Prediction of atrial fibrillation from at-home single-lead ECG signals without arrhythmias

Early identification of atrial fibrillation (AF) can reduce the risk of stroke, heart failure, and other serious cardiovascular outcomes. However, paroxysmal AF may not be detected even after a two-week continuous monitoring period. We developed a model to quantify the risk of near-term AF in a two-week period, based on AF-free ECG intervals of up to 24 h from 459,889 patch-based ambulatory single-lead ECG (modified lead II) recordings of up to 14 days. A deep learning model was used to integrate ECG morphology data with demographic and heart rhythm features toward AF prediction. Observing a 1-day AF-free ECG recording, the model with deep learning features produced the most accurate prediction of near-term AF with an area under the curve AUC = 0.80 (95% confidence interval, CI = 0.79-0.81), significantly improving discrimination compared to demographic metrics alone (AUC 0.67; CI = 0.66-0.68). Our model was able to predict incident AF over a two-week time frame with high discrimination, based on AF-free single-lead ECG recordings of various lengths. Application of the model may enable a digital strategy for improving diagnostic capture of AF by risk stratifying individuals with AF-negative ambulatory monitoring for prolonged or recurrent monitoring, potentially leading to more rapid initiation of treatment.

Magnetic resonance elastography (MRE) shows significant reduction of thigh muscle stiffness in healthy older adults

Determining the effect of ageing on thigh muscle stiffness using magnetic resonance elastography (MRE) and investigate whether fat fraction and muscle cross-sectional area (CSA) are related to stiffness. Six healthy older adults in their eighth and ninth decade and eight healthy young men were recruited and underwent a 3 T MRI protocol including MRE and Dixon fat fraction imaging. Muscle stiffness, fat fraction and muscle CSA were calculated in ROIs corresponding to the four quadriceps muscles (i.e. vastus lateralis (VL), vastus medialis (VM), vastus intermedius (VI), rectus femoris (RF)), combined quadriceps, combined hamstrings and adductors and whole thigh. Muscle stiffness was significantly reduced (p < 0.05) in the older group in all measured ROIs except the VI (p = 0.573) and RF (p = 0.081). Similarly, mean fat fraction was significantly increased (p < 0.05) in the older group over all ROIs with the exception of the VI (p = 0.059) and VL muscle groups (p = 0.142). Muscle CSA was significantly reduced in older participants in the VM (p = 0.003) and the combined quadriceps (p = 0.001), hamstrings and adductors (p = 0.008) and whole thigh (p = 0.003). Over the whole thigh, stiffness was significantly negatively correlated with fat fraction (r = − 0.560, p = 0.037) and positively correlated with CSA (r = 0.749, p = 0.002). Stepwise regression analysis revealed that age was the most significant predictor of muscle stiffness (p = 0.001). These results suggest that muscle stiffness is significantly decreased in healthy older adults. Muscle fat fraction and muscle CSA are also significantly changed in older adults; however, age is the most significant predictor of muscle stiffness.

Fast Robust Dejitter and Interslice Discontinuity Removal in MRI Phase Acquisitions: Application to Magnetic Resonance Elastography

MRI phase contrast imaging methods that assemble slice-wise acquisitions into volumes can contain interslice phase discontinuities (IPDs) over the course of the scan from sources, including unavoidable physiological activity. In magnetic resonance elastography (MRE), this can alter wavelength and tissue stiffness estimates, invalidating the analysis. We first model this behavior as jitter along the z-axis of the phase of 3D complex-valued wave volumes. A two-step image processing pipeline is then proposed that removes IPDs. First, constant slicewise phase shift is removed with a novel, non-convex dejittering algorithm. Then, regional physiological noise artifacts are removed with novel filtering of 3D wavelet coefficients. Calibration of two pipeline coefficients, the dejitter parameter α and the wavelet band high-pass coefficient ω_c , was first performed on a finite-element method brain phantom. A comparative investigation was then performed, on a cohort of 48 brain acquisitions, of four approaches to IPDs: 1) the proposed method; 2) a "control" condition of neglect of IPDs; 3) an anisotropic wavelet-based method; and 4) a method of in-plane (2D) processing. The present method showed medians of [Formula: see text] Pa for a multifrequency wave inversion centered at 40 Hz which was within 6% of methods 3) and 4), while neglect produced [Formula: see text] estimates a mean of 17% lower. The proposed method reduced the value range of the cohort against methods 3) and 4) by 29% and 31%, respectively. Such reduction in variance enhances the ability of brain MRE to predict subtler physiological changes. Our theoretical approach further enables more powerful applications of fundamental findings in noise and denoising to MRE.

