Non-Invasive Biosensing for Healthcare Using Artificial Intelligence: A Semi-Systematic Review

The rapid development of biosensing technologies together with the advent of deep learning has marked an era in healthcare and biomedical research where widespread devices like smartphones, smartwatches, and health-specific technologies have the potential to facilitate remote and accessible diagnosis, monitoring, and adaptive therapy in a naturalistic environment. This systematic review focuses on the impact of combining multiple biosensing techniques with deep learning algorithms and the application of these models to healthcare. We explore the key areas that researchers and engineers must consider when developing a deep learning model for biosensing: the data modality, the model architecture, and the real-world use case for the model. We also discuss key ongoing challenges and potential future directions for research in this field. We aim to provide useful insights for researchers who seek to use intelligent biosensing to advance precision healthcare.

Effect of Mobile Internet on Attitude and Self-Efficacy of Patients with Coronary Heart Disease Diagnosed by 12-Lead Holter ECG

Objective

To explore the effect of mobile Internet on attitude and self-efficacy of patients with coronary heart disease (CHD) diagnosed by 12-lead Holter ECG.

Methods

The clinical data of 62 patients with CHD who underwent routine ECG examination (control group I) and 12-lead dynamic electrocardiogram (control group II) in our hospital (June 2017–December 2020) were retrospectively analyzed, and the clinical data of another 62 patients with CHD who received 12-lead Holter ECG examination combined with mobile Internet in our hospital at the same time (study group) were retrospectively analyzed. The clinical observation indexes of the three groups were compared.

Results

No obvious difference in general data among groups (P > 0.05). Compared with the control group I, the positive detection rate (PDR) of the study group and the control group II was obviously higher (P < 0.05), and the PDR of the study group was obviously higher than that of the control group II, without remarkable difference between both groups (P > 0.05). Compared with the control group, the scores of CAS-R of the study group were obviously higher (P < 0.05), and self-efficacy of daily life, health behaviors, medication compliance, and compliance behavior of the study group was obviously better (P < 0.05). The diagnostic efficacy was derived by ROC curve analysis, 12-lead Holter ECG combined with mobile Internet + routine ECG > 12-lead Holter ECG combined with mobile Internet > 12-lead Holter ECG > routine ECG.

Conclusion

Compared with the routine ECG, the sensitivity of 12-lead Holter ECG in the diagnosis of CHD is conspicuously higher. Meanwhile, 12-lead Holter ECG combined with mobile Internet can enhance the diagnostic efficiency and improve patients' perceived control attitude and self-efficacy.

Electrocardiogram generation with a bidirectional LSTM-CNN generative adversarial network

Heart disease is a malignant threat to human health. Electrocardiogram (ECG) tests are used to help diagnose heart disease by recording the heart’s activity. However, automated medical-aided diagnosis with computers usually requires a large volume of labeled clinical data without patients' privacy to train the model, which is an empirical problem that still needs to be solved. To address this problem, we propose a generative adversarial network (GAN), which is composed of a bidirectional long short-term memory(LSTM) and convolutional neural network(CNN), referred as BiLSTM-CNN,to generate synthetic ECG data that agree with existing clinical data so that the features of patients with heart disease can be retained. The model includes a generator and a discriminator, where the generator employs the two layers of the BiLSTM networks and the discriminator is based on convolutional neural networks. The 48 ECG records from individuals of the MIT-BIH database were used to train the model. We compared the performance of our model with two other generative models, the recurrent neural network autoencoder(RNN-AE) and the recurrent neural network variational autoencoder (RNN-VAE). The results showed that the loss function of our model converged to zero the fastest. We also evaluated the loss of the discriminator of GANs with different combinations of generator and discriminator. The results indicated that BiLSTM-CNN GAN could generate ECG data with high morphological similarity to real ECG recordings.

Comparative TQT analysis with three fully-automated platforms: comparison to core laboratory semi-automated results

BACKGROUND AND PURPOSE

Technological advances in machine-read QT measurement now enable detailed and precise cardiac repolarization assessments. This study assessed the applicability of three state-of-art ECG measurement applications to provide reliable continuous analyses from data obtained in a positive thorough QT study previously characterized with sparse semi-automated measurements performed by an ECG core laboratory.

METHODS

Continuous RR, QT, QTc measurements, and individual QT/RR relationships and their associated intra- and inter-subject variability were derived in parallel with BioQT, Ponemah PRO, and WinAtrec analysis software.

RESULTS

Despite slight vendor-specific differences in measurement variability and QTc, all machine-read methods demonstrated requisite assay sensitivity and yielded similar conclusions in accordance with SA analysis.

CONCLUSIONS

Three commercially available ECG analytical software applications reliably detected the drug-induced QT prolonging effects and replicated the SA core-laboratory conclusions, with greatly improved temporal resolution and reduced analytical costs. With broader experience, these data suggest that current SA methodologies could be effectively replaced by fully automated ECG analysis.