Skip to main content

Could Ultrasound Replace the Stethoscope?

Miniaturization, experimentation, and A.I. have unlocked revolutionary potential in an old technology.

Illustration by Nicholas Konrad / The New Yorker

The patient, a man in his early twenties, hobbled into the E.R. on a Wednesday morning, anxious and gasping, his shirt covered in blood. Minneapolis in the nineteen-eighties was experiencing an increase in violent crime that would later earn it the nickname Murderapolis; at Hennepin County Medical Center, the city’s safety-net hospital, stabbings and gunshot wounds had become commonplace. Doctors there had treated dozens of patients with wounds to the chest, and the outcomes had been dismal: roughly half had died, and many survivors suffered brain damage.

The chest contains the heart, the lungs, and the body’s largest blood vessels. The challenge for a doctor is figuring out which, if any, organs have been injured, since each must be treated differently. For decades, medical texts had advocated using a stethoscope for this task: in theory, doctors could use a patient’s breath pattern to detect a collapsed lung, or hear the muffled sounds of a heart filling with blood. But in reality the stethoscope performed poorly in the emergency room. It was dangerous to just treat and hope for the best: by acting without a clear diagnosis, a doctor could harm or even kill a patient, who might turn out to have only a superficial injury.

If the bloodied man at Hennepin had arrived a day earlier, he might have died while his doctors continued to monitor him. But he had stumbled into an experiment. A small group of Hennepin doctors had decided to place an ultrasound machine in the E.R.’s trauma bay, to see if they could quickly diagnose hemorrhaging in the heart. Ultrasound lets clinicians see inside the body in much the same way echolocation allows bats to navigate at night: a probe emits sound waves at a frequency far beyond human hearing, and these waves bounce off bone but pass through fluid, allowing the probe, which is also a receiver, to sense the body’s interior. On an ultrasound screen, bones appear bright white, flowing blood looks black, and most other bodily tissues are visible in different shades of gray.

As doctors and nurses descended on the injured man, someone rolled the half-ton ultrasound machine close and placed its probe on his chest. Sound waves spread imperceptibly through his body, and an instant later his heart filled the screen. It was surrounded by light gray: blood was beginning to suffocate it. The man was rushed to the operating room, where surgeons quickly drained the encroaching blood and repaired the wounds to his heart. He recovered without significant disability.

Ultrasound is an old technology, with roots in the sonar scanners used during the Second World War. For decades, it’s been used mainly to inspect fetuses while they’re still in the womb, and to examine diseased hearts. But, in the past few decades, rapid advances in computer technology, combined with the trial-and-error work of clinicians, have transformed ultrasound into a powerful diagnostic instrument for everything from damaged organs to tuberculosis. If ultrasound’s evangelists are correct, it may soon replace the stethoscope as the quintessential doctor’s tool. Its rise, meanwhile, reveals something about how technology works. In some cases, inventions arrive fully formed. But others reveal their true potential slowly, truly coming into their own with the passage of time.

Sonar uses pings that humans can hear. Ultrasonic frequencies, which are higher and inaudible, were first employed in metal-flaw detectors—machines that shipbuilders used to spot defects in their hulls. At first, it wasn’t obvious how to adapt the technology for medicine. One pioneer tried to use ultrasound to look at the brain; unfortunately, that’s one of the organs least conducive to ultrasonic imaging, since it’s encased in a skull of reflective bone. The first ultrasound machines were enormous, in part because, since air causes ultrasonic waves to scatter, patients had to be submerged in water. (Today, clinicians use gel to create an air-free interface between probe and patient.)

Most ultrasound trailblazers were engineer-physicians with a thirst for experimentation. As a young medical officer in the Royal Air Force during the Second World War, Ian Donald, a British obstetrician, witnessed firsthand the power of both sonar and radar; later, he wondered if ultrasound might be more effective than a physical exam at distinguishing benign cysts from cancerous masses. He persuaded a Glasgow boilermaker to let him turn its metal-flaw detector on two trunkfuls of recently removed tumors, cysts, and fibroids. In 1956, Donald and another young physician, John MacVicar, used a primitive ultrasound machine of their own design on a patient who’d been diagnosed with inoperable cancer. The diagnosis had been based on X-rays and physical exams. The ultrasound, by contrast, suggested that the mass was a large ovarian cyst —a benign growth that could be removed easily through surgery. Doctors removed the cyst, and the patient’s symptoms disappeared.

“From this point, there could be no turning back,” Donald reportedly said. But his colleagues were not convinced. Early ultrasound machines were hard to use and created murky pictures. Donald’s team took the positive step of replacing the water bath with a probe, but used olive oil to bridge the gap between probe and body—a messy proposition for both patient and practitioner. To many doctors, ultrasound seemed like a crutch for those who hadn’t mastered the art of the physical exam. One physician told MacVicar that ultrasound would only be of value “to a gynecologist who was blind and had lost the use of both hands.”

