Everyone snores occasionally, but for some, snoring can indicate a more serious issue: sleep disordered breathing. This refers to a range of sleep-related respiratory conditions that include obstructive sleep apnea, which is estimated to affect a billion people worldwide.

One common treatment is positive airway pressure (PAP) therapy, which involves sleeping patients wearing a mask connected to a machine that blows a steady stream of air to their upper airway, keeping it open to enable normal breathing during sleep.

Despite its effectiveness in treating sleep disordered breathing, PAP therapy is a challenge for many patients in the long term. “Many patients are intolerant to PAP treatment due to poor mask fit,” explained Song Peng, Assistant Professor at the Singapore University of Technology and Design (SUTD). He highlighted that poor mask fit leads to not only patient discomfort, but also air leakage issues that reduce the effectiveness of the treatment.

In response to these challenges, Assistant Prof Song led a team comprising researchers from Chinese universities and clinicians from Singapore’s Changi General Hospital to use computational modeling to design PAP masks that optimize comfort while ensuring minimal air leakage. The research is published in the journal Computers & Graphics.

The team first modeled a custom-fit PAP mask by making use of two components to form the mask interface with a patient—the mask cushion that comes in contact with the patient and a connector that joins the cushion to a generic mask frame. Having previously observed that the shape of most commercially available PAP masks can be represented by something known as a “swept surface,” the team then modeled the mask cushion using parameters from a 3D swept trajectory.

To optimize the custom-fit PAP masks for different faces, the team began by scanning the faces of four volunteers to create 3D models. These models were then combined with various PAP mask designs to simulate how each mask would fit in its “equilibrium state,” which represents the mask’s final shape after accounting for the forces applied by the straps and the patient’s facial features.

Due to the computational cost of accurately simulating the mask-face fitting process, the researchers chose to focus on the human face and the mask cushion in their design. The human face was assumed to be rigid, while the mask cushion was treated as a deformable body. By using a method known as quasi-static finite element modeling, and creating a tetrahedral mesh for the cushion to support the simulation, the researchers could then simulate a fitted mask.

The masks were evaluated based on two key criteria: potential air leakage and comfort level. These factors were quantified by analyzing the forces acting on the mask in its equilibrium state. Thereafter, the team further optimized the mask’s geometry to minimize leakage and maximize comfort while ensuring that the cushion remained easy to fabricate.

The final step in the research was to validate these custom-fit PAP masks against commercially available ones. The volunteers were subjected to PAP therapy with three different masks: a commercially available mask, a baseline custom-fit mask designed by inverting the subject’s face shape, and the team’s custom-fit mask, which was fabricated with a silicone mold. Air leakage was also measured at different air pressures during the treatment.

At the end of the treatment, the volunteers filled in a questionnaire on their experience with the different masks. The results proved highly promising, as the custom-fit masks demonstrated reduced leakage while also being more comfortable than the other masks.

Reflecting on the use of design and technology in the study, Assistant Prof Song believes his findings will benefit both health care professionals and computational design researchers. For respiratory sleep professionals, the custom-fit masks offer a promising solution to the challenges posed by ill-fitting traditional masks. Meanwhile, the computational modeling approach employed by the team could prove interesting to researchers working in computational design and/or computer graphics.

“Our computational design approach closes the design loop in computers—that is, automating the design process—by performing mask evaluation via simulation-based analysis and mask design via optimization,” he said.

The team is eager to expand their research beyond PAP masks and delve into the computational design of personalized wearable artifacts, such as oxygen masks, swimming goggles, and even headsets for virtual and augmented reality platforms.