Perfusion alters stiffness of deep gray matter

Viscoelastic properties of the brain reflect tissue architecture at multiple length scales. However, little is known about the relation between vital tissue functions, such as perfusion, and the macroscopic mechanical properties of cerebral tissue. In this study, arterial spin labelling is paired with magnetic resonance elastography to investigate the relationship between tissue stiffness and cerebral blood flow (CBF) in the in vivo human brain. The viscoelastic modulus, | G*|, and CBF were studied in deep gray matter (DGM) of 14 healthy male volunteers in the following sub-regions: putamen, nucleus accumbens, hippocampus, thalamus, globus pallidus, and amygdala. CBF was further normalized by vessel area data to obtain the flux rate q which is proportional to the perfusion pressure gradient. The striatum (represented by putamen and nucleus accumbens) was distinct from the other DGM regions by displaying markedly higher stiffness and perfusion values. q was a predictive marker for DGM stiffness as analyzed by linear regression | G*| =  q·(4.2 ± 0.6)kPa·s + (0.80 ± 0.06)kPa ( R²= 0.92, P = 0.006). These results suggest a high sensitivity of MRE in DGM to perfusion pressure. The distinct mechano-vascular properties of striatum tissue, as compared to the rest of DGM, may reflect elevated perfusion pressure, which could explain the well-known susceptibility of the putamen to hemorrhages.

Finite element analysis to investigate variability of MR elastography in the human thigh

PURPOSE

To develop finite element analysis (FEA) of magnetic resonance elastography (MRE) in the human thigh and investigate inter-individual variability of measurement of muscle mechanical properties.

METHODS

Segmentation was performed on MRI datasets of the human thigh from 5 individuals and FEA models consisting of 12 muscles and surrounding tissue created. The same material properties were applied to each tissue type and a previously developed transient FEA method of simulating MRE using Abaqus was performed at 4 frequencies. Synthetic noise was applied to the simulated data at various levels before inversion was performed using the Elastography Software Pipeline. Maps of material properties were created and visually assessed to determine key features. The coefficient of variation (CoV) was used to assess the variability of measurements in each individual muscle and in the groups of muscles across the subjects. Mean measurements for the set of muscles were ranked in size order and compared with the expected ranking.

RESULTS

At noise levels of 2% the CoV in measurements of |G^*| ranged from 5.3 to 21.9% and from 7.1 to 36.1% for measurements of ϕ in the individual muscles. A positive correlation (R² value 0.80) was attained when the expected and measured |G^*| ranking were compared, whilst a negative correlation (R² value 0.43) was found for ϕ.

CONCLUSIONS

Created elastograms demonstrated good definition of muscle structure and were robust to noise. Variability of measurements across the 5 subjects was dramatically lower for |G^*| than it was for ϕ. This large variability in ϕ measurements was attributed to artefacts.

MR elastography measurement of the effect of passive warmup prior to eccentric exercise on thigh muscle mechanical properties

Purpose

To investigate the effect of warmup by application of the thermal agent Deep Heat (DH) on muscle mechanical properties using magnetic resonance elastography (MRE) at 3T before and after exercise‐induced muscle damage (EIMD).

Materials and Methods

Twenty male participants performed an individualized protocol designed to induce EIMD in the quadriceps. DH was applied to the thigh in 50% of the participants before exercise. MRE, T₂‐weighted MRI, maximal voluntary contraction (MVC), creatine kinase (CK) concentration, and muscle soreness were measured before and after the protocol to assess EIMD effects. Five participants were excluded: four having not experienced EIMD and one due to incidental findings.

Results

Total workload performed during the EIMD protocol was greater in the DH group than the control group (P < 0.03), despite no significant differences in baseline MVC (P = 0.23). Shear stiffness |G*| increased in the rectus femoris (RF) muscle in both groups (P < 0.03); however, DH was not a significant between‐group factor (P =  0.15). MVC values returned to baseline faster in the DH group (5 days) than the control group (7 days). Participants who displayed hyperintensity on T₂‐weighted images had a greater stiffness increase following damage than those without: RF; 0.61 kPa vs. 0.15 kPa, P < 0.006, vastus intermedius; 0.34 kPa vs. 0.03 kPa, P = 0.06.

Conclusion

EIMD produces increased muscle stiffness as measured by MRE, with the change in |G*| significantly increased when T₂ hyperintensity was present. DH did not affect CK concentration or soreness; however, DH participants produced greater workload during the EIMD protocol and exhibited accelerated MVC recovery.

Level of Evidence: 1

Technical Efficacy: Stage 2

J. Magn. Reson. Imaging 2017;46:1115–1127.

Nonlinear multiscale regularisation in MR elastography: Towards fine feature mapping

Fine-featured elastograms may provide additional information of radiological interest in the context of in vivo elastography. Here a new image processing pipeline called ESP (Elastography Software Pipeline) is developed to create Magnetic Resonance Elastography (MRE) maps of viscoelastic parameters (complex modulus magnitude |G^*| and loss angle ϕ) that preserve fine-scale information through nonlinear, multi-scale extensions of typical MRE post-processing techniques.

