What is the clinical evidence from this study?

Study design: Cohort. Population: Pulmonary embolism (n=21183). Intervention: Fusion machine learning model (clinical data + ECG waveform) vs. ECG-only model, EHR-only model, and standard clinical scores. Primary outcome: Pulmonary embolism likelihood (AUROC).

November 25, 2021Open Access

Development of a machine learning model using electrocardiogram signals to improve acute pulmonary embolism screening

Key Result

A Fusion machine learning model integrating clinical data and ECG waveforms outperformed an EHR-only model for predicting pulmonary embolism likelihood (AUROC 0.81 vs 0.65).

Study Design

Type

Cohort (n=21,183)

Structured PICO

Does a Fusion machine learning model integrating ECG waveforms and clinical data improve the specificity and overall performance of acute pulmonary embolism screening compared to traditional clinical scores?

Population

21,183 patients at moderate- to high suspicion of pulmonary embolism (PE), associated with 23,793 CTPAs (10.0% PE-positive) and 320,746 ECGs and encounter-level clinical data.

Intervention

Fusion machine learning model integrating clinical data and an embedded representation of the ECG waveform to predict PE likelihood.

Comparator

ECG model alone, EHR model alone, and four clinical scores (Wells' Criteria, Revised Geneva Score, Pulmonary Embolism Rule-Out Criteria, and 4-Level Pulmonary Embolism Clinical Probability Score).

Outcome

Prediction of pulmonary embolism likelihood measured by area under the receiver-operating characteristic curve (AUROC) and specificity.surrogate

A machine learning model combining ECG waveforms with clinical data significantly improves the specificity and overall performance of acute pulmonary embolism screening compared to traditional clinical scores.

Main Result

Absolute Event Rate: 0.81% vs 0.65%

Abstract

Aims: Clinical scoring systems for pulmonary embolism (PE) screening have low specificity and contribute to computed tomography pulmonary angiogram (CTPA) overuse. We assessed whether deep learning models using an existing and routinely collected data modality, electrocardiogram (ECG) waveforms, can increase specificity for PE detection. Methods and results: We create a retrospective cohort of 21 183 patients at moderate- to high suspicion of PE and associate 23 793 CTPAs (10.0% PE-positive) with 320 746 ECGs and encounter-level clinical data (demographics, comorbidities, vital signs, and labs). We develop three machine learning models to predict PE likelihood: an ECG model using only ECG waveform data, an EHR model using tabular clinical data, and a Fusion model integrating clinical data and an embedded representation of the ECG waveform. We find that a Fusion model area under the receiver-operating characteristic curve (AUROC) 0.81 ± 0.01 outperforms both the ECG model (AUROC 0.59 ± 0.01) and EHR model (AUROC 0.65 ± 0.01). On a sample of 100 patients from the test set, the Fusion model also achieves greater specificity (0.18) and performance (AUROC 0.84 ± 0.01) than four commonly evaluated clinical scores: Wells' Criteria, Revised Geneva Score, Pulmonary Embolism Rule-Out Criteria, and 4-Level Pulmonary Embolism Clinical Probability Score (AUROC 0.50-0.58, specificity 0.00-0.05). The model is superior to these scores on feature sensitivity analyses (AUROC 0.66-0.84) and achieves comparable performance across sex (AUROC 0.81) and racial/ethnic (AUROC 0.77-0.84) subgroups. Conclusion: Synergistic deep learning of ECG waveforms with traditional clinical variables can increase the specificity of PE detection in patients at least at moderate suspicion for PE.