Abstract 14900: Human Atrial Fibrillation Drivers Revealed by Machine Learning Using Multicomponent Electrograms and 3D CE-MRI Structural Features

Circulation ◽  
2020 ◽  
Vol 142 (Suppl_3) ◽  
Author(s):  
Aleksei Mikhailov ◽  
Brian Hansen ◽  
Matthew Fazio ◽  
Stanislav Zakharkin ◽  
Jichao Zhao ◽  
...  
Keyword(s):  
Machine Learning ◽  
Atrial Fibrillation ◽  
Gold Standard ◽  
Near Infrared ◽  
Structural Features ◽  
Multiple Features ◽  
Mri Features ◽  
Magnetic Resonance Imaging Mri ◽  
Driver Identification

Introduction: Conventional multielectrode mapping is not sufficient to reveal subsurface intramural activation. Thus, atrial fibrillation (AF) driver identification remains challenging. To overcome these limitations we utilized machine learning (ML) to identify AF drivers based on the combination of electrogram (EGM) and 3D structural magnetic resonance imaging (MRI) features. Hypothesis: Detailed electrogram features analysis, including minor deflections, combined with local structural features, can be used to define AF driver. Methods: Sustained AF was mapped in coronary perfused explanted human atria (n=7) with near-infrared optical mapping (NIOM) (0.3-0.9mm 2 resolution) and 64-electrode mapping catheter (3mm 2 resolution). Unipolar EGMs were analyzed for multiple features of the steepest negative deflection and the 2nd-4th steepest deflections in multicomponent EGMs. Atria underwent 9.4T MRI (154-180μm 3 resolution) with gadolinium enhancement and histology validation of fibrosis. Both 3D structural and EGM data from NIOM defined driver and non-driver regions were processed by ML algorithms (LR; PLSDA; GBM; CRF; PSVM; RSVM) using double cross-validation. Results: AF drivers’ reentrant tracks were defined by NIOM activation mapping, the gold-standard, and confirmed by targeted ablation. The best performing ML algorithm (PLSDA) correctly classified mapped driver region with 76.1% accuracy on the testing data. The most important features included sub-endocardial fibrosis, sub-epicardial fiber orientation, local wall thickness, beat-to-beat variability of multicomponent EGM deflections. Conclusions: The ML models pre-trained on combined EGM and structural features allow efficient classification of AF driver vs non-driver regions defined by the NIOM gold-standard. The results suggest that AF driver substrates formed by the combination of 3D fibrotic structural features, which correlate with local EGM characteristics.

scholarly journals Optical Mapping-Validated Machine Learning Improves Atrial Fibrillation Driver Detection by Multi-Electrode Mapping

2020 ◽  
Vol 13 (10) ◽  
Author(s):  
Alexander M. Zolotarev ◽  
Brian J. Hansen ◽  
Ekaterina A. Ivanova ◽  
Katelynn M. Helfrich ◽  
Ning Li ◽  
...  
Keyword(s):  
Machine Learning ◽  
Atrial Fibrillation ◽  
Near Infrared ◽  
Fourier Spectrum ◽  
Optical Mapping ◽  
Binary Classification ◽  
Future Application ◽  
Data Set ◽  
Frequency Spectra

