Location, Location, Location

Mayo Clinic’s Richard Robb explains how advances in three-dimensional digital imaging help surgeons locate their targets with pinpoint accuracy.

“Instead of seeing a 2D projection we have the true 3D relationships of all the organs”
-Richard Robb, Mayo Clinic

It all starts with a 3D image. Modern scanners are wonderfully capable of providing the data for making 3D images because they scan very fast adjacent, thin slices. These slices might be less than a millimeter apart around the anatomic region of the body of interest. I then get these stacks of slices of CT scans and MR scans in my lab – one way to think of them is like stacks of pancakes.

We then render three-dimensional images from this using computer algorithms, which make the fundamental transformation from this stack of slices into a recognizable three-dimensional image that then can be displayed in a variety of ways, rotated around, measured and manipulated.

That’s where we start, with the development of a 3D rendition of this stack of slices provided by the medical imaging scanners.

How it’s different from conventional X-ray imaging, which we’ve had available for a long time, is that it’s quantitative. When you take a typical X-ray you get a visual depiction of the inner organs superimposed upon each other, but the CT scanner and the MR scanner and other 3D scanners take images at individual little pixels or boxels inside the body, so we have truly a three-dimensional array of data that is a bag of numbers.

In three-dimensional space we have numbers for every small piece of tissue that was seen by the scanner. That may seem like a simple thing, but it’s a tremendously powerful advantage that digital imaging has over conventional X-ray imaging, because now we can put these data into a computer, which understands and can make measurements from numbers, and can create new displays from them.

Modern scanners are calibrated well enough, and understood well enough, that these numbers correlate to certain things about the tissues that we’re imaging, including as important an issue as differentiating normal from abnormal tissues, and detecting disease and how bad the disease is. This gives us a quantitative representation of what’s going on inside of the body, not just a picture.

Another advantage of having a digital image and having it in the computer is we can manipulate it. The computer can turn it around or can cut into it and can display any view of interest to the physician, to the surgeon, to the scientist or engineer, or whoever might be using the image.

And finally, images are very valuable in medicine for providing two major new capabilities that we didn’t have two decades ago. One of these is computer-aided diagnosis – we can assist the physician in making a more accurate, sensitive and specific diagnosis of the disease from these quantitative images.

The images themselves can also be used in the therapy process. They are increasingly used to navigate and target surgery or radiation treatment, for example, with much more precision and reproducibility in less time than we before.

In the neighborhood
Let’s take one of the examples that is pretty familiar to everybody: a brain tumor. When a patient presents with headaches and has a brain tumor, two or three decades ago you took X-ray pictures, and the surgeon often had to do exploratory surgery to find the tumor and to remove it.

Now with 3D imaging and the high resolution we get with CT scanners and MRI scanners, we can look inside that patient painlessly and non-invasively and eliminate exploratory surgery. A patient who has a brain tumor now can have an MRI and a CT scan and maybe also a PET scan, which tells us something about the function. Before the surgeon even operates he or she knows precisely where the tumor is, how large it is, what kind it is, and what are the adjacent critical structures that need to be considered and avoided in attempts to remove the tumor.

Imaging is high resolution and fast and able to discriminate a brain tumor and other kinds of abnormal tissues from the normal tissues. This information is now used to plan the surgery. The big benefit of that is not only a better outcome to the patient, because the tumor is removed, but it also reduces morbidity, which means there’s less risk and less chance that you’ll do damage you don’t want to do in the operation.

It’s done faster because you know exactly what you’re going to do before you go in, there’s no time wasted on exploring around and figuring out what you’re going to do after you get into the operating site.

Targeting is a term we use that says we know not only where the bad guy is, we also know where the good guys are. That is, we have the bad tumor but we have the good tissue that has not been invaded by the tumor yet. I sometimes think of surgeons as being like real estate agents – everything to them is location, location, location.

It isn’t just knowing where the tumor is, it’s knowing where it isn’t, and where the other critical structures are – in the brain it could be language centers or cognitive centers, and you don’t want to accidentally cut or remove that brain tissue if you don’t have to.

This is a tremendous benefit to the patient, but it’s of tremendous benefit to the healthcare industry, too, because this whole procedure has taken less time, and it has less morbidity so you don’t have the same length of hospital stay. And the whole thing costs less.

I won’t get on a soapbox about this, but sometimes technology gets blamed for the high cost of healthcare and I think you have to be very careful about that presumption and evaluate what would have been the real cost of the morbidity and extended hospital stays if you didn’t have the technology. My view is that three-dimensional imaging technology helps in diagnosis and helps in therapy and also reduces healthcare costs.

