CT takes different directions to tackle age-old problems

August 5, 2008

The future of CT used to be easy to predict. So-called slice wars dominated the discussion as vendors wheeled out their latest multidetector machines. CT systems were ranked according to a kind of geometric progression: single-slice, two-slice, four-slice, 16-slice . . . It was a simple form of classification and one that was easy to appreciate.


The future of CT used to be easy to predict. So-called slice wars dominated the discussion as vendors wheeled out their latest multidetector machines. CT systems were ranked according to a kind of geometric progression: single-slice, two-slice, four-slice, 16-slice . . . It was a simple form of classification and one that was easy to appreciate.

Today the picture is less straightforward. The numbers game is not yet finished, but comparisons cannot be based on detector rows alone. Assumptions of what a CT scanner should look like and how the modality should behave are being challenged from all directions.

"It is a fun time to be working in CT," said Dr. Konstantin Nikolaou, associate chair of radiology and section chief of CT at the University Hospital Grosshadern in Munich. "There are new things coming up all the time."

The continuing evolution of CT technology will ultimately be dictated by what radiologists want and how much they want it. Vendors could undoubtedly create the perfect, must-have scanner today. The price tag would be so high, however, that the system would never see the inside of a working clinical radiology department.

"The technology is moving very fast, and there are still more new developments on the horizon," said Dr. Christian Fink, head of CT at the University Medical Center Mannheim, part of the University of Heidelberg in Germany. "Let's see what they actually bring to clinical radiology."


Today's 64-slice systems may already have brought radiologists exactly what they asked for: sufficiently high spatial and temporal resolution to convert potential applications into tried and trusted diagnostic techniques. This status can be seen clearly in abdominal imaging, the bread and butter of any department's CT work list, according to Dr. Elliot Fishman, director of diagnostic imaging and body CT at Johns Hopkins Hospital in Baltimore, Maryland.

"To detect the source of gastrointestinal bleeding was very hard before we had 64-slice CT, whereas now it seems to be a robust study," he said.

Faster scanning and improved spatial resolution have benefited other vascular imaging applications as well. Smaller aneurysms can be detected and finer blood vessels studied. Tumor vascularity can also be quantified, with a view to determining the degree of angiogenesis.

Experts agree, however, that the main beneficiary of next-generation CT has been cardiac imaging. The higher scanning speeds and increased spatial and temporal resolution have greatly improved the quality of coronary CT studies.

"With 64-slice systems, cardiac CT has become a clinical tool that hospitals and clinics, large and small, are using routinely," Fishman said.

The launch of dual-source CT technology has upped the ante even further. In the first departure from the slice wars, Siemens has given radiologists a machine with not one but two x-ray tubes. With both tubes running at the same energy, the temporal resolution of a cardiac scan is effectively halved. The second tube can be turned off for routine applications and the system treated like a regular 64-slice scanner.

"The improved temporal resolution [of] dual-source CT holds a lot of promise for cardiac imaging," said Dr. Geoffrey Rubin, chief of cardiovascular imaging at Stanford University in Palo Alto, California. "As with any new technology, though, we have to be careful in touting new applications before they have been rigorously vetted in a scientific manner."


The availability of CT scanners with two x-ray tubes has reawakened interest in energy-dependent imaging. This technique involves separating the information gained on imaging from x-ray photons of differing energies.

Characteristic attenuation profiles are then used to separate materials such as bone, soft tissue, and contrast media.

With two x-ray tubes running simultaneously, one can be operated at a low kV and the other at a higher kV. First publications indicate the benefits of dual-energy CT in various clinical fields, such as the characterization of kidney stones and kidney tumors or the visualization of myocardial ischemia.

Not everyone is convinced that dual-energy imaging requires two x-ray tubes. At least two vendors, GE and Toshiba, are reinvestigating the early idea of rapidly switching the energy of a single x-ray source. Meanwhile, researchers from Philips are focusing on a type of layered detector that separates low- and high-energy x-rays produced by a single source. An alternative solution could incorporate photon-counting technology into the x-ray detector, allowing the energy of each photon to be measured individually.

"One would expect this to be a good approach," said Dr. Norbert Pelc, a professor of bioengineering and radiology at Stanford. "If the detector could count quickly enough and perform its energy discrimination fast enough, it would give the best performance in terms of delivering precise, energy-dependent information."

Dr. Ioannis Vlahos, an assistant professor of radiology at New York University Hospital, accepts that dual-energy imaging has yet to filter through to the wider radiology community. As more results appear, he expects that the clinical value of dual-energy applications will outweigh any lingering uncertainties.

"These techniques can be a bit overwhelming when people look at them for the first time," he said. "But I think in a few years, everybody will be talking about the applications of dual-energy imaging."

Others are more cautious, saying that the technology has yet to prove itself.

"The actual success of the technique could be impacted quite a bit by the approach that is being used," Rubin said.



