Deep Metric Learning for the Hemodynamics Inference with Electrocardiogram Signals

Hyewon Jeong,Collin M. Stultz,Marzyeh Ghassemi

2023-09-11

Abstract:Heart failure is a debilitating condition that affects millions of people worldwide and has a significant impact on their quality of life and mortality rates. An objective assessment of cardiac pressures remains an important method for the diagnosis and treatment prognostication for patients with heart failure. Although cardiac catheterization is the gold standard for estimating central hemodynamic pressures, it is an invasive procedure that carries inherent risks, making it a potentially dangerous procedure for some patients. Approaches that leverage non-invasive signals - such as electrocardiogram (ECG) - have the promise to make the routine estimation of cardiac pressures feasible in both inpatient and outpatient settings. Prior models trained to estimate intracardiac pressures (e.g., mean pulmonary capillary wedge pressure (mPCWP)) in a supervised fashion have shown good discriminatory ability but have been limited to the labeled dataset from the heart failure cohort. To address this issue and build a robust representation, we apply deep metric learning (DML) and propose a novel self-supervised DML with distance-based mining that improves the performance of a model with limited labels. We use a dataset that contains over 5.4 million ECGs without concomitant central pressure labels to pre-train a self-supervised DML model which showed improved classification of elevated mPCWP compared to self-supervised contrastive baselines. Additionally, the supervised DML model that uses ECGs with access to 8,172 mPCWP labels demonstrated significantly better performance on the mPCWP regression task compared to the supervised baseline. Moreover, our data suggest that DML yields models that are performant across patient subgroups, even when some patient subgroups are under-represented in the dataset. Our code is available at <a class="link-external link-https" href="https://github.com/mandiehyewon/ssldml" rel="external noopener nofollow">this https URL</a>

Machine Learning,Signal Processing,Quantitative Methods

What problem does this paper attempt to address?

### Problems Addressed by the Paper

This paper aims to address the issue of non-invasive estimation of cardiac pressure (such as mean pulmonary capillary wedge pressure, mPCWP) in heart failure patients. Specifically, the authors attempt to use electrocardiogram (ECG) signals to infer cardiac hemodynamic parameters as an alternative to the traditional invasive method—right heart catheterization.

#### Background and Challenges

1. **Impact of Heart Failure**:
- Heart failure is a serious condition affecting millions of people worldwide, significantly impacting patients' quality of life and mortality rates.
- Accurate assessment of cardiac pressure is crucial for the diagnosis and treatment of heart failure.

2. **Limitations of Traditional Methods**:
- Right heart catheterization is currently the gold standard for assessing central hemodynamic pressures, but it is invasive, carries inherent risks, and is not suitable for all patients.
- Obtaining large datasets with cardiac pressure labels is very challenging because it requires invasive procedures.

3. **Shortcomings of Existing Models**:
- Supervised learning models have shown good discriminative ability in estimating cardiac pressure, but these models rely on limited labeled datasets and perform inconsistently across different patient populations.
- Self-supervised contrastive learning methods perform poorly in regression tasks and exhibit poor predictive fairness across different gender groups.

#### Solution

1. **Application of Deep Metric Learning (DML)**:
- The authors propose a deep metric learning-based approach, leveraging large-scale unlabeled ECG datasets for pre-training, followed by fine-tuning on labeled datasets to improve model performance.
- They introduce a novel self-supervised DML method, enhancing model performance through distance-based mining techniques.

2. **Specific Methods**:
- **Supervised DML**: Joint training using labeled ECG datasets, optimizing for both classification and regression tasks.
- **Self-supervised DML**: Pre-training using unlabeled ECG datasets, followed by fine-tuning on labeled datasets.

3. **Experimental Results**:
- The supervised DML model showed significant performance improvement in the mPCWP regression task and achieved competitive results in classification tasks.
- The self-supervised DML model performed well in regression tasks, comparable to random initialization and baseline models, demonstrating the advantage of utilizing unlabeled datasets.
- These models exhibited more equitable performance across different gender and age subgroups, reducing performance disparities among subgroups.

### Summary

By introducing deep metric learning methods, particularly self-supervised DML, this paper successfully improves the accuracy of estimating cardiac pressure using ECG signals, especially in data-limited scenarios. This approach not only offers technical innovation but also provides a safer, more reliable non-invasive tool for clinical applications.