Yes, these scanners cost hundreds of thousands of dollars – some more than a million dollars. They’re a big, expensive capital equipment item. But you have to compare that cost of the initial capital investment with the savings in the services that this instrumentation provides.

Adding up
Another exciting area is multi-modality imaging: any one imaging system does not provide all the information that we’re able to get by combining imaging systems. For example, a CT scan tells us something a little different about the brain than an MRI scan. If we can put them together, it’s truly a synergistic situation. One plus one is greater than two.

Two images together tell us a lot more about the normal brain and the abnormal brain than either image on its own. We call this process image fusion: putting multi-modality images of the same patient together so we get more information about what’s wrong with the patient or where the problem is, so that we can more effectively diagnose and treat the patient.

Let me give you an example. For a long time, patients with epilepsy who had intractable seizures that didn’t respond to medication had a low quality of life because they had several seizures a day. The only treatment that was effective was removal of the small parts of their brains that were causing these seizures.

The diagnostic and therapeutic challenge and problem is where is that in the brain? What part of the brain is having aberrant electrical activity giving rise to these seizures? It isn’t dense like a tumor, so if you use a CT scan or an MR scan, it will look like normal brain tissue because it’s really electrical activity in a piece of brain tissue. The question is, how do you find that?

We do a functional image scan with a PET scanner or a SPECT scanner. These imaging systems show us function in the brain rather than morphology or anatomy.We inject a radionuclide into the bloodstream and while that radionuclide is circulating in the blood in the brain we take an image with a SPECT or a PET scanner, and it finds out where this radioactive tracer lodges in the brain. It turns out that if you inject this in a patient with epilepsy shortly after a seizure, this radionuclide bunches up and collects in the region of the brain that’s over-excited because of the increased blood flow there, and this is the region of the brain that’s causing the epileptic seizure.

This is a low-resolution image and it’s not an anatomic image. It’s like a flashlight in a tunnel. It’s like looking down in a cave and you see a flashlight somewhere – you’re in the brain and you see where the problem is, but you can’t see the rest of the brain. That’s what this functional image shows us. With fusion, we take that image and register it to combine it with an MR image, which gives us exquisite information about the anatomic detail of the brain.

We superimpose this flashlight, this region of interest, through our mathematical algorithms directly on to the brain surface where that radionuclide collected during the scan after the seizure. Now we have an image that we can use to show the neurosurgeon exactly what piece of brain tissue to remove.

Image fusion is a technique that was developed by my lab, in working with the neuroradiologists, neurologists and neurosurgeons here at Mayo Clinic in the 1990s. It is used routinely three or four times a week to treat a lot of pediatric patients successfully with highly improved outcomes and with lower costs. It is now used in almost every major medical center in the world to treat patients with intractable epilepsy.

Making models
There is another wonderful technology that has come about in the last decade, which we call 3D printing. It’s an actual printer, like you would attach to your computer, except it prints three-dimensional models. It makes models out of special materials, and the input to this printer is not a string of alphabetic characters or numbers, it’s the surfaces of an object that we have segmented from a three-dimensional CT scan.

For example, it could be the brain, it could be the liver, it could be the heart, it could be part of the skeleton system, like the spine, or any part of the body in which the surgeon is interested in operating.

Pre-operatively, we can take the CT scan not only to plan how to do the surgery but to see a replica of that patient’s skeleton. Let’s say a spine surgeon has a patient with a serious fracture in the lumbar spine and he’s going to put in screws and plates and braces to shore up the damage.

With a CT scan not only can he plan how to do that using the images on the computer screen, but we can print an exact model of that patient. We call this patient-specific imaging. We can make a life-sized model on this 3D printer. We feed it the CT, we process the CT scan where we have segmented out the bone: all of the spine in the region where the spine surgeon is going to operate.

We send that data to this 3D printer and it takes about an hour, depending on how large the model is, for it to lay down very thin layers of this special material. The material is soft, and when we take it out it takes a while to cure and dry. Then it’s an exact replica of that patient’s anatomy.

How does the surgeon use that? There are a variety of ways. Sometimes he or she just looks at it to confirm his or her suspicions. It’s one thing to look at a three-dimensional image on a computer screen, and it’s quite a different thing to hold in your hands the actual spine or liver or heart of that patient and understand how big it is, where the grooves are, where the curves are, where the damage might be, and see exactly and hold exactly what you’re going to see when you open up the patient to treat the condition. It’s a very valuable part of the rehearsal for the operation.

Let me give you another example. My lab has been very fortunate, on five different occasions since the early 1990s, to be involved with a Mayo team of physicians, radiologists, scientists and engineers in separating conjoined twins.

