Bursting Bubbles Breach Blood-Brain Barrier: Blogger Bequeaths Belated Boasts

There was a story in the news today about the blood-brain barrier being bypassed* in order to deliver chemotherapy drugs directly to a brain tumour, using a combination of microbubbles and focused ultrasound. I worked on this project for over a year between finishing one university and starting another university (2008-2009), and it was good to see it finally in use. Even though they have achieved a medical feat, the phenomena behind it are quite physicsy.

Bubbles oscillating near red blood cells. From one of the news articles.

Focused ultrasound is what it sounds like: applying high intensity sound waves to a specific part of the body. Constructive interference allows the sound intensity to be maximized within the body rather than at the surface, and if the waves are strong enough then the tissue will start heating up as some of the acoustic energy is absorbed. It often fills the same medical niche as radiation therapy, except without the typical side effects of radiation exposure. My first university summer job involved working on the electronics for a focused ultrasound treatment for prostate cancer. Having one's prostate burned by sound waves from the inside may sound unpleasant, but it's not as unpleasant as having the whole thing removed.

High intensity focused ultrasound thermal therapy. All other images I could find involved detailed 3D renderings of the rectum. There's more to HIFU than just butts. Image source.

After graduating from university, I got a full time job in the Focused Ultrasound Lab at Sunnybrook Hospital in Toronto. There, they had developed what was essentially a helmet full of hundreds ultrasound transducers, designed for constructively targeting sound waves inside the brain. In addition to designing this device, they had to solve such problems as "How much is the skull going to refract the waves?" and "How do we avoid burning bone before flesh?" This device has since been used to therapeutically zap through the skull, but that's not what the news is about today.

Inside the transducer dome helmet array. The head goes in the middle.

My specific project involved microbubbles: really small bubbles (duh) that are used as contrast agents during diagnostic ultrasound. They are injected into the bloodstream (they are far too small to cause embolisms, about the size of red blood cells), and when an ultrasound wave hits them, they contract and expand in phase with the applied pressure wave, re-emitting sound waves as they drive the surrounding fluid with their oscillation, which can then be detected. All the videos of this on youtube suck, so out of protest I won't post one. My supervisor, Kullervo Hynynen, wanted to move beyond bubble diagnostics and into bubble therapy. His plan, as we have seen, was to use bubbles to open the blood-brain barrier and deliver drugs to the brain.

Because most of the articles I write are about physics and math, I'll remind the readers that the blood-brain barrier is not a physical separation between the brain and the arterial network, but rather refers to the impermeable network of proteins that forms around the walls of blood vessels inside the brain, that prevents molecules from getting from the bloodstream into the brain. This is useful for preventing blood contagions from affecting the brain, but makes it hard to get drugs into it (cocaine is a notable exception).

Two diagrams of how this works, to hammer the point across. Bubbles are injected into the blood stream, focused ultrasound makes them oscillate and/or collapse, that collapse opens the vessel wall.

The general plan was to use the energy absorption of ultrasound by bubbles to raise the temperature in their vicinity, as well as to create shockwaves from their collapse. It was hoped that either the increased temperature would cause the proteins making up the barrier to relax their grip, or just to violently shear them away.  At the risk of repeating what I talked about in "My Journey into the Hyperbubble," My research involved developing a theory to describe bubble oscillation inside blood vessels, then apply that to a 3D model of the blood vessels in a rat brain, figuring out how much heat would be transferred to the brain based on bubbles oscillating in those blood vessels.

From my paper, a rendering of the heat distribution inside the 3D rat brain. This rotating gif is way cooler and you can see the hot-spots in the blood vessels from the bubbles, but I'll only include a link because it'll kill somebody's data plan: HERE

My solution involved solving a modified version of the Rayleigh-Plesset equation (which is itself derivable from the Navier-Stokes equation) to simulate the bubble oscillation dynamics, calculate the power radiated from those dynamics (through a thermal damping term), and use that power as an input for the Pennes bio-heat equation, which is like the regular heat equation except with a blood flow term that we chose to ignore anyway. The idea was that the results of my simulations would inform future neurologists how much ultrasound to use and at what frequency to get the best results and not fry the person's brain.

The results of one of my simulations from 2009, showing the temperature around the bubbles in the vessels increasing over time.

After I started grad school, my particular project (the localized heating simulations, not the whole research program) didn't really go anywhere. Oh well. However, the Focused Ultrasound Lab kept working on developing this treatment. They apparently used it to treat Alzheimer's in mice.

Today, some news articles cropped up on Facebook about how this treatment has been successfully used to bypass the blood-brain barrier in humans, which is a pretty big milestone for any biomedical development process. There is not yet a peer-reviewed journal article on the topic, but from the news articles, they used this procedure to deliver a chemotherapy agent to a patient's brain tumour (without having to flood the entire body with it, one of the main issues with chemo).

I do not know whether my calculations factored into the patient's treatment. I would hazard to guess that they did not, because they never verified them experimentally and I don't have enough faith in my simulations to recommend going directly from simulation output to sonic brain zapping. However, it is a good feeling nonetheless to see something that I worked on in its early stages finally come to fruition. It is a good example of how a good old fashioned physics problem and an application of fluid dynamics, acoustics, and heat transfer starts saving lives in less than a decade.

*this one is not intentional, this topic is just really alliterative.