Echocardiographic Detection of Regional Wall Motion Abnormalities Using Artificial Intelligence Compared to Human Readers

BACKGROUND

Although regional wall motion abnormality (RWMA) detection is foundational to transthoracic echocardiography, current methods are prone to interobserver variability. We aimed to develop a deep learning (DL) model for RWMA assessment and compare it to expert and novice readers.

METHODS

We used 15,746 transthoracic echocardiography studies-including 25,529 apical videos-which were split into training, validation, and test datasets. A convolutional neural network was trained and validated using apical 2-, 3-, and 4-chamber videos to predict the presence of RWMA in 7 regions defined by coronary perfusion territories, using the ground truth derived from clinical transthoracic echocardiography reports. Within the test cohort, DL model accuracy was compared to 6 expert and 3 novice readers using F1 score evaluation, with the ground truth of RWMA defined by expert readers. Significance between the DL model and novices was assessed using the permutation test.

RESULTS

Within the test cohort, the DL model accurately identified any RWMA with an area under the curve of 0.96 (0.92-0.98). The mean F1 scores of the experts and the DL model were numerically similar for 6 of 7 regions: anterior (86 vs 84), anterolateral (80 vs 74), inferolateral (83 vs 87), inferoseptal (86 vs 86), apical (88 vs 87), inferior (79 vs 81), and any RWMA (90 vs 94), respectively, while in the anteroseptal region, the F1 score of the DL model was lower than the experts (75 vs 89). Using F1 scores, the DL model outperformed both novices 1 (P = .002) and 2 (P = .02) for the detection of any RWMA.

CONCLUSIONS

Deep learning provides accurate detection of RWMA, which was comparable to experts and outperformed a majority of novices. Deep learning may improve the efficiency of RWMA assessment and serve as a teaching tool for novices.

Deep learning based classification of left ventricular function from two-dimensional echocardiographic images

Funding Acknowledgements

Type of funding sources: Private company. Main funding source(s): Philips Healthcare

Background. Echocardiographic quantification of left ventricular function (LVF) is mainly based on ejection fraction (EF) measurements, which relies on either manual or automated identification of endocardial boundaries followed by calculation of model-based end-systolic and end-diastolic LV volumes. Recent developments in artificial intelligence resulted in computer algorithms that allow fully automated detection of endocardial boundaries and measurement of LV volumes and EF. However, this methodology is prone to errors and inter-measurement variability. We hypothesized that a fully automated deep learning algorithm could be developed, which would accurately classify LVF while avoiding volume and EF measurements. This study was designed to test the accuracy of this approach.

Methods. Deep learning algorithm was developed (Philips Research) based on convolutional neural network (CNN) that uses as input dynamic sequences of apical 2- and 4-chamber echocardiographic views obtained without ultrasound enhancing agents. We used for CNN development a database of clinical DICOM studies: a training set of 14,427 studies with normal LV function and 6,135 abnormal, and a validation set of 2,898 normal and 1,081 abnormal studies, based on Philips IntelliSpace Cardiovascular (ISCV) codes found (defined by cardiologists) in the patients’ reports. The CNN was trained to automatically classify LVF into 3 categories: (1) normal, (2) mildly-to-moderately or moderately reduced, and (3) moderately-to-severely or severely reduced. In the validation set, the automated classifications were compared to those in the patients’ reports as a reference standard. Accuracy of the automated classification was tested using contingency tables, from which sensitivity, specificity, and negative and positive predictive values (NPV, PPV) and overall accuracy were calculated for each category of LVF. Additionally, the area under ROC curve (AUC) was calculated to assess the diagnostic accuracy of the automated classification for each LVF category.

Results.  Automated classification of LVF showed high levels of diagnostic accuracy in identifying cases with LVF in all 3 categories, reflected by high AUC values: (1) 0.94, (2) 0.87 and (3) 0.97 (Figure), and overall accuracy of 0.84 (Table). 

Conclusions. Deep learning algorithm based on CNN allowed accurate automated classification of LVF, when tested on ∼4,000 clinical studies and compared to ISCV codes found in the patients’ reports. This novel fully-automated methodology may become a useful aid in the interpretation of echocardiographic images by providing the reader with a preliminary assessment of LVF. Abstract Figure.

