Incorporating developments in the field of computational mechanics, UCLA researchers have created a computer model of the upper airway that may ultimately predict obstructive sleep apnea patients’ response to dental sleep appliances.
Whether a dental sleep appliance will be efficacious for a given patient’s obstructive sleep apnea (OSA) is a question that many have been trying to answer. The ideal predictive tool would be inexpensive and quick, at minimum more affordable and faster than creating and titrating the real appliance.
At the University of California, Los Angeles (UCLA), a professor of mechanical and aerospace engineering and his colleagues have achieved promising preliminary results with their novel 3-D upper airway simulation. Where previous simulations have treated the airway as rigid, Jeff D. Eldredge, PhD, MS, and team’s model depicts the tissue as it experiences elastic deformation and collapse with breath. This allows it to better show the pathophysiology of OSA and more realistically predict occlusion—with the ultimate goal of allowing practitioners to virtually “try out” therapies in a given patient’s simulation to find out if they will work in real life.
Eldredge explains: “We seek to simulate air flows in which the forces that the flow exerts on the adjacent tissues causes them to deform and therefore dynamically alter the geometry of the airway in the simulation. At its extreme, this will involve simulating the actual collapse of the airway, rather than just (imperfectly) predicting the potential for collapse if the airway is treated as rigid.”
The UCLA School of Dentistry initially approached Eldredge about finding a way to test whether mandibular devices for OSA would work without having to send patients home to test the product for weeks at a time. Working with Sanjay M. Mallya, BDS, MDS, PhD, associate professor of dentistry, and then-UCLA medical resident Susan M. White, DDS, Eldredge and his partners developed the preliminary computational tool.
To create a patient’s airway simulation, dental cone beam computed tomography (CBCT) scans are first taken. Mallya says, “When assessing patients for dental sleep appliances, we assess not only their airway geometry, but also the morphology of the jaw bones and the temporomandibular joint. We use a low-dose cone beam CT to provide us with 3-dimensional imaging of the jaws and the nasal and pharyngeal airways. This imaging procedure provides us valuable information that are not provided by standard 2-dimensional x-ray projections.” Mallya adds that all patients are also assessed visually with an intraoral and extraoral examination—”The dental cone beam CT scans supplement information that is not readily available by visual examination,” he says.
Once the CBCT scans are taken, the computational tool allows the researchers to alter the geometry of the model to show what effect a dental sleep appliance would have on any given patient.
Developing the Tool
Previously Eldredge worked on creating models that simulated the interactions between blood and vessel walls with Shao-Ching Huang from the UCLA Institute for Digital Research and Education. They recently presented a simulation of a leg being injured by shrapnel, with the goal of training combat medics.
Eldredge says, “Our extension of the tool comes in a larger context. There have been exciting developments in the field of computational mechanics in medicine, in which several different sets of physics are incorporated into the computational tool. Not surprisingly, this development has most notably been in heart modeling, where the elasticity of the heart chambers and valves and the electrophysiology of the heart muscle all have a role that is inseparable from the blood flow.”
Air and blood are governed by the same equations. “The techniques that are being developed for these problems are quite general, however, and there is surprisingly little difference in how one treats the interaction of blood with a deformable valve or chamber wall compared to how one simulates the interaction of air flow with the deformable tissues lining the upper airway,” Eldredge says. “So we are building on a lot of very recent work in order to extend our tool in upper airway simulations.”
Airway Simulation Applications: Mandibular Devices and Beyond
The development of UCLA’s upper airway simulation has been driven by a desire to predict response to mandibular devices, a priori, says associate professor of dentistry Sanjay M. Mallya, BDS, MDS, PhD. In a clinical setting, that could lead to the simulation being used with patients who have been diagnosed with OSA via a sleep study and are being considered for a dental sleep appliance.
But it could also be used to predict the outcomes of surgery for OSA. By altering the geometry of the model, such as shaving a little off a jawbone in the simulation, Jeff D. Eldredge, PhD, MS, a professor at the School of Engineering and Applied Sciences, says the computation could show what effect surgery could have on a patient. Mallya adds, “Perhaps these computational models may direct the design of hybrid therapeutic approaches that combine minimal surgery with a device, improving therapeutic outcome on a patient-specific basis.”
There may also be applications for non-OSA-related procedures that affect airway morphology, for example, cosmetic surgeries that may change upper airway dimensions. “Being able to predict the clinical impact of this change will be an important consideration for both the surgeon and the patient,” Mallya says.—SR [/sidebar]
The team has published a paper detailing the preliminary results.1 Now research student Chien-Jung Huang is implementing the full set of physics into the computational tool, a painstaking process. “At the end of that process, we will publish a paper on our full computational tool, and most likely pursue some studies on real patient data to assess our prediction success for OSA,” Eldredge says.
Sree Roy is editor of Sleep Review.
Photo by Joanne Leung/UCLA
1. Huang CJ, Huang SC, White SM, Mallya SM, Eldredge JD. Toward numerical simulations of fluid–structure interactions for investigation of obstructive sleep apnea. Theor Comput Fluid Dyn. April 2016, 30(1):87-104.