JUNE Project: Does the modeling of nasal airflows improve the pathophysiological understanding and the diagnosis of functional respiratory disorders? – JUNE

Chronic Nasal Obstruction (CNO) is not measurable in clinical practice and represents an important lack for the clinician. In addition, the components of the NO are poorly known. Simulating the passage of air in the nose from the computer processing of the patient's scanner is a technology inaccessible to the clinician. However, this simulation would make it possible to obtain the physical reference values of nasal respiratory comfort and to measure the NO. However, we need a detailed analysis, correlated with the gold standard and clinical practices, validated on a large number of patients. We want to know the values of the nasal respiratory comfort before and after treatment for normal to pathological situations and finally to know the veracity physical parameters recorded during the various simulations research works. The main stages of realization and techniques used as follow : 1. Building a 3D model of nasal airway from CT-scan including an automatic segmentation of nasal cavities close to the anatomy and the practitioner’s expertise - 2. to develop an automatic adaptive tetrahedral mesh generation suitable for airflow simulation of the previous interest volumes; - 3. To develop an automated construction of surface geometries of nasal cavity using a deep-learning-based approach. - 4. to develop a finite volume solver to simulate the fluid dynamic and optimize the computation;- 5. to correlate simulations results with the symptom (CNO).
This project combines the know-how of engineers and clinicians specialized in the different technological and clinical barriers. The rhinology department of the University Hospital of Bordeaux has the leadership in functional surgeries of the nose and obstructive sleep disorders. It works in close collaboration to the CIC-IT (CIC1401) and optifluides compagny since 10y to resolve numerous issues of rhinology and CFD. OptiFluides is a French private engineering company specialized in Research and Development related to Fluid Dynamics resulting from the work of the National Institut of Applied Sciences (INSA). This project benefits from its network of engineers specialized in biomedical imaging research (Créatis, INSA) and spectrum of computer science challenges of the digital world, including those posed by artificial intelligence (AI), big data analytics, computer vision, cyber security, digital transformation and human learning (LIRIS, University of Lyon1).
The clinical impact comes from the expectation of practitioners in the absence of a diagnostic tool. It is therefore a question of providing practitioners with a new complementary examination validated and correlated to the clinic, allowing them to better understand clinical situations, to argue the therapeutic decision and the follow-up. We want to measure the CNO as a hemoglobin rate, an audiogram or an electrocardiogram. The final objective is to provide a marketable and exploitable prototype which should resolve many uncertainties (architectural CNO, obstructive sleep disorders, treatment of post-cancer quality of life, follow-up of chronic rhinosinusitis, etc.). The impacts are also scientific and economic since no tool is available combining these technological steps automatically improved by deep learning and AI, and because such information would reduce the failure rate of surgeries by 25% sometimes useless, would reduce failures in the care of patients suffering from obstructive sleep disorders, thus reducing the financial burden on the health system.