Developing tools for assessing obstructive sleep apnea severity and consequences from existing clinical data

Project Summary/Abstract Obstructive sleep apnea (OSA) affects a significant portion of the middle-aged population in the US and has been associated with a number of health concerns including cardiovascular disease and cognitive dys- function. The development and progression of these health consequences is believed to be related to the severity of OSA. However, current clinical indicators of OSA severity, which include an intermittent hypoxia pattern at the tissues and the apnea-hypopnea index (AHI), fail to capture the impact of the consequences associated with the disease. The intermittent hypoxia exposure pattern at the tissue-level, a qualitative indicator of OSA severity, relies on pulse oximetry for measurement, which has limitations including inaccu- racies in recording and only a generalized representation of systemic arterial oxygen hemoglobin saturation. In addition, the AHI, a quantitative measure of OSA severity, has not been shown to be strongly correlated to disease development in previous clinical studies. Accordingly, there is a large volume of OSA sleep study data that needs to be re-analyzed. Therefore, to better understand the development of OSA-related con- sequences, there is a need to assess OSA severity with a method that avoids the aforementioned clinical limitations. This can be achieved with mathematical modeling and, in this project, we propose to develop a model for a more detailed clinical representation of OSA. In Aim 1, the model will be constructed using fundamental mass transfer equations to track the transport of oxygen and carbon dioxide throughout the body. Considering the lack of patient-specific approaches in current OSA modeling literature, our model will have the ability to use respiratory and heart rate data from polysomnography studies and will be feasible for clinical application. Therefore, to address the limitations of current measures of intermittent hypoxia expo- sure, the result of Aim 1 will be a presentation of blood gas concentration profiles at the arterial and venous ends of various target tissues, which will allow for a quantification of hypoxia burden. In Aim 2, the model from Aim 1 will be used with clinical data to run a correlation analysis between predicted oxygen decreases and patient characteristics such as daytime sleepiness, a potential indicator of OSA presentation, which could provide an alternative to the AHI. The patient results will also be organized into a multi-dimensional database for scientific rigor and ease of access and analysis.Therefore, the two primary outcomes of this project will be a clinically deployable mathematical model for OSA severity assessment and a large pool of data obtained from simulated cases of OSA and a re-analysis of existing sleep studies, which will play an important role in improving patient care. Additionally, this work will be invaluable for my training in clinically- relevant computational research, which I ultimately plan to pursue in my post-graduate career.