Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article

Degree

Master of Science

Program

Medical Biophysics

Supervisor

Diop, Mamadou

Affiliation

Lawson Health Research Institute and University of Western Ontario

Abstract

Patient neurological outcomes following cardiac surgery are improved when near-infrared spectroscopy (NIRS) is used to optimize intraoperative cerebral oxygen delivery. However, current NIRS analysis methods have difficulties monitoring adult brains due to contamination from the extracerebral layer (ECL).

The objective of this thesis is to develop a time-resolved (TR) NIRS data analysis method for monitoring adult cerebral oxygen saturation (ScO2) and total hemoglobin (HbT) by assuming the head is composed of two layers – the ECL and the brain. We tested the validity of this assumption using in silico data from an adult human head using two approaches; a few-wavelength, single detector method, and a hyperspectral, two-detector method that does not require prior knowledge of exact ECL thickness. Both methods were able to recover ScO2 and HbT with mean percent differences below 3%. Additionally, the hyperspectral method requires only 0.22 seconds per measurement, enabling quasi-real-time adult neuromonitoring.

Summary for Lay Audience

Cardiac surgery is associated with high incidence of brain injury during and after the operation, occurring in approximately 6% of patients. These injuries include stroke, seizures, and other types of brain damage, and can have lasting impacts on patient quality of life. There is growing evidence that patient neurological outcomes are improved when near-infrared spectroscopy (NIRS) is used to optimize cerebral oxygen delivery during the operation. NIRS uses near-infrared light to non-invasively monitor brain health by measuring cerebral oxygen saturation (ScO2) and total hemoglobin concentration (HbT). However, current NIRS analysis methods have difficulties monitoring adult brains due to contamination from the scalp, skull, and cerebrospinal fluid, collectively referred to as the extracerebral layers (ECL).

The objective of this thesis is to develop a time-resolved (TR) NIRS data analysis method for monitoring adult ScO2 and HbT by assuming the head is composed of two layers – the ECL and the brain. A TR-NIRS system releases pulses of light into the tissue and records how long the light takes to reach a detector. The detected light is then analyzed to measure the physiological parameters of interest. We tested the validity of the two-layer assumption using simulated data of light transport in an adult human head using two approaches. The approaches differ primarily in two ways: the number of wavelengths of light used, and the number of detectors. The first approach uses data from a single source-detector pair at four wavelengths to estimate ScO2 and HbT with a mean percent difference of 2.3% and 2.9%, respectively. The second method recovers the same parameters from 150 wavelengths (referred to as hyperspectral) and uses two detectors. The added information provided by using more wavelengths and a second detector allows the hyperspectral method to estimate the thickness of the ECL, making it more versatile. The accuracy of the hyperspectral method is comparable to the few-wavelength alternative, recovering ScO2 and HbT with mean percent differences of 2.4% each. The hyperspectral method can analyze TR-NIRS data in near-real-time, requiring only 0.22 seconds per measurement, which is a major step towards real-time monitoring of the brain during cardiac surgery.

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Available for download on Saturday, January 01, 2022

Share

COinS