Background: Atrial fibrillation (AF) can be maintained by localized intramural reentrant drivers. However, AF driver detection by clinical surface-only multielectrode mapping (MEM) has relied on subjective interpretation of activation maps. We hypothesized that application of machine learning to electrogram frequency spectra may accurately automate driver detection by MEM and add some objectivity to the interpretation of MEM findings. Methods: Temporally and spatially stable single AF drivers were mapped simultaneously in explanted human atria (n=11) by subsurface near-infrared optical mapping (NIOM; 0.3 mm 2 resolution) and 64-electrode MEM (higher density or lower density with 3 and 9 mm 2 resolution, respectively). Unipolar MEM and NIOM recordings were processed by Fourier transform analysis into 28 407 total Fourier spectra. Thirty-five features for machine learning were extracted from each Fourier spectrum. Results: Targeted driver ablation and NIOM activation maps efficiently defined the center and periphery of AF driver preferential tracks and provided validated annotations for driver versus nondriver electrodes in MEM arrays. Compared with analysis of single electrogram frequency features, averaging the features from each of the 8 neighboring electrodes, significantly improved classification of AF driver electrograms. The classification metrics increased when less strict annotation, including driver periphery electrodes, were added to driver center annotation. Notably, f1-score for the binary classification of higher-density catheter data set was significantly higher than that of lower-density catheter (0.81±0.02 versus 0.66±0.04, P <0.05). The trained algorithm correctly highlighted 86% of driver regions with higher density but only 80% with lower-density MEM arrays (81% for lower-density+higher-density arrays together). Conclusions: The machine learning model pretrained on Fourier spectrum features allows efficient classification of electrograms recordings as AF driver or nondriver compared with the NIOM gold-standard. Future application of NIOM-validated machine learning approach may improve the accuracy of AF driver detection for targeted ablation treatment in patients.

scholarly journals Classification of Mammograms using Various Feature Extraction Methods and Machine Learning

2019 ◽  
Vol 8 (2) ◽  
pp. 5401-5405
Keyword(s):  
Breast Cancer ◽  
Machine Learning ◽  
Extraction Methods ◽  
Machine Learning Algorithms ◽  
Support Vector ◽  
Multiple Features ◽  
Human Power ◽  
Hsv Histogram ◽  
Mammogram Images

Breast cancer is an alarming disease which takes millions of lives every year. In 2018, it was anticipated that 627,000 women died due to breast cancer – which is around 15% of all deaths caused due to different types of cancers among women. Currently, risk factors of breast cancer cannot be avoided, and early detection is the only way of survival. Automated detection of breast cancer with the help of image processing methods and machine learning algorithms helps in giving more accurate results and less human power. In the proposed system, multiple features are extracted using HSV histogram, LBP, GLCM, 2-D DWT. Support vector machine and LIBSVM classifiers are used for the classification of mammogram images if it’s benign or malign in nature. For classification, the INbreast dataset have been used which includes 115 cases containing 410 images. The dataset is divided into benign and malign category based upon BI-RAIDS scale. According to this partition we have 243 benign images and 100 malign images present in this dataset and a feature matrix of 595 features in total is generated for balanced and unbalanced datasets respectively and fed into SVM and LIBSVM to distinguish the data. The balanced datasets on LIBSVM gave best results with 92% accuracy, 84% sensitivity, 100% specificity and 91.30% F1 score followed by SVM which gave 75% accuracy, 73.61% sensitivity, 76.66% specificity and 75.8% F1 score.

scholarly journals Multivariate and machine learning approaches for honey botanical origin authentication using near infrared spectroscopy

2019 ◽  
Vol 27 (1) ◽  
pp. 65-74 ◽  
Author(s):  
Vittoria Bisutti ◽  
Roberta Merlanti ◽  
Lorenzo Serva ◽  
Lorena Lucatello ◽  
Massimo Mirisola ◽  
...  
Keyword(s):  
Machine Learning ◽  
Discriminant Analysis ◽  
Least Squares ◽  
Partial Least Squares ◽  
Near Infrared ◽  
Canonical Discriminant Analysis ◽  
Learning Approaches ◽  
Botanical Origin ◽  
Infrared Spectral

