in the classification. Our results show that AD patients have significantly lower HFD values in the parietal areas. HFD values for channels in these areas were used to discriminate between AD and normal subjects with a sensitivity and specificity values of 100% and 80%, respectively.

Discovery of Novel Biomarkers for Alzheimer's Disease from Blood.

Blood-based biomarkers for Alzheimer's disease would be very valuable because blood is a more accessible biofluid and is suitable for repeated sampling. However, currently there are no robust and reliable blood-based biomarkers for practical diagnosis. In this study we used a knowledge-based protein feature pool and two novel support vector machine embedded feature selection methods to find panels consisting of two and three biomarkers. We validated these biomarker sets using another serum cohort and an RNA profile cohort from the brain. Our panels included the proteins ECH1, NHLRC2, HOXB7, FN1, ERBB2, and SLC6A13 and demonstrated promising sensitivity (>87%), specificity (>91%), and accuracy (>89%).

Changes in the EEG amplitude as a biomarker for early detection of Alzheimer's disease.

The rapid increase in the number of older people with Alzheimer's disease (AD) and other forms of dementia represents one of the major challenges to the health and social care systems. Early detection of AD makes it possible for patients to access appropriate services and to benefit from new treatments and therapies, as and when they become available. The onset of AD starts many years before the clinical symptoms become clear. A biomarker that can measure the brain changes in this period would be useful for early diagnosis of AD. Potentially, the electroencephalogram (EEG) can play a valuable role in early detection of AD. Damage in the brain due to AD leads to changes in the information processing activity of the brain and the EEG which can be quantified as a biomarker. The objective of the study reported in this paper is to develop robust EEG-based biomarkers for detecting AD in its early stages. We present a new approach to quantify the slowing of the EEG, one of the most consistent features at different stages of dementia, based on changes in the EEG amplitudes (ΔEEGA). The new approach has sensitivity and specificity values of 100% and 88.88%, respectively, and outperformed the Lempel-Ziv Complexity (LZC) approach in discriminating between AD and normal subjects.

Tsallis entropy as a biomarker for detection of Alzheimer's disease.

Alzheimer's disease (AD) and other forms of dementia are one of the major public health and social challenges of our time because of the large number of people affected. Early diagnosis is important for patients and their families to get maximum benefits from access to health and social care services and to plan for the future. EEG provides useful insight into brain functions and can play a useful role as a first line of decision-support tool for early detection and diagnosis of dementia. It is non-invasive, low-cost and has a high temporal resolution. The functions of brain cells are affected by damage caused by dementia and this in turn causes changes in the features of the EEG. Information theoretic methods have emerged as a potentially useful way to quantify changes in the EEG as biomarkers of dementia. Tsallis entropy has been shown to be one of the most promising information theoretic methods for quantifying changes in the EEG. In this paper, we develop the approach further. This has yielded an enhanced performance compared to existing approaches.

Principal Dynamic Mode Analysis of EEG Data for Assisting the Diagnosis of Alzheimer's Disease.

We examine whether modeling of the causal dynamic relationships between frontal and occipital electroencephalogram (EEG) time-series recordings reveal reliable differentiating characteristics of Alzheimer's patients versus control subjects in a manner that may assist clinical diagnosis of Alzheimer's disease (AD). The proposed modeling approach utilizes the concept of principal dynamic modes (PDMs) and their associated nonlinear functions (ANF) and hypothesizes that the ANFs of some PDMs for the AD patients will be distinct from their counterparts in control subjects. To this purpose, global PDMs are extracted from 1-min EEG signals of 17 AD patients and 24 control subjects at rest using Volterra models estimated via Laguerre expansions, whereby the O1 or O2 recording is viewed as the input signal and the F3 or F4 recording as the output signal. Subsequent singular value decomposition of the estimated Volterra kernels yields the global PDMs that represent an efficient basis of functions for the representation of the EEG dynamics in all subjects. The respective ANFs are computed for each subject and characterize the specific dynamics of each subject. For comparison, signal features traditionally used in the analysis of EEG signals in AD are computed as benchmark. The results indicate that the ANFs of two specific PDMs, corresponding to the delta-theta and alpha bands, can delineate the two groups well.