These new developments and new directions don't necessarily mean an end to CT's slice wars. For evidence, look no further than the vendors' stands at RSNA 2007. Philips used this platform to unveil a 256-slice CT scanner, using 128 collimated detector slices and a z-flying focal spot, while Toshiba captured delegates' attention with a work-in-progress 320-slice machine. Size clearly still matters for CT developers.

A key selling point of these latest CT systems is the size of the detection area. Toshiba's 320-slice scanner, for example, can cover up to 16 cm of anatomy. This makes it feasible to image an entire organ, such as the brain or the heart, in a single rotation lasting one second or less.

One obvious beneficiary of wide area imaging will be perfusion CT. Radiologists will be able to track contrast as it moves throughout an organ without moving the table. The amount of contrast required could also be reduced if, for example, arterial, venous, and whole-organ perfusion exams can be performed in a single study.

This new breed of CT scanner has yet to convince everyone. Some radiologists have queried the need for such machines, given the cost of the detector technology. Others have questioned the quality of reconstructed images. They point to the likelihood of conebeam artifacts where data to reconstruct the edges of the axial volume are insufficient. Although algorithms designed to deal with these artifacts have been developed, skeptics will not be convinced until they have seen the results. Another problem of large area detectors is scattered radiation that degrades the signal-to-noise ratio of the images.

"We are still waiting to see whether you can employ a 256- or 320-slice scanner for general CT applications without compromising imaging quality," said Dr. Dennis Foley, a professor of radiology at the Medical College of Wisconsin in Milwaukee.

Spiral scanning remains the best option for attaining high-quality images over lengths greater than 10 cm, said Dr. Willi Kalender, director of the Institute of Medical Physics at the University of Erlangen in Germany. You are asking for trouble if you go for further and further extensions of detector coverage," he said. "This seems to be confirmed by the fact that standard spiral scans on the 320-slice CT system are performed in 64-slice mode."

One solution to the problem of cone-angle artifacts could be turning the technology on its head. Researchers at Stanford, in collaboration with GE Healthcare, are investigating an alternative method of wide field-of-view imaging that uses a wide array of x-ray sources and a much smaller detector array.

The idea behind the inverse-geometry CT (IGCT) system is to switch the x-ray point sources on and off rapidly while the array is rotated around the patient. Because the detector and x-rays have the same axial extent, conebeam artifacts should no longer be an issue.

"We have received funding from the U.S. National Institutes of Health to build a test IGCT system that will be mounted on a rotating gantry and will be capable of performing in vivo scanning," Pelc said.

Another option is to keep the spiral scanning mode but move patients back and forth through the gantry when performing perfusion CT or dynamic CT angiographic studies. Coverage length can then be tailored to the organ of interest, say, 10 cm for the brain or 25 cm for the liver. This is what Siemens is proposing with its work-in-progress adaptive scanning technology.


What is it, then, that radiologists want from CT? Despite clear gains in temporal resolution, faster scanning still tops the wish list. This appetite for speed is being driven by a desire for further improvements in cardiac CT.

Practitioners welcome the boost in image quality that 64-slice CT and dual-source scanning have brought but believe there is still room for improvement.

Radiologists fully expect that this demand will be met. Today's CT machines rotate with speeds that were previously unimaginable. So what is to stop tomorrow's CT systems from going that little bit faster still?

"I was always told when a new machine came out that it was now at its mechanical limit in terms of rotation, that it wouldn't be able to stand any greater forces," Becker said. "I was told this when the gantry was rotating around its own axis in 0.5 seconds. But the new machine we have has a rotation speed of 0.3 seconds."

CT experts will not be surprised to see the launch of more systems with two x-ray tubes. As systems become more complex, however, radiologists will be paying increasing attention to reliability and maintenance issues.

The need for more sophisticated postprocessing tools will become increasingly pertinent as CT moves beyond 64-slice technology and data sets become still larger. Emerging applications such as dual-energy CT will also require user-friendly software.

"It doesn't make any sense if you acquire 4000 perfusion images of the liver but then can't easily and quickly get a color-coded image that tells you the difference in tumor perfusion compared with three weeks ago," Nikolaou said.

Future developments in CT hardware will increasingly be guided by concerns over patients' medical x-ray burden, according to experts in the field. As scanner technology has advanced, and issues with equipment overheating have been addressed, it has become all too easy to up the examination dose.

Vendors are clearly taking note of this message, as information on radiation reduction strategies has become de rigueur in marketing literature. Toshiba, for example, has been keen to stress that its 320-slice CT system produces less radiation than a 64-slice CT scanner. Advocates of dual-source CT have been equally quick to dispel the myth that two x-ray tubes means double the dose. Instead, several clinical publications have demonstrated that cardiac CT with a dual-source CT can be performed at equal or lower dose than with a 64-slice CT.

Paula Gould is a contributing editor to Diagnostic Imaging Europe.