In this work the feasibility of near infrared spectroscopy was evaluated combined with chemometric approaches, as a tool for the botanical origin prediction of 119 honey samples. Four varieties related to polyfloral, acacia, chestnut, and linden were first characterized by their physical–chemical parameters and then analyzed in triplicate using a near infrared spectrophotometer equipped with an optical path gold reflector. Three different classifiers were built on distinct multivariate and machine learning approaches for honey botanical classification. A partial least squares discriminant analysis was used as a first approach to build a predictive model for honey classification. Spectra pretreatments named autoscale, standard normal variate, detrending, first derivative, and smoothing were applied for the reduction of scattering related to the presence of particle size, like glucose crystals. The values of the descriptive statistics of the partial least squares discriminant analysis model allowed a sufficient floral group prediction for the acacia and polyfloral honeys but not in the cases of chestnut and linden. The second classifier, based on a support vector machine, allowed a better classification of acacia and polyfloral and also achieved the classification of chestnut. The linden samples instead remained unclassified. A further investigation, aimed to improve the botanical discrimination, exploited a feature selection algorithm named Boruta, which assigned a pool of 39 informative averaged near infrared spectral variables on which a canonical discriminant analysis was assessed. The canonical discriminant analysis accounted a better separation of samples according to the botanical origin than the partial least squares discriminant analysis. The approach used has permitted to achieve a complete authentication of the acacia honeys but not a precise segregation of polyfloral ones. The comparison between the variables important in projection and the Boruta pool showed that the informative wavelengths are partially shared especially in the middle and far band of the near infrared spectral range.

scholarly journals Automating the assessment of biofouling in images using expert agreement as a gold standard

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Nathaniel J. Bloomfield ◽  
Susan Wei ◽  
Bartholomew A. Woodham ◽  
Peter Wilkinson ◽  
Andrew P. Robinson
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Gold Standard ◽  
Automated Analysis ◽  
Marine Species ◽  
Shipping Industry ◽  
Classification Of Images ◽  
Fuel Costs ◽  
Deep Learning Model

AbstractBiofouling is the accumulation of organisms on surfaces immersed in water. It is of particular concern to the international shipping industry because it increases fuel costs and presents a biosecurity risk by providing a pathway for non-indigenous marine species to establish in new areas. There is growing interest within jurisdictions to strengthen biofouling risk-management regulations, but it is expensive to conduct in-water inspections and assess the collected data to determine the biofouling state of vessel hulls. Machine learning is well suited to tackle the latter challenge, and here we apply deep learning to automate the classification of images from in-water inspections to identify the presence and severity of fouling. We combined several datasets to obtain over 10,000 images collected from in-water surveys which were annotated by a group biofouling experts. We compared the annotations from three experts on a 120-sample subset of these images, and found that they showed 89% agreement (95% CI: 87–92%). Subsequent labelling of the whole dataset by one of these experts achieved similar levels of agreement with this group of experts, which we defined as performing at most 5% worse (p $$=$$ = 0.009–0.054). Using these expert labels, we were able to train a deep learning model that also agreed similarly with the group of experts (p $$=$$ = 0.001–0.014), demonstrating that automated analysis of biofouling in images is feasible and effective using this method.

scholarly journals Diffuse midline gliomas, H3 K27M-mutant are associated with less peritumoral edema and contrast enhancement in comparison to glioblastomas, H3 K27M-wildtype of midline structures

PLoS ONE ◽  
2021 ◽  
Vol 16 (8) ◽  
pp. e0249647
Author(s):  
Rouzbeh Banan ◽  
Arash Akbarian ◽  
Majid Samii ◽  
Amir Samii ◽  
Helmut Bertalanffy ◽  
...  
Keyword(s):  
Contrast Enhancement ◽  
Lower Probability ◽  
Imaging Features ◽  
Peritumoral Edema ◽  
Mri Features ◽  
Histological Criteria ◽  
Magnetic Resonance Imaging Mri ◽  
Lower Contrast ◽  
Diffuse Midline Glioma