Changes in the MEG background activity in patients with positive symptoms of schizophrenia: spectral analysis and impact of age.

The frequency spectrum of the magnetoencephalogram (MEG) background activity was analysed in 15 schizophrenia (SCH) patients with predominant positive symptoms and 17 age-matched healthy control subjects using the following variables: median frequency (MF), spectral entropy (SpecEn) and relative power in delta (RPδ), theta (RPθ), lower alpha (RPα1), upper alpha (RPα2), beta (RPß) and gamma (RPγ) bands. We found significant differences between the two subject groups in the average level of MF and RPγ in some regions of the scalp. Additionally, the MF, SpecEn, RPß and RPγ values of SCH patients with positive symptoms had a different dependence on age as compared with the results of control subjects, suggesting that SCH affects the way in which the brain activity evolves with age. Moreover, we also classified the MEG signals by means of a cross-validated feature selection process followed by a logistic regression. The subjects were classified with 71.3% accuracy and an area under the ROC curve of 0.741. Thus, the spectral and classification analysis of the MEG in SCH may provide insights into how this condition affects the brain activity and may help in its early detection.

Machine learning-based method for personalized and cost-effective detection of Alzheimer's disease.

Diagnosis of Alzheimer's disease (AD) is often difficult, especially early in the disease process at the stage of mild cognitive impairment (MCI). Yet, it is at this stage that treatment is most likely to be effective, so there would be great advantages in improving the diagnosis process. We describe and test a machine learning approach for personalized and cost-effective diagnosis of AD. It uses locally weighted learning to tailor a classifier model to each patient and computes the sequence of biomarkers most informative or cost-effective to diagnose patients. Using ADNI data, we classified AD versus controls and MCI patients who progressed to AD within a year, against those who did not. The approach performed similarly to considering all data at once, while significantly reducing the number (and cost) of the biomarkers needed to achieve a confident diagnosis for each patient. Thus, it may contribute to a personalized and effective detection of AD, and may prove useful in clinical settings.

Bioprofile analysis: a new approach for the analysis of biomedical data in Alzheimer's disease.

This article presents a new approach for the analysis of biomedical data to support the management and care of patients with Alzheimer's disease (AD). The increase in prevalence of neurodegenerative disorders such as AD has led to a need for objective means to assist clinicians with the analysis and interpretation of complex biomedical data. To this end, we propose a "Bioprofile" analysis to reveal the pattern of disease in the subject's biodata. From the Bioprofile, personal "Bioindices" that indicate how closely a subject's data resemble the pattern of AD can be derived. We used an unsupervised technique (k-means) to cluster variables of the ADNI database so that subjects are divisible into those with the Bioprofile of AD and those without it. Results revealed that there is an "AD pattern" in the biodata of most AD and mild cognitive impairment (MCI) patients and some controls. This pattern agrees with a recent hypothetical model of AD evolution. We also assessed how the Bioindices changed with time and we found that the Bioprofile was associated with the risk of progressing from MCI to AD. Hence, the Bioprofile analysis is a promising methodology that may potentially provide a complementary new way of interpreting biomedical data. Furthermore, the concept of the Bioprofile could be extended to other neurodegenerative diseases.

Early detection and characterization of Alzheimer's disease in clinical scenarios using Bioprofile concepts and K-means.

Alzheimer's Disease (AD) is the most common neurodegenerative disease in elderly people. There is a need for objective means to detect AD early to allow targeted interventions and to monitor response to treatment. To help clinicians in these tasks, we propose the creation of the Bioprofile of AD. A Bioprofile should reveal key patterns of a disease in the subject's biodata. We applied k-means clustering to data features taken from the ADNI database to divide the subjects into pathologic and non-pathologic groups in five clinical scenarios. The preliminary results confirm previous findings and show that there is an important AD pattern in the biodata of controls, AD, and Mild Cognitive Impairment (MCI) patients. Furthermore, the Bioprofile could help in the early detection of AD at the MCI stage since it divided the MCI subjects into groups with different rates of conversion to AD.