And imaging, which was never designed specifically for the purpose of aiding in the process of planning a separation of conjoined twins, is in fact the primary modality and capability that we have to make these surgeries successful in every instance, and the end of the story is these were successful surgical separations at Mayo.

What our laboratory did was take these twin babies’ scans while they were together in the scanner and process them in many of the same 3D ways that I’ve been describing, to show the pediatric surgeons exactly what the babies looked like inside, how they were joined, where they were joined, what organs they might share and how far away certain vital structures were in one baby versus the other.

By contrast
Contrast material always plays an important role in the imaging. Contrast materials may have certain toxicities and disadvantages, but we do use them. They are used to tag tissues, when we can’t see with our normal imaging devices the things that the contrast material will enhance in the image.

Examples might be iodine injected in the bloodstream to see the coronary arteries, or gadolinium injected into the bloodstream to highlight tissues with MRI scanning. Or the radionuclide tracer that will flow through the body and preferentially be taken up by excited tissues, which we couldn’t see otherwise without the injection of the contrast material.

There are times when we don’t have to use contrast at all, but sometimes it’s still necessary and there has been an improvement in contrast materials so that you can use low doses of the contrast, and it tags the information, the tissues, the functions that you want to see better than previous contrast materials or higher doses. This is a very important advance in medicine.

The whole field of biomarkers is also exciting, where you selectively inject substances into the body and they only light up when they get to their target. They’re targeted to certain kinds of cells, certain kinds of biological elements inside the body, and these substances roam around in the body until they find that and then turn on the flashlight and we can image that.

If you think of the surgeon as a person who has to navigate through the body in three dimensions, going around, under and missing things to get to the target, then knowing where the target is is one thing. You know it’s in there and you can figure out a path to it, but you have to get to it efficiently.

That oversimplification explains how 3D imaging helps reduce the procedure time because you don’t have to spend much time figuring out the most safe path or running into things that you didn’t expect and having to back out and do something else. You have this precise 3D roadmap of pathways that you can pursue that will get you to the target efficiently and safely. And this reduces time.

That reduction in time not only improves the outcome, but it also reduces the risk of collateral damage, which historically has always been the challenge of surgery; even in a good surgeon’s hands sometimes that happens.

Improved  care
The way digital imaging technology has improved patient care compared to non-digital technologies like film, is quantitation – numerical, quantitative assessment of the patient’s disease.

The second thing is that it’s 3D, and even 4D. Instead of seeing a 2D representation or projection like an X-ray film has, we have the true three-dimensional relationships of all of the organs and structures in the body that are provided by the imaging scanner.

And the third is the digital image means that we can use the power of computers and the creativity of human beings who know how to program those computers to visualize that anatomy and pathology in ways that cannot be visualized or seen with plain film, with X-rays or non-digital techniques.

It provides these wonderful opportunities for what I call computer-aided diagnosis and image-guided interventions, and both of those things have significantly impacted and improved clinical diagnosis and treatment.

Images are often used in therapy monitoring, too. You’re not necessarily done imaging the patient after you’ve treated them, especially if they have cancer or another serious condition and they come back, or if you didn’t treat them with surgery or they were given a drug or radiation to try to reduce the tumor or remove the tumor. The way you assess that treatment regimen over time is through imaging.

You image the area again to see what the effects of the treatment are. This means that if the treatments aren’t effective, we don’t have to submit patients to a long drug trial, for example, or have them stay on the same drug that isn’t making any difference. We image them every month or six weeks or whatever the protocol requires and if it’s not working we change the dose, or change the drug or change the radiation orientation. Imaging plays a significant role in deciding if whatever therapy is chosen is effective and then modifying that therapy if necessary.

What has driven me and my laboratory in all of this over the years is a comment that a neurosurgeon made to me when I came to Mayo Clinic in the early 1970s, when this type of 3D imaging first became available here in the US.

I began working with imaging in Dr. Earl Woods’ lab here at Mayo, and I remember showing a famous neurosurgeon, Thoralf Sundt, these images and how we might be able to show him tumors in 3D. They weren’t very good quality then, but he got it right away, and he said to me something that I’ll never forget: “If I can see it, I can fix it.”

That became a mandate for my career. I thought, “Okay, I want to show it to you. I want to show it to you better and better and better and more often and accurately, so you can fix it.”

Richard Robb is Director of Mayo Clinic’s Biomedical Imaging Resource. He has been involved in the development and application of computer systems for processing, analysis and display of biomedical image data for more than 30 years.

© Images provided courtesy of Mayo Clinic