Correlation between Preoperative and Postoperative Nasal Cavity Symmetry for Empty Nose Syndrome

Empty nose syndrome (ENS) is referred to as a clinical phenomenon in which people with clear nasal passages experience a range of discomforts. Patients with ENS may complain physical or emotional distress, such as nasal congestion, drainage, headache, paranasal aching, dyspnea, nasal and pharyngeal dryness, hyposmia, rhinorrhea or postnasal drip, nasal crusting, inability to concentrate, chronic fatigue, frustration, irritability, anger, anxiety, or depression and even hyperventilation syndrome. Since the terminology first used by Eugene Kern and Monika Stenkvist in 1994, ENS remains a poorly recognized but crippling disease for rhinologists and patients. The diagnostic criteria for ENS have not been established due to the lack of reliable objective measurement, neither have defined consistent causes. One shared characteristic of ENS is, however, loss of nasal turbinates, resulting into asymmetric nasal septum and the lateral wall. In the proposed project, we plan to quantitatively characterize the symmetry of the nasal cavity and correlate with the patient’s clinical complaints by constructing a deep-learning framework. The aims of the proposed project are threefold: (1) to quantify nasal cavity symmetry/asymmetry from three-dimensional Computed Tomography (3DCT) images; (2) to conduct preoperative and postoperative questionnaire surveys to characterize the patient’s ENS complaints; (3) to construct a symmetry/asymmetry classifier, in form of a deep-learning framework, that can morphometrically categorizes nasal septum and lateral wall and determine its relationship with the patient’s complaints. In the first year of the project, a preliminary deep-learning algorithm shall be designed and developed. The questionnaire surveys will be continuously conducted, with clinical helps of the co-PIs, during the two-year’s span. The questionnaire results, verified and validated, will serve as the ground truth to adjust the proposed model in most part of the second year of the project. The principal investigator expects to publish two academic journal articles.

Project ID:PB10901-1775
External Project ID:MOST108-2221-E182-015-MY2