The stethoscope, medicine’s most totemic object, had faced similar obstacles. In 1816, a physician named René Laennec was treating a young woman with cardiac disease; worried about the impropriety of putting his ear directly to her chest, he rolled a piece of paper into a tube, placing his ear at one end and his patient at the other. To his surprise, he found he could hear heart and lung sounds more clearly than with his ear alone. Laennec spent years refining and improving his stethoscope—the name derives from the Greek words for “looking” and “thorax”—before publishing a book describing his findings. But adoption was slow. Critics argued that the tool was too difficult to use, and that the training required was too specialized. Even the Scottish physician John Forbes, who translated Laennec’s treatise into English, wrote that he doubted the stethoscope would “ever come into general use.” It took numerous revisions to the device’s design—early models still resembled rolled-up tubes—and the demonstration of replicable and meaningful results for Laennec and his acolytes to overcome these objections.

If you like this article, please sign up for Snapshot, Portside's daily summary.

(One summary e-mail a day, you can change anytime, and Portside is always free.)

In his book “The Diffusion of Innovations,” from 1962, the sociologist Everett Rogers identifies five characteristics that explain the success or failure of new technologies. The most obvious is relative advantage: a new invention must offer a clear improvement over what has come before. But it must also mesh with current practice patterns, be simple to use, and be easy to try out. On those scores, early ultrasound failed miserably. Even into the nineteen-sixties, ultrasound machines remained large and difficult to transport, and required specially trained operators. They produced grainy still images, initially captured on Polaroid film. Obstetricians were open to ultrasound, because they wanted to avoid exposing fetuses to the radiation created by X rays. Other doctors adopted an attitude of wait and see.

The first wave of substantial improvements came through digitization. As silicon chips replaced vacuum tubes, ultrasound benefitted from Moore’s Law; image quality improved dramatically even as the size of the machines shrank. Manufacturers simplified their user interfaces, making the machines accessible to non-techies. In the nineteen-nineties, darpa, the Defense Advanced Research Projects Agency, awarded a grant to design an ultrasound unit that was portable and durable enough to be carried onto the battlefield. In 1999, a company called Sonosite released a commercial version—the first handheld ultrasound device. The race toward miniaturization continued: today, there are ultrasound machines that can plug into your smartphone.

As a technology spreads, experimentation ensues, and new ideas get refined and regularized. In the early nineteen-nineties, Grace Rozycki, then a surgeon at Grady Memorial, a hospital in Atlanta, studied how ultrasound could be used in evaluating trauma patients. “Surgeons recognized rapidity as ultrasound's most valuable quality,” Rozycki told me. She and her colleagues helped pioneer the use of the fast exam—for Focused Assessment for Sonography with Trauma—to allow them to make treatment decisions sooner.

I learned to perform the fast exam as an emergency-medicine intern. I’ll never forget my first patient with a positive scan—a person in their fifties who’d been struck by a car after they lay down in the road, in a likely suicide attempt. The stretcher came careening through the double doors of the ambulance entrance; as it passed the threshold, a nurse rushed to put an I.V. in the patient’s arm, while another connected them to a monitor that began displaying their vitals. In a worrisome sign, the patient was becoming increasingly confused.

I rolled the ultrasound machine to the bedside, squirted some gel across the probe, and placed it on the right side of the patient’s abdomen. Most probes radiate ultrasound waves outward in an arc, and as a result the images have a phantasmagoric quality, as though a flashlight is being shone into murky waters. When the patient’s kidney came into view, it was surrounded by a pool of black—an abdominal hemorrhage. In an instant, we knew that surgery and a blood transfusion could make a life-changing difference.

This versatility has proved particularly valuable in places with limited access to medical care. “I basically bring the ultrasound into every patient visit,” Ashley Weisman, an emergency-medicine doctor who practices primarily in rural areas, told me. For a time, Weisman worked at a small hospital in Kotzebue, Alaska—one of the most remote hospitals in the United States. She began using ultrasound for home visits and in village clinics. “You might have a patient in their sixties or seventies with shortness of breath, but to get them to a clinic—you have no ambulance, some of these villages don’t even have roads,” she said. “But you can go to their house and do lung ultrasound at the bedside, and figure out it’s their heart failure that flared up, and change their meds. You don’t necessarily have to strap them to the back of someone’s A.T.V. or put them on a plane.” Tapiwa Kumwenda, who practices at a hospital in Lilongwe, Malawi, where tuberculosis is endemic, told me that clinicians there routinely use ultrasound to diagnose the disease. Traditional TB testing can take days, and is often difficult to perform in remote settings. “With the ultrasound, you can see the microabscesses and the lymph nodes, and you know it is TB,” Kumwenda said. “You start them on treatment and then two weeks later most of them will be stabilized.”

The price of a handheld ultrasound today is around three to five thousand dollars—low enough that a number of American medical schools have begun giving handheld ultrasound probes to their first-year students, in addition to the traditional gift of a stethoscope. “I keep joking we’re going to have a big bonfire, and we’re going to take all those stethoscopes and burn them, because there aren’t many times a stethoscope helps us today,” Diku Mandavia, an emergency-medicine doctor and early pioneer of point-of-care ultrasound, told me. “But ultrasound—it’s low-cost, no radiation, has so much value for patient care . . . it’s going to be ubiquitous.”

