The ability to non-invasively measure left atrial pressure would facilitate the identification of patients at risk of pulmonary congestion and guide proactive heart failure care. Wearable cardiac monitors, which record single-lead electrocardiogram data, provide information that can be leveraged to infer left atrial pressures. We developed a deep neural network using single-lead electrocardiogram data to determine when the left atrial pressure is elevated. The model was developed and internally evaluated using a cohort of 6739 samples from the Massachusetts General Hospital (MGH) and externally validated on a cohort of 4620 samples from a second institution. We then evaluated model on patch-monitor electrocardiographic data on a small prospective cohort. The model achieves an area under the receiver operating characteristic curve of 0.80 for detecting elevated left atrial pressures on an internal holdout dataset from MGH and 0.76 on an external validation set from a second institution. A further prospective dataset was obtained using single-lead electrocardiogram data with a patch-monitor from patients who underwent right heart catheterization at MGH. Evaluation of the model on this dataset yielded an area under the receiver operating characteristic curve of 0.875 for identifying elevated left atrial pressures for electrocardiogram signals acquired close to the time of the right heart catheterization procedure. These results demonstrate the utility and the potential of ambulatory cardiac hemodynamic monitoring with electrocardiogram patch-monitors. Heart failure is a common disorder that is challenging to manage. Appearance of symptoms can be subtle and dangerous and there are few tools for clinicians to estimate when a patient is likely to experience an episode of heart failure. Current methods to detect elevated pressure in the heart (one sign of oncoming failure) are invasive and can only be performed in an inpatient setting. A non-invasive, quick method for detecting higher heart pressure would be helpful for identifying worsening heart failure in the home environment. For this reason, we developed a computer method to detect elevated pressures inside the heart using a non-invasive signal from a wearable patch monitor device, the electrocardiogram (ECG, or EKG). Our results show our method provides a reliable, non-invasive way to measure heart pressures using data that can be obtained in the outpatient setting. Schlesinger and Alam et al. utilize a deep neural network and single-lead electrocardiogram data to determine elevated left atrial pressure in patients. This work aims to identify patients at risk of pulmonary congestion and guide proactive heart failure care.

AbstractQuantifying hemodynamic severity in patients with heart failure (HF) is an integral part of clinical care. A key indicator of hemodynamic severity is the mean Pulmonary Capillary Wedge Pressure (mPCWP), which is ideally measured invasively. Accurate non-invasive estimates of the mPCWP in patients with heart failure would help identify individuals at the greatest risk of a HF exacerbation. We developed a deep learning model, HFNet, that uses the 12-lead electrocardiogram (ECG) together with age and sex to identify when the mPCWP > 18 mmHg in patients who have a prior diagnosis of HF. The model was developed using retrospective data from the Massachusetts General Hospital and evaluated on both an internal test set and an independent external validation set, from another institution. We developed an uncertainty score that identifies when model performance is likely to be poor, thereby helping clinicians gauge when to trust a given model prediction. HFNet AUROC for the task of estimating mPCWP > 18 mmHg was 0.8 $$2\pm$$ 2 ± 0.01 and 0.$$81\pm$$ 81 ± 0.01 on the internal and external datasets, respectively. The AUROC on predictions with the highest uncertainty are 0.50 $$\pm$$ ± 0.02 (internal) and 0.$$56\pm$$ 56 ± 0.04 (external), while the AUROC on predictions with the lowest uncertainty were 0.86 ± 0.01 (internal) and 0.82 ± 0.01 (external). Using estimates of the prevalence of mPCWP > 18 mmHg in patients with reduced ventricular function, and a decision threshold corresponding to an 80% sensitivity, the calculated positive predictive value (PPV) is 0.$$89\pm$$ 89 ± 0.01when the corresponding chest x-ray (CXR) is consistent with interstitial edema HF. When the CXR is not consistent with interstitial edema, the estimated PPV is 0.$$78\pm$$ 78 ± 0.02, again at an 80% sensitivity threshold. HFNet can accurately predict elevated mPCWP in patients with HF using the 12-lead ECG and age/sex. The method also identifies cohorts in which the model is more/less likely to produce accurate outputs.

Background: Central hemodynamic parameters are typically measured via pulmonary artery catherization-an invasive procedure that involves some risk to the patient and is not routinely available in all settings. Objectives: This study sought to develop a noninvasive method to identify elevated mean pulmonary capillary wedge pressure (mPCWP). Methods: We leveraged data from 248,955 clinical records at the Massachusetts General Hospital to develop a deep learning model that can infer when the mPCWP >15 mmHg using the 12-lead electrocardiogram (ECG). Of these data, 242,216 records were used to pre-train a model that generates useful ECG representations. The remaining 6,739 records contain encounters with direct measurements of the mPCWP. Eighty percent of these data were used for model development and testing (5,390), and the remaining records comprise a holdout set (1,349) that was used to evaluate the model. We developed an associated unreliability score that identifies when model predictions are likely to be untrustworthy. Results: The model achieves an area under the receiver operating characteristic curve (AUC) of 0.80 +/- 0.02 (test set) and 0.79 +/- 0.01 (holdout set). Model performance varies as a function of the unreliability, where patients with high unreliability scores correspond to a subgroup where model performance is poor: for example, patients in the holdout set with unreliability scores in the highest decile have a reduced AUC of 0.70 +/- 0.06. Conclusions: The mPCWP can be inferred from the ECG, and the reliability of this inference can be measured. When invasive monitoring cannot be expeditiously performed, deep learning models may provide information that can inform clinical care.

Purpose of review Although deep learning represents an exciting platform for the development of risk stratification models, it is challenging to evaluate these models beyond simple statistical measures of success, which do not always provide insight into a model’s clinical utility. Here we propose a framework for evaluating deep learning models and discuss a number of interesting applications in light of these rubrics. Recent findings Data scientists and clinicians alike have applied a variety of deep learning techniques to both medical images and structured electronic medical record data. In many cases, these methods have resulted in risk stratification models that have improved discriminatory ability relative to more straightforward methods. Nevertheless, in many instances, it remains unclear how useful the resulting models are to practicing clinicians. Summary To be useful, deep learning models for cardiovascular risk stratification must not only be accurate but they must also provide insight into when they are likely to yield inaccurate results and be explainable in the sense that health care providers can understand why the model arrives at a particular result. These additional criteria help to ensure that the model can be faithfully applied to the demographic for which it is most accurate.

Triggering shape-memory functionality under clinical hyperthermia temperatures could enable the control and actuation of shape-memory systems in clinical practice. For this purpose, we developed light-inducible shape-memory microparticles composed of a poly(d,l-lactic acid) (PDLLA) matrix encapsulating gold nanoparticles (Au@PDLLA hybrid microparticles). This shape-memory polymeric system for the first time demonstrates the capability of maintaining an anisotropic shape at body temperature with triggered shape-memory effect back to a spherical shape at a narrow temperature range above body temperature with a proper shape recovery speed (37 < T < 45 °C). We applied a modified film-stretching processing method with carefully controlled stretching temperature to enable shape memory and anisotropy in these micron-sized particles. Accordingly, we achieved purely entanglement-based shape-memory response without chemical cross-links in the miniaturized shape-memory system. Furthermore, these shape-memory microparticles exhibited light-induced spatiotemporal control of their shape recovery using a laser to trigger the photothermal heating of doped gold nanoparticles. This shape-memory system is composed of biocompatible components and exhibits spatiotemporal controllability of its properties, demonstrating a potential for various biomedical applications, such as tuning macrophage phagocytosis as demonstrated in this study.