• Login
• Admin

Stockwell, Simon John (2024) Improvement of reflection-mode pulse oximetry for measurement of SpO2 in newborns. PhD thesis, University of Nottingham.

PDF (Thesis - as examined) - Repository staff only until 18 July 2026. Subsequently available to Anyone - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader Available under Licence Creative Commons Attribution. Download (27MB)

Abstract

Measurement of the photoplethysmogram in the immediate moments after birth is a historically challenging setting to obtain rapid and accurate data in these critical early minutes. Motion artefacts, low perfusion and noisy signals all contribute to an inability to measure the photoplethysmogram to derive the necessary vital signs to show improvement in the condition of the newborn. Previous clinical trials from 2017 on infants born by elective caesarean section showed an ability to accurately and rapidly track the R ratio of an infant (via the forehead) with a strong correlation (0.98) when compared against a commercial pulse oximeter in the delivery room. A novel algorithm for calculation of the perfusion index is presented.

In 2020 further clinical trials using a modified forehead sensor were conducted to study the ability of the Surepulse VS (2020) sensor to track oxygen saturations in two cohorts: neonatal intensive care and in the delivery room with a new technique called delayed cord clamping. In neonatal intensive care the sensor performed moderately well with an ability to track R ratio changes as oxygen saturations change at a patient level and across a cohort (r2 = 0.45). However, in the labour suite with delayed cord clamping the signal tracked was much smaller than in previous delivery room trials. This made the probes inaccurate (median r2 = 0.006, interquartile range 0.001 - 0.150) when trying to track an R ratio with changing SpO2. Despite a constant SpO2/SaO2 a wide range of R ratios were produced which was labelled the ’flat-R’ problem.

Following this, the reasons behind the flat-R problem were investigated. The main cause of concern was reflections from interface materials between the LED and the skin - this being the main hardware changes between the 2017 and 2020 trials. Investigations into whether reflections were causing the ’flat-R’ problem were undertaken. It was found that the addition of a “plexiglass interface layer” increased the amount of reflected light and was contributing to this problem. In an attempt to increase the accuracy of the sensor, further experiments were conducted in an attempt to improve the signal collected from the sensor whilst simultaneously reducing the proportion of reflected light. This was in an attempt to increase the perfusion index (and hence SNR) of the sensor.

An investigation into the optimal source-sensor separation distance was undertaken, using a halogen light source, spectrometer and fibre optic cables. This was an attempt to find the maximal AC power and perfusion index collected at each wavelength. A custom holder was 3D printed to allow for separation distances between 2mm and 10mm, in 2mm increments, to be investigated. Here, results showed that the magnitude and wavelength of the peak perfusion index values changed with separation distance, with shorter wavelengths having a larger perfusion index at small separation distances and longer wavelengths having larger perfusion index at larger separation distances.

Finally, a new sensor, the Surepulse VSP patch sensor was developed in 2023. This sensor was tested under the BS-EN-ISO-80601-2-61:2019 standard to attempt to generate an SpO2 calibration curve with greater than 200 data points and meeting demographic requirements. 13 participants underwent induced hypoxia whilst PPG signals were acquired with a Sure pulse VS (2020) and Surepulse VSP (2023) sensors. SaO2 data was collected via blood gas analysis and SpO2 data was collected via a commercial transmission mode pulse oximeter. SaO2 and SpO2 R curves were constructed with polynomials fitted to each curve. Following acception and rejection criteria an SaO2 R curve with an RMSE of 4.03% was achieved, albeit with less than 200 data points with only 162 data points passing the acceptance criteria.

Item Type:	Thesis (University of Nottingham only) (PhD)
Supervisors:	Hayes-GIll, Barrie Morgan, Steve
Keywords:	PPG; photoplethysmography; forehead; neonatal intensive care unit
Subjects:	R Medicine > R Medicine (General) > R855 Medical technology. Biomedical engineering. Electronics
Faculties/Schools:	UK Campuses > Faculty of Engineering > Department of Electrical and Electronic Engineering
Item ID:	78154
Depositing User:	Stockwell, Simon
Date Deposited:	18 Jul 2024 04:40
Last Modified:	18 Jul 2024 04:40
URI:	https://eprints.nottingham.ac.uk/id/eprint/78154

Actions (Archive Staff Only)

Edit View