Before Pierre Curie met the chemist Marie Sklodowska; before they married and she took his name; before he abandoned his physics work and moved into her laboratory on Rue Lhomond where they would discover the radioactive elements polonium and radium, Curie discovered something called piezoelectricity. Some materials, he found—like quartz and certain kinds of salts and ceramics—build up an electric charge when you squeeze them. Sure, it’s no nuclear power. But thanks to piezoelectricity, US troops could locate enemy submarinesduring World War I. Thousands of expectant parents could see their baby’s face for the first time. And one day soon, it may be how doctors cure disease.

Ultrasound, as you may have figured out by now, runs on piezoelectricity. Applying voltage to a piezoelectric crystal makes it vibrate, sending out a sound wave. When the echo that bounces back is converted into electrical signals, you get an image of, say, a fetus, or a submarine. But in the last few years, the lo-fi tech has reinvented itself in some weird new ways.

Researchers are fitting people’s heads with ultrasound-emitting helmets to treat tremors and Alzheimer’s. They’re using it to remotely activate cancer-fighting immune cells. Startups are designing swallowable capsules and ultrasonically vibrating enemas to shoot drugs into the bloodstream. One company is even using the shockwaves to heal wounds—stuff Curie never could have even imagined.

So how did this 100-year-old technology learn some new tricks? With the help of modern-day medical imaging, and lots and lots of bubbles.

Bubbles are what brought Tao Sun from Nanjing, China to California as an exchange student in 2011, and eventually to the Focused Ultrasound Lab at Brigham and Women’s Hospital and Harvard Medical School. The 27-year-old electrical engineering grad student studies a particular kind of bubble—the gas-filled microbubbles that technicians use to bump up contrast in grainy ultrasound images. Passing ultrasonic waves compress the bubbles’ gas cores, resulting in a stronger echo that pops out against tissue. “We’re starting to realize they can be much more versatile,” says Sun. “We can chemically design their shells to alter their physical properties, load them with tissue-seeking markers, even attach drugs to them.”

Nearly two decades ago, scientists discovered that those microbubbles could do something else: They could shake loose the blood-brain barrier. This impassable membrane is why neurological conditions like epilepsy, Alzheimer’s, and Parkinson’s are so hard to treat: 98 percent of drugs simply can’t get to the brain. But if you station a battalion of microbubbles at the barrier and hit them with a focused beam of ultrasound, the tiny orbs begin to oscillate. They grow and grow until they reach the critical size of 8 microns, and then, like some Grey Wizard magic, the blood-brain barrier opens—and for a few hours, any drugs that happen to be in the bloodstream can also slip in. Things like chemo drugs, or anti-seizure medications.

This is both super cool and not a little bit scary. Too much pressure and those bubbles can implode violently, irreversibly damaging the barrier.

That’s where Sun comes in. Last year he developed a device that could listen in on the bubbles and tell how stable they were. If he eavesdropped while playing with the ultrasound input, he could find a sweet spot where the barrier opens andthe bubbles don’t burst. In November, Sun’s team successfully tested the approach in rats and mice, publishing their results inProceedings in the National Academy of Sciences.

“In the longer term we want to make this into something that doesn’t require a super complicated device, something idiot-proof that can be used in any doctor’s office,” says Nathan McDannold, co-author on Sun’s paper and director of the Focused Ultrasound Lab. He discovered ultrasonic blood-brain barrier disruption, along with biomedical physicist Kullervo Hynynen, who is leading the world’s first clinical trial evaluating its usefulness for Alzheimer’s patients at the Sunnybrook Research Institute in Toronto. Current technology requires patients to don special ultrasound helmets and hop in an MRI machine, to ensure the sonic beams go to the right place. For the treatment to gain any widespread traction, it’ll have to become as portable as the ultrasound carts wheeled around hospitals today.

More recently, scientists have realized that the blood-brain barrier isn’t the only tissue that could benefit from ultrasound and microbubbles. The colon, for instance, is pretty terrible at absorbing the most common drugs for treating Crohn’s disease, ulcerative colitis, and other inflammatory bowel diseases. So they’re often delivered via enemas—which, inconveniently, need to be left in for hours.

But if you send ultrasound waves waves through the colon, you could shorten that process to minutes. In 2015, pioneering MIT engineer Robert Langer and then-PhD student Carl Schoellhammer showed that mice treated with mesalamine and one second of ultrasound every day for two weeks were cured of their colitis symptoms. The method also worked to deliver insulin, a far larger molecule, into pigs.

Since then, the duo has continued to develop the technology within a start-up called Suono Bio, which is supported by MIT’s tech accelerator, The Engine. The company intends to submit its tech for FDA approval in humans sometime later this year.

Instead of injecting manufactured microbubbles, Suono Bio uses ultrasound to make them in the wilds of the gut. They act like jets, propelling whatever is in the liquid into nearby tissues. In addition to its backdoor approach, Suono is also working on an ultrasound-emitting capsule that could work in the stomach for things like insulin, which is too fragile to be orally administered (hence all the needle sticks). But Schoellhammer says they have yet to find a limit on the kinds of molecules they can force into the bloodstream using ultrasound.

“We’ve done small molecules, we’ve done biologics, we’ve tried DNA, naked RNA, we’ve even tried Crispr,” he says. “As superficial as it may sound, it all just works.”

Earlier this year, Schoellhammer and his colleagues used ultrasound to deliver a scrap of RNA that was designed to silence production of a protein called tumor necrosis factor in mice with colitis. (And yes, this involved designing 20mm-long ultrasound wands to fit in their rectums). Seven days later, levels of the inflammatory protein had decreased sevenfold and symptoms had dissipated.

Now, without human data, it’s a little premature to say that ultrasound is a cure-all for the delivery problems facing gene therapies using Crispr and RNA silencing. But these early animal studies do offer some insights into how the tech might be used to treat genetic conditions in specific tissues.

Even more intriguing though, is the possibility of using ultrasound to remotely control genetically-engineered cells. That’s what new research led by Peter Yingxiao Wang, a bioengineer at UC San Diego, promises to do. The latest craze in oncology is designing the T-cells of your immune system to better target and kill cancer cells. But so far no one has found a way to go after solid tumors without having the T-cells also attack healthy tissue. Being able to turn on T-cells near a tumor but nowhere else would solve that.

Wang’s team took a big step in that direction last week, publishing a paper that showed how you could convert an ultrasonic signal into a genetic one. The secret? More microbubbles.

This time, they coupled the bubbles to proteins on the surface of a specially designed T-cell. Every time an ultrasonic wave passed by, the bubble would expand and shrink, opening and closing the protein, letting calcium ions flow into the cell. The calcium would eventually trigger the T-cell to make a set of genetically encoded receptors, directing it it to attack the tumor.

“Now we’re working on figuring out the detection piece,” says Wang. “Adding another receptor so that we’ll known when they’ve accumulated at the tumor site, then we’ll use ultrasound to turn them on.”

In his death, Pierre Curie was quickly eclipsed by Marie; she went on to win another Nobel, this time in chemistry. The discovery for which she had become so famous—radiation—would eventually take her life, though it would save the lives of so many cancer patients in the decades to follow. As ultrasound’s second act unfolds, perhaps her husband’s first great discovery will do the same.