Professor Garry Duffy and Rachel Beatty show the soft robotic implant developed by University of Galway and MIT. [Martina Regan]

Incorporating artificial intelligence and soft robotics into an implantable drug delivery device allows the consistent release of medication, bypassing problems caused by the body forming scar tissue around it.

The FibroSensing Dynamic Soft Reservoir (FSDSR), described in Science Robotics, senses when a fibrotic capsule is being formed as a result of the device being recognized as a foreign body.

The biocompatible device then uses AI to change shape so it can maintain consistent drug dosing, allowing it to circumvent the fibrosis that can interfere with its action and cause it to fail.

The system, developed by a transatlantic team of researchers, holds promise for fine-tuning the delivery of therapeutics in chronic conditions, such as insulin in diabetes.

“The technology which we have developed, by using soft robotics, advances the potential of implantable devices to be in a patient’s body for extended periods, providing long-lasting therapeutic action,” explained co-lead author Rachel Beatty, PhD, from the University of Galway.

“Imagine a therapeutic implant that can also sense its environment and respond as needed using AI – this approach could generate revolutionary changes in implantable drug delivery for a range of chronic diseases.”

The foreign body response is a significant hurdle in the development of long-term implantable devices.

Myofibroblasts lay down collagen around the implanted material causing fibrotic encapsulation that stops the device from interacting with surrounding tissues, often affecting its action and leading to failure.

The smart FSDSR device includes a conductive porous membrane that can detect scar tissue, which interferes with electrical signals traveling through it. Studies showed that electrical impedance correlated with the thickness and volume of the fibrotic capsule formed.

The system then deploys mechanotherapy, in which the soft robotic implant makes movements such as inflating and deflating. Timing, repeating, or varying these movements helps prevent scar tissue from forming.

The University of Galway and MIT team assessed their FSDSR device in vivo, in vitro and in silico.

They found that a machine learning algorithm was able to predict the number and force of actuations to enable consistent drug delivery, regardless of how much fibrosis there was.

Even when significant fibrosis was simulated, the device was able to determine the optimal regime to deliver a consistent dose by changing the way the pumps delivered medication.

The researchers say that the system has the potential to be used in implantable systems for infusion pumps and for repeated targeted drug delivery to localized areas, which could be useful in peritoneal cancers.

Additional sensors, such as a continuous glucose monitor, could also input data for a closed-loop insulin delivery system for the treatment of diabetes.

“If we can sense how the individual’s immune system is responding to an implanted therapeutic device and modify the dosing regime accordingly, it could have great potential in personalized, precision drug delivery, reducing off-target effects and ensuring the right amount of drug is delivered at the right time,” said researcher Ellen Roche, PhD, professor of mechanical engineering at MIT.

“The work presented here is a step towards that goal.”

Also of Interest