Machine learning in cardiovascular flows modeling: Predicting arterial blood pressure from non-invasive 4D flow MRI data using physics-informed neural networks

AbstractContinuous vital signal monitoring is becoming more relevant in preventing diseases that afflict a large part of the world’s population; for this reason, healthcare equipment should be easy to wear and simple to use. Non-intrusive and non-invasive detection methods are a basic requirement for wearable medical devices, especially when these are used in sports applications or by the elderly for self-monitoring. Arterial blood pressure (ABP) is an essential physiological parameter for health monitoring. Most blood pressure measurement devices determine the systolic and diastolic arterial blood pressure through the inflation and the deflation of a cuff. This technique is uncomfortable for the user and may result in anxiety, and consequently affect the blood pressure and its measurement. The purpose of this paper is the continuous measurement of the ABP through a cuffless, non-intrusive approach. The approach of this paper is based on deep learning techniques where several neural networks are used to infer ABP, starting from photoplethysmogram (PPG) and electrocardiogram (ECG) signals. The ABP was predicted first by utilizing only PPG and then by using both PPG and ECG. Convolutional neural networks (ResNet and WaveNet) and recurrent neural networks (LSTM) were compared and analyzed for the regression task. Results show that the use of the ECG has resulted in improved performance for every proposed configuration. The best performing configuration was obtained with a ResNet followed by three LSTM layers: this led to a mean absolute error (MAE) of 4.118 mmHg on and 2.228 mmHg on systolic and diastolic blood pressures, respectively. The results comply with the American National Standards of the Association for the Advancement of Medical Instrumentation. ECG, PPG, and ABP measurements were extracted from the MIMIC database, which contains clinical signal data reflecting real measurements. The results were validated on a custom dataset created at Neuronica Lab, Politecnico di Torino.

Download Full-text

Continuous cardiac output monitoring in humans by non-invasive arterial blood pressure waveform analysis

Computers in Cardiology, 2005 ◽

10.1109/cic.2005.1588236 ◽

2005 ◽

Author(s):

Z. Lu ◽

R. Mukkamala

Keyword(s):

Blood Pressure ◽

Cardiac Output ◽

Arterial Blood Pressure ◽

Waveform Analysis ◽

Cardiac Output Monitoring ◽

Pressure Waveform ◽

Non Invasive ◽

Arterial Blood ◽

Blood Pressure Waveform ◽

Output Monitoring

Download Full-text

Analysis for the non-invasive determination of arterial properties and for the transcutaneous continuous monitoring of arterial blood pressure

Medical & Biological Engineering & Computing ◽

10.1007/bf02442452 ◽

1978 ◽

Vol 16 (6) ◽

pp. 715-726 ◽

Cited By ~ 7

Author(s):

Dhanjoo N. Ghista ◽

G. Jayaraman ◽

H. Sandler

Keyword(s):

Blood Pressure ◽

Arterial Blood Pressure ◽

Continuous Monitoring ◽

Non Invasive ◽

Arterial Blood

Download Full-text

Non - invasive right ventricular efficiency using 4D flow MRI

Journal of Cardiovascular Magnetic Resonance ◽

10.1186/1532-429x-17-s1-q58 ◽

2015 ◽

Vol 17 (S1) ◽

Author(s):

Alejandro Roldán-Alzate ◽

Scott W Grogan ◽

Heidi B Kellihan ◽

Alessandro Bellofiore ◽

Naomi C Chesler ◽

...

Keyword(s):

Right Ventricular ◽

4D Flow Mri ◽

4D Flow ◽

Non Invasive ◽

Flow Mri ◽

Ventricular Efficiency

Download Full-text

NON-INVASIVE MODELING OF INTRACRANIAL HYPERTENSION FROM PHYSIOLOGICAL CHANNELS

10.32920/ryerson.14654556 ◽

2021 ◽

Author(s):

Parisa Naraei ◽

Alireza Sadeghian

Keyword(s):

Machine Learning ◽

Blood Pressure ◽

Intracranial Hypertension ◽

Clinical Decision Making ◽

Machine Learning Techniques ◽

Non Invasive ◽

Arterial Blood ◽

Secondary Damage ◽

Invasive Techniques ◽

Present Thesis

Intracranial pressure (ICP), the pressure within the cranium reflects three elements: cerebrospinal fluid, brain tissue and blood pressure. High ICP (above 20 mmHg) is called intracranial hypertension (ICH) which is due to the tumour, swelling or the internal bleeding of brain and may cause secondary damage to the brain. ICP is a crucial parameter in diagnosis of brain injuries. Two models which utilize machine learning techniques to anticipate ICH and assist in clinical decision making were developed in the present thesis. ICP can be monitored through the invasive techniques (i.e., inserting an intraventricular catheter through the skull). Despite the high accuracy, the episodes of ICH can also be manually identified only after placement of catheter which is accompanied by lots of technical difficulties. Furthermore, the ICP signal might not be available continuously or may include unwanted noise that could introduce more complication to the diagnosis and treatment procedure. Considering the difficulties of the invasive techniques, a non-invasive model, capable to predict the ICH helps to save time, estimate the missing ICPs, predict the ICP in advance and accelerate medical intervention. The present thesis introduces two machine learning models to resolve the current limitations: 1- Non-invasive prediction of ICP labels 10 minutes in advance where the status of ICP (normal / ICH) is predicted based on the two components extracted from the physiological signals such as mean arterial blood pressure and respiration rate. 2- Wavelet – clustering where a machine learning solution for ICP estimation using a hybrid wavelet clustering is proposed. The episodes of ICP and derived from ICP (such as cerebral perfusion pressure) are excluded from the second model.

Download Full-text