CT applications expand through thick and thin
Increasingly thin slices speak volumes about CT's cardiac potential
By: Deborah R. Dakins
Since its introduction 30 years ago, CT has evolved from a primarily anatomic imaging tool to one capable of volumetric scanning. The impact of that change has been felt in everything from clinical applications to workflow, and it has even affected the CT vernacular.
Today's multidetector scanners represent a huge technical leap over the first dedicated brain imager unveiled in 1972. The clinical advances made possible by these devices, which are capable of acquiring four CT slices per tube rotation, have been impressive. Foremost among them is CT angiography. New 16-slice units in production by GE, Siemens, Philips, and Toshiba promise to push CT into a new frontier: the heart.
"CT angiography is the biggest clinical advance made possible by MDCT technology," said Dr. Elliot K. Fishman, director of diagnostic radiology and body CT at Johns Hopkins Hospital. "In our practice, we are doing 15 to 20 of them per day now. The clinicians love it, we like doing it, and it's excellent patient care."
Sixteen-slice scanners are only going to improve that picture, Fishman said. At Johns Hopkins, CTA with a four-slice scanner has already replaced more invasive studies in applications ranging from oncology to atherosclerotic disease.
"With 16 slices, the things we do now we'll do even better, because we'll have higher resolution and faster speeds, so runoff studies become very reasonable to do," he said. "The possibilities for cardiac scanning also expand. Cardiac with four slices is hard; it's not fast enough. With 16 slices, coronary angiography becomes a real possibility."
CT THEN AND NOW
One need not look far to appreciate the dramatic difference in capability between the CT of yesterday and today. By comparison, today's multidetector units offer as much as an eight-fold increase in speed over single-slice systems. A lumbar CT study that might have taken 45 minutes 10 years ago can now be completed in less than 10 seconds.
CT's introduction sparked a new era in the way radiologists evaluate patients. Slices that cut through the body augmented 2D coronal images, common in both radiography and nuclear medicine. These slices offered immediate benefits, allowing multiple points to be recorded from different angles. A series of digitally stored slices could be reconstructed in multiple views.
CT quickly earned a niche as the primary technique for abdominal disorders and the mainstay of body and chest imaging. Its specialties included the study of cortical bone, examining fractures of the temporal bone and pelvis, and providing high-contrast spatial resolution in the upper abdomen. By the late 1980s, cine CT had emerged as a cost-effective way to evaluate cardiac structure, function, and reperfusion.
Prior to MDCT, the last major innovation in CT came with the introduction of spiral scanning in 1989. This moved CT from step-and-shoot technology, in which patients were moved incrementally between slices, to continuous data acquisition in a spiraling pattern as patients slid through a rotating gantry.
The ability to acquire a constant flow of data was CT's first step toward volumetric data collection-and represented the first shift in thinking about CT data-away from "slices" and toward the concept of volume.
CENTER STAGE: ANGIOGRAPHY
Spiral scanning enjoyed a heyday through most of the 1990s. It changed the utility of CT by opening the door to CTA, which has evolved exponentially since the introduction of quad-slice scanners in 1998.
Prior to MDCT, there was only one way to scan an abdomen, said Dr. Jay Cinnamon, vice chair of radiology and chief of neuroradiology at Long Island Jewish Medical Center in New Hyde Park, NY. The versatility of these devices and the numerous factors that can be manipulated-pitch, gantry rotation time, slice thickness, reconstruction increment, total scan time, reconstruction algorithms, image enhancement, and table speed-create multiple options.
"There's no question that vascular imaging has benefited the most
from this new technology," said Dr. Geoffrey Rubin, chief of cardiovascular imaging and codirector of the 3D Laboratory at Stanford University. "We can now image areas of the body that we could not image sufficiently with CT in the past, such as the upper and lower extremities."
New 16-slice units, with which Rubin has logged research experience, will only serve to increase CT's vascular expertise.
Extremity studies conducted at Stanford in 2000 using four-slice scanners required up to 70 seconds and resulted in sections measuring 2.5 mm, Rubin said. That's both longer and thicker than his group would prefer to use, and it meant stretching the contrast bolus.
"With the new scanners, we're routinely acquiring 1.25-mm-thick sections, and the scans are completed in about 20 seconds," he said.
Another difference: With vessels that are even smaller than those in the lower extremities, upper extremity imaging couldn't even be attempted with four-slice technology.
"We needed higher spatial resolution than was possible with four-row," Rubin said. "With 16-slice, we have it. We frequently need to scan from the heart through the fingertips. These are new applications that we just could not do before."
MSCT IN PRACTICE
At sites like Stanford, Hopkins, and Case Western Reserve University- all early adopters-MDCT is the technique of choice for most cardiovascular cases. From virtual endoscopy of the aorta to surgical reconstruction or evaluation of pulmonary embolism, radiologists swear by the fast and consistent imaging possible with MDCT.
But that's not the case everywhere, Fishman said.
At radiology gatherings across the country, Fishman has found, only about 15% of the physicians with MDCT at their facilities say they are performing CTA. The reasons are primarily technology-related. To do CTA well involves more than the exam itself; postprocessing of the data is key, he said.
"People have been uncomfortable with some of the workstations and software that perform 3D processing," he said. "That's because some of the early software was not that good, or because people have had bad experiences with poor software, and they are reluctant to get involved with it now."
But strong hardware platforms offer real-time rendering and are fast and relatively easy to use, he said.
NEXT FRONTIER
The ultrafast subsecond speeds of four-slice CT rotations make it possible to isolate and image single cardiac cycles. Inherent cardiac motion artifact is markedly reduced. Improved capability to perform single breath-hold scans has opened a new frontier for CT in the heart.
