Smart inhaler tailors drug delivery to each patient

The powerful applications of machine learning are ushering in an era of personalized medicine. Able to adapt to conditions specific to each circumstance, these modern medical tools can accurately assess individuals’ needs to deliver more effective treatment.

One such example is the machine learning-enabled inhaler designed by Islam et al. to treat juvenile-onset recurrent respiratory papillomatosis (JORRP), a chronic condition that can cause life-threatening airway blockages in the larynx and glottis. Their inhaler prototype employs Computational Fluid Particle Dynamics simulations and artificial intelligence to determine how and where to target medication.

“The algorithm will use patient-specific data, such as breathing profiles and drug particle sizes, as its input,” said author Yu Feng. “The inputs can be processed in the algorithm to determine the optimal nozzle diameter and location for drug release, which are the outputs to the hardware. The hardware system adjusts the nozzle accordingly, ensuring precise drug delivery to the larynx and glottis while minimizing deposition on healthy tissues, thus enhancing the effectiveness of JORRP treatment on a patient-specific level.”

By reducing unnecessary medication deposition, the inhaler also reduces unwanted side effects. This study demonstrates the success of their inhaler, as well as their computation model of the airway. Using the same principles, the team can modify their inhaler to treat other respiratory conditions.

“With minimal modifications, it can be adapted to treat other diseases requiring targeted drug delivery at the larynx and glottis, like laryngopharyngeal reflux, an inflammatory condition of the upper aerodigestive tract,” said Yu. “The smart inhaler can also be trained to achieve targeted drug delivery to small airways, to treat chronic obstructive pulmonary disease or asthma.”

Source: “A user-centered smart inhaler algorithm for targeted drug delivery in juvenile onset recurrent respiratory papillomatosis treatment integrating computational fluid particle dynamics and machine learning,” by Mohammad Rashedul Islam, Chenang Liu, Chanjie Cai, Jindal Shah, and Yu Feng, Physics of Fluids (2024). The article can be accessed at https://doi.org/10.1063/5.0186786 .