Test of Lepton-Flavor Universality in B→K^{*}ℓ^{+}ℓ^{-} Decays at Belle

We present a measurement of R_{K^{*}}, the branching fraction ratio B(B→K^{*}μ^{+}μ^{-})/B(B→K^{*}e^{+}e^{-}), for both charged and neutral B mesons. The ratio for the charged case R_{K^{*+}} is the first measurement ever performed. In addition, we report absolute branching fractions for the individual modes in bins of the squared dilepton invariant mass q^{2}. The analysis is based on a data sample of 711  fb^{-1}, containing 772×10^{6}  BB[over ¯] events, recorded at the ϒ(4S) resonance with the Belle detector at the KEKB asymmetric-energy e^{+}e^{-} collider. The obtained results are consistent with standard model expectations.

Search for Axionlike Particles Produced in e^{+}e^{-} Collisions at Belle II

We present a search for the direct production of a light pseudoscalar a decaying into two photons with the Belle II detector at the SuperKEKB collider. We search for the process e^{+}e^{-}→γa, a→γγ in the mass range 0.2<m_{a}<9.7  GeV/c^{2} using data corresponding to an integrated luminosity of (445±3)  pb^{-1}. Light pseudoscalars interacting predominantly with standard model gauge bosons (so-called axionlike particles or ALPs) are frequently postulated in extensions of the standard model. We find no evidence for ALPs and set 95% confidence level upper limits on the coupling strength g_{aγγ} of ALPs to photons at the level of 10^{-3}  GeV^{-1}. The limits are the most restrictive to date for 0.2<m_{a}<1  GeV/c^{2}.

Search for an Invisibly Decaying Z^{'} Boson at Belle II in e^{+}e^{-}→μ^{+}μ^{-}(e^{±}μ^{∓}) Plus Missing Energy Final States

Theories beyond the standard model often predict the existence of an additional neutral boson, the Z^{'}. Using data collected by the Belle II experiment during 2018 at the SuperKEKB collider, we perform the first searches for the invisible decay of a Z^{'} in the process e^{+}e^{-}→μ^{+}μ^{-}Z^{'} and of a lepton-flavor-violating Z^{'} in e^{+}e^{-}→e^{±}μ^{∓}Z^{'}. We do not find any excess of events and set 90% credibility level upper limits on the cross sections of these processes. We translate the former, in the framework of an L_{μ}-L_{τ} theory, into upper limits on the Z^{'} coupling constant at the level of 5×10^{-2}-1 for M_{Z^{'}}≤6  GeV/c^{2}.

Lepton-Flavor-Dependent Angular Analysis of B→K^{*}ℓ^{+}ℓ^{-}

We present a measurement of angular observables and a test of lepton flavor universality in the B→K^{*}ℓ^{+}ℓ^{-} decay, where ℓ is either e or μ. The analysis is performed on a data sample corresponding to an integrated luminosity of 711  fb^{-1} containing 772×10^{6} BB[over ¯] pairs, collected at the ϒ(4S) resonance with the Belle detector at the asymmetric-energy e^{+}e^{-} collider KEKB. The result is consistent with standard model (SM) expectations, where the largest discrepancy from a SM prediction is observed in the muon modes with a local significance of 2.6σ.

Observation of e+e- → π+π-π(0)(χbJ) and Search for X(b) → ωϒ(1S) at sqrt[s] = 10.867 GeV

The e(+)e(-) → π(+)π(-)π(0)χ(bJ) (J = 0,1,2) processes are studied using a 118 fb(-1) data sample acquired with the Belle detector at a center-of-mass energy of 10.867 GeV. Unambiguous π(+)π(-)π(0)χ(bJ) (J = 1,2), ωχ(b1) signals are observed, and indication for ωχ(b2) is seen, both for the first time, and the corresponding cross section measurements are presented. No significant π(+)π(-)π(0)χ(b0) or ωχ(b0) signals are observed, and 90% confidence level upper limits on the cross sections for these two processes are obtained. In the π(+)π(-)π(0) invariant mass spectrum, significant non-ω signals are also observed. We search for the X(3872)-like state (named X(b)) decaying into ωϒ(1S); no significant signal is observed with a mass between 10.55 and 10.65 GeV/c(2).