Each year, about 5-18% of babies are born preterm, accounting for over 0.5M births in the US and 15M globally. Many of these babies are admitted to Neonatal Intensive Care Units (NICUs) where the medical staff generally have the option of using either invasive arterial lines (IALs) or inflatable-cuff non-invasive blood pressure (NIBP) monitoring. The former introduces the risk of infection, tissue and nerve damage, and the latter is less accurate, especially for hypotensive infants, and may add the risk of ischemic and nerve damage upon repeated measurement. There is a clear need for a safer, continuous, and cost effective form of NIBP measurement to meet the challenge of managing unhealthy blood pressures for neonates. PyrAmes Inc. has developed a novel capacitive sensor technology that is paper thin and flexible and can accurately detect blood pressure (BP). This sensor technology is part of a unique continuous BP monitoring platform that provides accurate, lightweight and comfortable BP monitoring in a wireless, wrist-worn package that is easy to use. The system uses lightweight neural networks to analyze pulse waveform data to provide continuous determination of systolic, diastolic, and mean BP, heart rate, and their variabilities. The sensor is easy to apply non-invasively and records pulsatile data similar to an arterial line, while avoiding the difficulties of placing and maintaining an arterial line. This device can provide gold standard BP monitoring without perturbing the patients for more accurate and relevant measurements. The objective of this project is to extend the platform for use with term and pre-term neonates. This goal will be accomplished through redesign of the sensor hardware and optimization of the data analytics software. We will validate these modifications with clinical data from the NICU at Stanford University Medical Center. In Phase I, we will miniaturize the electronics and modify the sensor array of a wrist-worn pulse wave monitor to be sized more appropriately for neonates. We will validate the new device design by collecting NICU clinical data from patients who have IALs in place in an IRB-approved study. From IAL and sensor data taken simultaneously, we will determine ground truth values on a pulse-by-pulse basis and use these data in conjunction with additional IAL data from historical databases to improve our sensor quality and predictive BP models. Our success metric will be to equal or exceed the quality and accuracy of our data for adults. Phase II will be a follow-up IRB-approved pivotal study using the device from Phase 1 to position our device for FDA submission and clearance and scale up to pilot production of this device.
The goal of this Phase 1 project is to confirm that machine learning can be used to extract blood pressure values for critically ill neonates from pulse waveform data collected with a wearable, non-invasive device that is comfortable, low-cost and easy to use. The proposed device will significantly reduce the need for frequent cuff-based measurements and/or invasive arterial lines, thereby decreasing morbidity, risk of complications, patient discomfort, and overall cost of care. This project is based on the pioneering efforts of Prof. Zhenan Bao?s lab at Stanford on thin film sensors for electronic skin and includes the design and use of a miniaturized device to collect clinical data from neonates that will be used to validate a model which derives blood pressure values from the pulse waveform without external calibration.