Purpose The entity ‘diffuse midline glioma, H3 K27M-mutant (DMG)’ was introduced in the revised 4th edition of the 2016 WHO classification of brain tumors. However, there are only a few reports on magnetic resonance imaging (MRI) of these tumors. Thus, we conducted a retrospective survey focused on MRI features of DMG compared to midline glioblastomas H3 K27M-wildtype (mGBM-H3wt). Methods We identified 24 DMG cases and 19 mGBM-H3wt patients as controls. After being retrospectively evaluated for microscopic evidence of microvascular proliferations (MVP) and tumor necrosis by two experienced neuropathologists to identify the defining histological criteria of mGBM-H3wt, the samples were further analyzed by two experienced readers regarding imaging features such as shape, peritumoral edema and contrast enhancement. Results The DMG were found in the thalamus in 37.5% of cases (controls 63%), in the brainstem in 50% (vs. 32%) and spinal cord in 12.5% (vs. 5%). In MRI and considering MVP, DMG were found to be by far less likely to develop peritumoral edema (OR: 0.13; 95%-CL: 0.02–0.62) (p = 0.010). They, similarly, were associated with a significantly lower probability of developing strong contrast enhancement compared to mGBM-H3wt (OR: 0.10; 95%-CL: 0.02–0.47) (P = 0.003). Conclusion Despite having highly variable imaging features, DMG exhibited markedly less edema and lower contrast enhancement in MRI compared to mGBM-H3wt. Of these features, the enhancement level was associated with evidence of MVP.

scholarly journals Automated Electronic Phenotyping of Cardioembolic Stroke

Stroke ◽  
2021 ◽  
Vol 52 (1) ◽  
pp. 181-189 ◽  
Author(s):  
Wyliena Guan ◽  
Darae Ko ◽  
Shaan Khurshid ◽  
Ana T. Trisini Lipsanopoulos ◽  
Jeffrey M. Ashburner ◽  
...  
Keyword(s):  
Machine Learning ◽  
Atrial Fibrillation ◽  
Ischemic Stroke ◽  
Positive Predictive Value ◽  
Predictive Value ◽  
Regular Expression ◽  
Massachusetts General Hospital ◽  
Cardioembolic Stroke ◽  
Stroke Registry

Background and Purpose: Oral anticoagulation is generally indicated for cardioembolic strokes, but not for other stroke causes. Consequently, subtype classification of ischemic stroke is important for risk stratification and secondary prevention. Because manual classification of ischemic stroke is time-intensive, we assessed the accuracy of automated algorithms for performing cardioembolic stroke subtyping using an electronic health record (EHR) database. Methods: We adapted TOAST (Trial of ORG 10172 in Acute Stroke Treatment) features associated with cardioembolic stroke for derivation in the EHR. Using administrative codes and echocardiographic reports within Mass General Brigham Biobank (N=13 079), we iteratively developed EHR-based algorithms to define the TOAST cardioembolic stroke features, revising regular expression algorithms until achieving positive predictive value ≥80%. We compared several machine learning-based statistical algorithms for discriminating cardioembolic stroke using the feature algorithms applied to EHR data from 1598 patients with acute ischemic strokes from the Massachusetts General Hospital Ischemic Stroke Registry (2002–2010) with previously adjudicated TOAST and Causative Classification of Stroke subtypes. Results: Regular expression-based feature extraction algorithms achieved a mean positive predictive value of 95% (range, 88%–100%) across 11 echocardiographic features. Among 1598 patients from the Massachusetts General Hospital Ischemic Stroke Registry, 1068 had any cardioembolic stroke feature within predefined time windows in proximity to the stroke event. Cardioembolic stroke tended to occur at an older age, with more TOAST-based comorbidities, and with atrial fibrillation (82.3%). The best model was a random forest with 92.2% accuracy and area under the receiver operating characteristic curve of 91.1% (95% CI, 87.5%–93.9%). Atrial fibrillation, age, dilated cardiomyopathy, congestive heart failure, patent foramen ovale, mitral annulus calcification, and recent myocardial infarction were the most discriminatory features. Conclusions: Machine learning-based identification of cardioembolic stroke using EHR data is feasible. Future work is needed to improve the accuracy of automated cardioembolic stroke identification and assess generalizability of electronic phenotyping algorithms across clinical settings.

Sign in / Sign up

Export Citation Format

Share Document