Some of the surprise that used to accompany the birth of a baby is no longer there. That's because imaging technology developed during the last decade or so gives physicians and parents-to-be - if they choose to know - a preview of what's ahead. Ultrasound is now routinely used during prenatal care to create images of the developing fetus. Months ahead of delivery, these images help track the gestational age of the fetus, alert the obstetrician to potential medical problems, and even offer clues about the sex of the child.

While ultrasound may be best known for providing a "window into the womb," it is now used routinely in other medical fields. Ultrasound can produce images of blood flow, for instance, giving cardiologists insight into blocked vessels in the heart. It helps diagnose cysts, tumors, and malformations throughout the abdominal region.

It was only during the 1980s that the first useful pictures could be produced with ultrasound, says Matthew O'Donnell, professor of electrical engineering and computer science and an expert in ultrasound technology. He is working toward making this imaging technology a stand-in for the eyes and hands of a physician.

"The physics of ultrasound hasn't changed, but the technology that we can bring to bear is advancing," says O'Donnell. "New electronics and probe technology permit us to draw more and more information using ultrasound."

In O'Donnell's engineering lab, research is underway to squeeze the best images possible out of existing technology, figure out ways different body characteristics could be imaged with ultrasound, and to build what seems to be impossibly small probes. O'Donnell is trying to adapt ultrasound to view the inside of blood vessels, create "virtual biopsies," and provide the ability to "feel" tissues inside the body, among other projects.

The principle behind ultrasound's ability to make pictures is simple and familiar to most people. "Ultrasound works by a process called echo ranging," says O'Donnell, the same process at work in sonar and used by bats for navigation. An electronic device called a transducer both generates the sound and detects echoes. The technology directs pulses of high-frequency sound at parts of the body. As these sound pulses - thousands each second - reach tissues within the body, some of the sound is reflected back to the sound source. Each tissue reflects a bit differently, depending on the density and composition of the tissue. The "listening" transducer detects these echoes and records the time lag since the initiation of each pulse.

The time lag and intensity of each echo varies depending on the specific tissue that causes the reflection, explains O'Donnell. "A lucky property of tissue is that the amount of sound that bounces back is small, but easily detectable, so that the rest of a pulse continues through the tissue to other reflective structures further along the pulse's path," he explains.

The record of sounds received by the transducer is passed along to a computer. There, lightning-fast calculations are used to "draw" the pictures shown to the expectant parents, or to other patients, and their physicians.

The sound is merely the "vector that is carrying information," O'Donnell says. "The ultrasound is not involved in creating images. That is done by computers which interpret the detected sound reflections."

The detail in these pictures can be quite fine. On a microscopic level, ultrasound detects cell wall boundaries "exquisitely," says O'Donnell. "Inside some larger cells, you can even see organelles [internal cell structures]."

But as anyone who has undergone an ultrasound will agree, not every image is simple to interpret. "Like any other [imaging technique], there can be confusing features," explains O'Donnell.

One constant in O'Donnell's work is a search for new ways to improve the images and expand ultrasound's uses. "To have a good imaging technology, you need a physical parameter to detect, something that is sensed and used to make a picture," he explains. "What I am constantly looking for are these new physical parameters that will provide contrast within the body."

Looking inside blood vessels

A physical condition that confronts some 1.8 million Americans according to the American Heart Association is a narrowing of the arteries to the heart muscle caused by build-up of arteriosclerotic plaque. This restricts blood flow to the heart muscle, leading to pain, damage, and sometimes death. One way to open those arteries is through a procedure called balloon angioplasty. A catheter tipped with a tiny balloon is threaded through blood vessels to the spaghetti thin arteries that feed the heart muscle. There the balloon is inflated and the plaque is pushed aside, opening up a channel for blood to flow again.

Traditionally, the catheter was guided to the arteries "blind" using an external fluoroscope, which produces X-ray images of the catheter and circulatory system. O'Donnell has been part of an industry-university collaboration that has put ultrasound technology at the catheter tip to guide the balloon on its way. "The ultrasound probe becomes the eyes of the interventional procedure," says O'Donnell.

Endosonics is a California-based medical equipment company that developed the miniaturized hardware to go inside the artery. O'Donnell's contribution has been to improve the quality of the images produced from the ultrasound data. [See "Industry-University Collaboration: Mutual Aid" on page 13.]

The next step is to put the ultrasound array inside the angioplasty balloon. That should allow even better navigation of the catheter to the correct locations, and aid in the proper inflation of the balloon. Researchers hope this will reduce the incidence of incomplete inflation. Then perhaps fewer angioplasties will have to be repeated.

Sound and Touch

For now, the ultrasound array on the catheter tip is providing a guided tour of the arteries. Next O'Donnell and his collaborators want to use ultrasound to give physicians one of the oldest medical "tools"- touch. "We like to say that palpation has been around since Hippocrates, but that it's been used only for diagnosis, not imaging," says O'Donnell. Ultrasound offers a way to develop "a remote sense of touch."

The key to using sound to obtain "touch" is ultrasound's ability to track motion. If exploited properly, ultrasound is very sensitive to motion, explains O'Donnell. Physicians could measure how stiff or elastic the artery wall is. The angioplasty balloon provides a way to do this. "While using the ultrasound to generate images, you deform tissue by inflating the balloon to press on the artery wall," O'Donnell continues.

By relating how much motion occurs to how much pressure was used to move the artery wall, researchers can determine the rigidity of the wall

Researchers can mathematically construct the motion because of the presence of artifacts in the image, called speckle. These are small imperfections in the image caused by a variety of structures that scatter the ultrasound in different ways. These are not random fluctuations- not just noise- but data used to track motion," explains O'Donnell.