And yet the path forward isn’t free of obstacles. Since the nineteen-eighties, many radiologists have argued that ultrasound, by ceding interpretive responsibility to non-radiologists, could lead to incorrect diagnoses. At Hennepin County Hospital, radiologists objected so strongly to ultrasound’s early use in the E.R. that the affiliated medical school threatened to pull its residents if it continued. The American Medical Association has made it much more difficult for radiologists to block other specialties from using ultrasound—but the fact remains that most new physicians don’t get significant ultrasound training, and most specialties don’t yet emphasize its use.

Ultrasound boosters argue the integration of artificial intelligence will give the technology the momentum it needs to reach into primary care and other areas of medicine where it’s not yet widespread. A hospital where I worked last year purchased new ultrasound devices that automatically calculate how much blood the heart pumps out to the body with each beat—a critical piece of information in determining whether a patient’s heart is failing. I used to estimate how much blood was being pumped by watching the screen, or, if I wanted to be more quantitative, by carefully measuring the movement of one of the heart valves—a relatively time-consuming process. “I think the key will be smart machines,” Michael Blaivas, an emergency-medicine physician who focusses on ultrasound, said. “The machine has to meet us partway. That’s how we really increase access on a mass scale.”

Ultrasound may also benefit from what Richard Hoppmann, a rheumatologist who has spent years teaching ultrasound to medical students, calls its “awe phenomenon.” “The first time these students use ultrasound to look under the skin to see inside the body, they have this incredible excitement,” he told me. Recently, I taught an ultrasound seminar to a group of new medical students and witnessed a similar reaction. It’s one thing to read about heart valves, and another to put a probe to your chest and see your own flapping with each passing beat. Patients can share this sense of awe. Hoppmann helmed a project in which patients with high blood pressure received periodic ultrasounds and were shown their own heart muscle thickening—an indication that their uncontrolled blood pressure might soon lead to heart failure. “The value of being able to see the heart muscle thickening as it works against the blood pressure and to point at what you’re talking about—that started making a difference in terms of patients being compliant with their medications,” he said. Now, whenever I perform a bedside ultrasound, I try to show the patient and family what I am seeing and what I am looking for. People in pain seek relief, but also answers; to be able to see them along with your doctor, your faces upturned to the same screen, is a gift.

Medicine has long grappled with an internal debate about how diagnosis should happen. Some physicians argue that today’s doctors rely too much on advanced imaging and lab tests, neglecting physical exam techniques, such as listening for heart murmurs; they maintain that more thorough physical exams could curb the use of medical testing, which is expensive, slow, and, in some cases, radioactive. They also lament the effect of remote diagnostics on the physician-patient encounter. Today’s physicians, they note, often spend more time placing orders and reviewing charts in the electronic medical record than they do interacting with patients.

Other doctors cite concerns about the reliability and reproducibility of some parts of the physical exam. Even feeling for a pulse during a potential cardiac arrest can be difficult. Ultrasound, unlike other tests, can be used wherever the patient happens to be, in real time. The physician and author Abraham Verghese, a vocal proponent of increasing time spent with patients, views ultrasound as a way to bring physicians back to the bedside. “The waning of the bedside exam,” he said last year, might “find new life” through ultrasound.

Improvements in A.I. and telemedicine raise another intriguing possibility: patients could be given their own ultrasound machines, which they could use to monitor chronic conditions at home. Guided by A.I., a heart-failure patient who’s having difficulty breathing could use a small ultrasound to see if fluid is accumulating in his lungs, or to examine his heart; the images could be relayed to a physician, who could decide whether to call the patient into the hospital or update a medication. Blood-pressure devices, glucometers, and dialysis machines all once resided exclusively in the hospital or clinic, and now patients routinely employ them at home.

In fact, this approach is currently being tested two hundred and fifty miles above Earth’s surface, where crew members aboard the International Space Station are using ultrasound to assess themselves for eye injuries, fractures, and other ills. With guidance from physicians on the ground, the astronauts are capturing high-quality images that help inform their own care, and also contribute to research on how life in orbit affects the human body. “I don’t have a medical background and didn’t know anything about sonography when we started,” Leroy Chiao, a now retired astronaut who helped run early ultrasound experiments aboard the I.S.S., told me. “But, with a few hours of training, we were able to begin recognizing a good ultrasound image versus a bad one.” Chiao said that he and the other crew members quickly learned how to “image each other’s internal organs, bones, and eyes in space.” Scott Dulchavsky, a surgeon, recently led a number of nasa’s ultrasound research investigations. If artificial intelligence progresses, he said, and if the price of portable ultrasound devices continues to fall, “it’s going to be ultrasound for dummies pretty soon.” Stethoscopes were just for doctors. Ultrasound could be for everyone. ♦


Clifford Marks is an emergency-medicine physician at Beth Israel Deaconess Medical Center and an instructor at Harvard Medical School.

The New Yorker. Subscribe and enjoy unlimited digital access to The New Yorker. Cancel anytime.