UCSF

Respiratory diseases are a leading cause of death worldwide. Despite modern antibiotics,

treatment of pneumonia and other lung diseases is often limited by the

tools available to diagnose these disorders in low resource settings. While tools such

as chest x-ray and CT scans are highly accurate, their high cost provides a high barrier

for many patient populations. On the other hand, the physical exam provides

a time-honed method for diagnosis of many common lung diseases. Unfortunately,

due to limited sensitivity, the pulmonary physical exam is often insufficient for diagnosis.

The goal of the research presented in this dissertation is to take advantage of

the simplicity of the clinical physical exam, but to quantify its findings using modern

sensors. We present a standardized approach to analysis which characterizes

healthy from diseased lung with 91.7% accuracy. For pneumonia specifically, we

demonstrated 92.3% accuracy in distinguishing between healthy subjects and pneumonia

subjects in a pilot study. In addition to these findings, we also review this

work in the context of the current work in the field and provide suggestions for next

steps to continue quantified acoustic analysis of the lungs.