Coronary artery scanning is the first target. Research conducted in Germany has found MDCT just as accurate as electron-beam CT for detecting severe stenosis in the coronary arteries. But MDCT still lags behind conventional invasive coronary angiography in imaging both distal segments and vessels that move a great deal, such as those serving the heart.
"In imaging the heart, the greatest interest clinically is in the coronary arteries," Rubin said. "There is currently no effective way to noninvasively image coronary artery disease. It's a huge opportunity, and MDCT has shown great promise."
To date, however, limitations have dogged four- and eight-slice technology in terms of imaging quality and consistency, barring its ability to supplant conventional angiography. It's still unclear whether 16-slice scanners will solve the problem, Rubin said.
"One development we're still waiting for is something that will substantially improve temporal resolution so that coronary arteries can be captured in a time frame on the order of 100 msec, rather than what is possible currently, about 250 msec," he said.
Ultimately, proving MDCT's superiority to conventional angiography in the coronaries will be a tricky challenge, Fishman said.
"In peripheral disease, you can prove superiority because you have an initial angiogram or surgery," he said. "The only way to get proof with the coronaries is to have coronary angiography to prove MDCT is better."
CONEBEAM THEORY
The number of scanning detectors a CT device has directly affects slice thickness--or thinness, as the case may be. But in some applications, such as the heart, thicker may be better.
In shifting from four-detector rows to eight to 16, manufacturers have created detectors with thinner segments without increasing overall detector width. That results in the ability to acquire narrower sections faster, Rubin said.
"But we are not able to get thicker sections faster," he said. "It would be nice to have a scanner that could image the entire heart, which is about 12 cm, in every single rotation, at maybe 5-mm-thick sections. You could do that with a 24-row scanner."
As vendors work on solving these problems, the stage is being set for the next generation of CT scanners, featuring a more advanced reconstruction capability known as conebeam. These scanners will use large-area detectors made from solid-state arrays based on the same flat-panel technology employed in digital radiography. The impact of such devices on CT's clinical applications could be significant.
Experimental work with conebeam scanners boasting a wide field-of-view is conceivable within a five-year time frame, he said. The resulting scanner will allow capture of very high resolution data sets. In addition, this flexible device will be capable of performing both radiography and fluoroscopy.
"Incorporating these novel flat-panel detectors will revolutionize the CT scanner," Rubin said.
CHANGING TERMINOLOGY
Like the shift from step-and-shoot imaging to spiral CT, upgrading to MSCT has prompted changes in workflow, image storage, and even the way radiologists refer to CT data.
"In the past, if you bought a new scanner, it was a little better than the one before, and you didn't have to change much," Fishman said. "With multidetector CT, you change everything. Protocols change. The speed of scanning changes. Applications, throughput, and the need for postprocessing change." Tools such as 3D and volume rendering are no longer options with MSCT. They become mandatory, if for no other reason than to adequately manage the image data and communicate them to referring physicians.
"A referring physician is not going to look at 500 images," he said. "They want the information. And you have to be able to give them the information in the most efficient way. More is not necessarily better, if you can't use it."
The large data sets that make MSCT an exemplary clinical tool are creating challenges for archiving, access, and display. With most PACS, it's not possible to cycle through hundreds of cross-sectional images per MSCT study.
"You can no longer talk about slices," Fishman said. "If CT was invented today with a four-slice or a 16-slice device, you would talk about volume visualization; no one would be thinking about slices."
Ms. Dakins is a freelance writer in Ben Lomond, CA.
---
A brief history of CT: From brain scanner to volume rendering
Mid-1970s
Versatile whole-body scanners replace dedicated brain imagers, spurring market adoption.1971
Researcher Godfrey Hounsfield of U.K.-based EMI invents CT scanner.
1972
CT scanners are introduced at RSNA meeting.
1978
Annual CT sales in U.S. begin dramatic six-year rise, from 200 systems annually to nearly 800 in 1983.
1979
Nobel Prize in medicine is awarded to two early CT researchers.
1980
Researchers advance CT's clinical standing by demonstrating its accuracy in lung cancer diagnosis. CT installed base reaches 5000 units.
1981
CT makes headlines during trial of John W. Hinckley when scans showing brain atrophy are used to diagnose the shooter's schizophrenia.
Early to mid-1980s
CT acquisition speeds decrease from five minutes to 20 seconds, thanks to fanbeams, array detectors, etc.
1985
Industry experts predict U.S. market for CT is nearing saturation. Researchers at Johns Hopkins develop volume rendering techniques.
1986
Despite being overshadowed by MRI, CT holds its own in studying cortical bone and fractures, and in providing high contrast and spatial resolution in the upper abdomen.
1989
Spiral scanning is introduced. CT remains primary screening technique for abdominal disorders and the mainstay for body and chest imaging.
Mid-1990s
Radiologists hone 3D applications. Spiral scanning technology evolves, migrating from high-end to midrange tiers; end of the spiral era is in sight.
1992
Elscint debuts CT Twin, the first multislice scanner. The multislice concept languishes until 1998.
1998
GE, Toshiba, Marconi, and Siemens all unveil versions of quad-slice CT scanners with subsecond rotation.
2000
Multidetector CT units offer eight-fold increase in speed. A study that might have taken 45 minutes a decade ago can be completed in under 10 seconds.
2001
Radiologists make dramatic clinical leaps with four-slice scanners; CT angiography emerges as the biggest beneficiary.
2002
Eight- and 16-slice units begin shipping.