METHODS

A new MRE image processing pipeline was developed that incorporates wavelet-domain denoising, image-driven noise estimation, and feature detection. ESP was first validated using simulated data, including viscoelastic Finite Element Method (FEM) simulations, at multiple noise levels. ESP images were compared with MDEV pipeline images, both in the FEM models and in three ten-subject cohorts of brain, thigh, and liver acquisitions. ESP and MDEV mean values were compared to 2D local frequency estimation (LFE) mean values for the same cohorts as a benchmark. Finally, the proportion of spectral energy at fine frequencies was quantified using the Reduced Energy Ratio (RER) for both ESP and MDEV.

RESULTS

Blind estimates of added noise (σ) were within 5.3% ± 2.6% of prescribed, and the same technique estimated σ in the in vivo cohorts at 1.7 ± 0.8%. A 5 × 5 × 5 truncated Gabor filter bank effectively detects local spatial frequencies at wavelengths λ ≤ 10px. For FEM inversions, mean |G^*| of hard target, soft target, and background remained within 8% of prescribed up to σ=20%, and mean ϕ results were within 10%, excepting hard target ϕ, which required redrawing around a ring artefact to achieve similar accuracy. Inspection of FEM |G^*| images showed some spatial distortion around hard target boundaries and inspection of ϕ images showed ring artefacts around the same target. For the in vivo cohorts, ESP results showed mean correlation of R=0.83 with MDEV and liver stiffness estimates within 7% of 2D-LFE results. Finally, ESP showed statistically significant increase in fine feature spectral energy as measured with RER for both |G^*| (p<1×10^-9) and ϕ (p<1×10^-3).

CONCLUSION

Information at finer frequencies can be recovered in ESP elastograms in typical experimental conditions, however scatter- and boundary-related artefacts may cause the fine features to have inaccurate values. In in vivo cohorts, ESP delivers an increase in fine feature spectral energy, and better performance with longer wavelengths, than MDEV while showing similar stability and robustness.

Tomoelastography by multifrequency wave number recovery from time-harmonic propagating shear waves

Palpation is one of the most sensitive, effective diagnostic practices, motivating the quantitative and spatially resolved determination of soft tissue elasticity parameters by medical ultrasound or MRI. However, this so-called elastography often suffers from limited anatomical resolution due to noise and insufficient elastic deformation, currently precluding its use as a tomographic modality on its own. We here introduce an efficient way of processing wave images acquired by multifrequency magnetic resonance elastography (MMRE), which relies on wave number reconstruction at different harmonic frequencies followed by their amplitude-weighted averaging prior to inversion. This results in compound maps of wave speed, which reveal variations in tissue elasticity in a tomographic fashion, i.e. an unmasked, slice-wise display of anatomical details at pixel-wise resolution. The method is demonstrated using MMRE data from the literature including abdominal and pelvic organs such as the liver, spleen, uterus body and uterus cervix. Even in small regions with low wave amplitudes, such as nucleus pulposus and spinal cord, elastic parameters consistent with literature values were obtained. Overall, the proposed method provides a simple and noise-robust strategy of in-plane wave analysis of MMRE data, with a pixel-wise resolution producing superior detail to MRE direct inversion methods.

Real-time 4D phase unwrapping applied to magnetic resonance elastography

PURPOSE

Phase amplitude is a source of signal in magnetic resonance elastography (MRE) experiments but its exploitation in experimental design has been limited due to the challenges of phase wrap. This study addressed this aspect of MRE through new developments in algorithms, heuristic strategy, and user interface.

METHODS

A test dataset with systematic variation of three parameters-nested wrap, gradient, and noise level-was developed to choose phase-unwrapping algorithms and to analyze their performance. A new application, PhaseTools, was developed that implemented three phase-unwrapping algorithms that adhere to a "real-time" criterion of less than 3 min for a four-dimensional MRE acquisition. Two of the algorithms extend previously published algorithms and one was newly developed. The algorithms were then applied to five datasets from MRE, two typical cases and three edge cases that were particularly challenging in one of the three parameters.

RESULTS

The performance of the PhaseTools algorithms on the test dataset was comparable to two widely cited algorithms that take hours or days to complete. Guidelines for the optimal use of each algorithm are established.

CONCLUSION

PhaseTools enables the substantial increase of signal-to-noise in MRE experiments at negligible additional computational cost. PhaseTools is freely released with this study, making robust real-time phase unwrapping available to any group using phase-based imaging.