The O'Donnell research team now has an imaging system that takes about four seconds to get an image they can use to assess the "feel" of the tissue. To make this technique useful for physicians, they need a system that works still more quickly.

A Clearer Image

In addition to developing new uses of ultrasound, O'Donnell is also trying to make the equipment do a better job of showing images. There can be a great deal of variability from patient to patient in the quality of the images. Today physicians and ultrasound technicians must use their best judgment to determine what is real in an image and what is an artifact. O'Donnell is trying to make the ultrasound equipment do a better job of discriminating the real from the distortion.

This is a well-known problem that no one has been able to solve satisfactorily, he says. "It occurs because the body is acoustically a heterogeneous medium, yet ultrasound imaging assumes a homogeneous medium." In its journey through tissue, sound encounters varying environments, but the ultrasound technology works as if the pulses all travel through an identical medium. The calculations used to form the image make that assumption. It results in ambiguous features in the image.

"It's a seductive problem, trying to model the distortion correctly and quickly," says O'Donnell. "We can't spend a lot of time calculating the distortion, since it can change with each image, sixty times a second.

"We have developed two new adaptive algorithms, basically two big mathematical tricks that the system employs to distinguish distortion from a 'real image'," O'Donnell explains.

Electronics plays a big part in the solution. As computing power increases, mathematical corrections for distortion can get more complex. O'Donnell has in mind to use the imaging system itself to detect imperfections, "filtering" the sound signals to account for the distortions.

This research is part of a $12 million contract with the Department of Defense's Advanced Research Projects Agency and GE. O'Donnell's group hopes to demonstrate the feasibility of the filtering system in about three years, and have a working prototype built in about five years.

"Although this is an old problem, the research project gives us a chance to build a new gizmo that may make a significant difference in the application of ultrasound imaging," says O'Donnell.

Star Trek "Scanman"

Another "gizmo" that TV watchers may wish every doctor had is that nifty all-knowing, hand-held scanner that the doctor in the original Star Trek series waved over the patient to get instantaneous and detailed diagnostic information.

O'Donnell is working on a project that while not exactly in the Star Trek league of diagnostics would nevertheless provide ultrasound imaging in the field. This so-called "scanman" would be a hand-held, but otherwise conventional, ultrasound imaging system. "The question is, can you really do it?" says O'Donnell .

Funding for this research comes from the Department of Defense because America's military planners are interested in developing a portable imaging system to use on a battlefield. Today once someone gets to a military hospital, their chances for survival are very good. However, the Department of Defense claims the ability to decide in the field who to transfer first is no better than it was a century and a half ago during the Crimean War. Triage requires very fast decisions, and portable ultrasound equipment would contribute greatly to speedy decisions. The "scanman" would also be a valuable addition to civilian ambulance equipment.

"The challenge for this project is to make the systems and circuits do complex imaging in a small device," says O'Donnell. The UM researchers are developing an imaging system architecture that can achieve good image quality in a device about as large as a paperback book.

"We need to create a precise ultrasound 'lens' that is small, and then be able to do a very efficient calculation to produce a good image. Our part of this project is to find the 'tricks' that will generate the image. We are developing circuit architectures that are very different from those used to generate images in existing devices," O'Donnell explains. He labels this "a pure electronics technology problem." The work is being done with Q-Dot, a small Colorado electronics company. O'Donnell's goal is to have a prototype ready in about three years.

Virtual Biopsies

While he works to miniaturize conventional ultrasound equipment, O'Donnell is also developing a new way of using high frequency sound to reduce some of the trauma of needle biopsies. In these procedures, hollow needles are inserted into "suspicious-looking" regions to extract bits of tissue. Such biopsies are done commonly in the breast or prostate to check for cancer cells.

In the ultrasound version, the needle would still be inserted into the suspect area but no tissue would need to be removed. "You might think of this as a virtual or phantom biopsy needle," says O'Donnell.

Ultrasound can make fine distinctions among cell types, but a biopsy requires a three-dimensional image. The idea is to create an ultrasound device on the tip of a biopsy probe that can generate three-dimensional images. The ultrasound biopsy needle would measure microscopic changes in density and compressibility of the tissue, and then display these differences in a way that allows for three-dimensional analysis of the tissue.

The difficulty is in attaining sufficient spatial resolution using a device that is small enough to be loaded on the tip of a 1 mm needle. O'Donnell hopes to accomplish this by using laser light.

"We are looking at some new means of generating and detecting sound using lasers," he explains. But rather than send a laser beam through an instrument, the system itself would be the laser. Then one optical fiber, with one light source, might emulate a much more complex traditional ultrasound probe.

O'Donnell has developed a "tabletop instrument" that operates about one-twentieth of the speed the system will eventually need to work.

"This is a classic 'systems' project: all of the individual technologies that we will use have been developed," explains O'Donnell. "What we are doing is assembling them in a novel way. We didn't invent the idea-it had been proposed elsewhere. We didn't do the physics research to demonstrate theoretical feasibility, what we are doing is trying to make this a reality-no one else has."

The first hurdle is to prove that light can generate ultrasound in a useful way, then transmit this data to a computer to generate the image, "even if [this image calculation] takes a weekend," he says. "Then others will get excited and may be willing to do the miniaturization that will be required to turn this into a medical technology."

As O'Donnell considers the variety of ultrasound projects he works on, he notes, "You can see that some involve investigating new modalities; some are more basic studies; and some focus on applying knowledge. I think you need to do all of them-be involved both in new ways to make pictures, and in trying to really make devices that function and are practical."