Machine Learning classification of MRI features of Alzheimer's disease and mild cognitive impairment subjects to reduce the sample size in clinical trials.

There is a need for objective tools to help clinicians to diagnose Alzheimer's Disease (AD) early and accurately and to conduct Clinical Trials (CTs) with fewer patients. Magnetic Resonance Imaging (MRI) is a promising AD biomarker but no single MRI feature is optimal for all disease stages. Machine Learning classification can address these challenges. In this study, we have investigated the classification of MRI features from AD, Mild Cognitive Impairment (MCI), and control subjects from ADNI with four techniques. The highest accuracy rates for the classification of controls against ADs and MCIs were 89.2% and 72.7%, respectively. Moreover, we used the classifiers to select AD and MCI subjects who are most likely to decline for inclusion in hypothetical CTs. Using the hippocampal volume as an outcome measure, we found that the required group sizes for the CTs were reduced from 197 to 117 AD patients and from 366 to 215 MCI subjects.

A new method for modeling preoperative diagnosis of ovarian tumors.

In this paper, we present a sequential nonuniform procedure, an inference method which combines feature selection based on the Kullback information gain and a step-wise classification procedure to produce a reliable, interpretable, and robust model. We applied the model to an ovarian tumor data set to distinguish between malignant and benign tumors. The performance of the model was assessed using receiver operating characteristic (ROC) analysis and gave an overall accuracy over 85%, and area under the curve (AUC) of 0.887 which compares well with existing methods. The method presented here is significant because of its ability to handle missing values, and it only uses a small number of variables which are graded according to their discriminative relevance. This, together with the fact that the resulting model is interpretable and has good performance, is likely to lead to widespread clinical acceptance of the method. The method is also generic and can be readily adapted for other classifications problems in biomedicine.

Characterization of EEGs in Alzheimer's disease using information theoretic methods.

The number of people that now go on to develop Alzheimer's disease (AD) and other types of dementia is rapidly rising. For maximum benefits from new treatments, the disease should be diagnosed as early as possible, but this is difficult with current clinical criteria. Potentially, the EEG can serve as an objective, first line of decision support tool to improve diagnosis. It is non-invasive, widely available, low-cost and could be carried out rapidly in the high-risk age group that will develop AD. Changes in the EEG due to the dementing process could be quantified as an index or marker. In this paper, we investigate two information theoretic methods (Tsallis entropy and universal compression algorithm) as a way to generate potentially robust markers from the EEG. The hypothesis is that the information theoretic makers for AD are significantly different to those of normal subjects. An attraction of the information theoretic approach is that, unlike most existing methods, there may be a natural link between the underlying ideas of information theoretic methods, the physiology of AD and its impact on brain functions. Data compression has not been investigated as a means of generating EEG markers before and is attractive because it does not require a priori knowledge of the source model. In this paper, we focus on the LZW algorithm because of its sound theoretical foundation. We used the LZW algorithm and Tsallis model to compute the markers (compression ratios and normalized entropies, respectively) from two EEG datasets.

Preoperative prediction of malignancy of ovarian tumours using modified sequential non-uniform procedure.

In this paper, we present an extension of sequential non-uniform procedure (SNuP) with application of the method to ovarian tumour data, obtained during multicentre study by the International Ovarian Tumour Analysis Group (IOTA). The inference method combines feature selection based on the Kullback information gain and a step-wise classification procedure to produce a reliable, interpretable and robust model. In particular, we extend SNuP to enable it to handle continuous variables without the need for manual specification of thresholds. We applied the extended model to an ovarian tumour data set to distinguish between malignant and benign tumours. The performance of the model was assessed using ROC analysis and gave 86.9% of sensitivity and 84.3% of specificity with overall accuracy level of 84.9%.