AAPM: Wide-area detector CCTA technology opens doors of possibility
PHILADELPHIA—What began as single-slice CT scanning has now evolved into dynamic volume scanning, utilizing as many as 320 detectors. The possibilities for wide-area detector technology include consistently low radiation exposure for coronary CT angiography (CCTA), as well as myocardial perfusion imaging, kinetic opacification slopes that distinguish between normal and diseased arteries and vascular profiling, according to Frank Rybicki, MD, PhD, who spoke July 21 at the annual meeting of the American Association of Physicists in Medicine (AAPM).
Conventional 64-slice retrospectively-gated CCTA scans the heart over the course of six to seven heartbeats and delivers on average 15 mSv of radiation exposure. In addition, the overlapping imaging produces stair-step artifacts.
The step-and-shoot technique, a prospectively-gated method that turns the x-ray tube on and off during assigned phases, has reduced the dose by half. The one disadvantage of this method, however, is it does not deliver functional information.
With wide-area detector row scanners, retrospectively-gated acquisition can occur over a single heartbeat, eliminating stair-step and motion artifacts. Additionally, cardiac function can be evaluated. "We do this in about 10 percent of patients, when we have to look at a valve or at a moving cardiac structure," said Rybicki, director of cardiac CT and vascular CT/MRI at Brigham and Women's Hospital in Boston.
When Rybicki and colleagues first acquired the 320-row scanner (Toshiba America Medical Systems), the average radiation exposure was around 12 mSv to 13 mSv. They cut the dose in half by tweaking several parameters including those associated with single and multiple segment reconstructions.
To reduce the dose even further, imagers can lower the kV and have dose in the range of 2 mSv. But unless the patients are thin, these images tend to be noisy.
So, Rybicki and colleagues looked at manipulating the phase window setting. "In prospective gating, if you narrow the window down even 1 percent so you have only one opportunity to look at the coronary windows, you will lower dose." Researchers retrospectively examined 41 patient scans that had a relatively wide window phase acquisition, between 60 to 95 percent. They reconstructed them with the window phase narrowed to 1 percent. While the dose was lowered, image quality was excellent in only two-thirds of the scans. They then increased the window phase to 10 percent, which increased the dose somewhat, but still lower than the original scans, and the rate of excellent image quality jumped to 93 percent.
They did not get a noticeable increase in the rate of excellent image quality at 20 percent. "The real power of coronary CT is in its negative predictive power. We want to exclude coronary artery disease without having to perform a catheterization. The 10 percent phase window is the right way to go," he said.
The next step was to capitalize on some efficiencies gained by cone beam processing, meaning the data acquired to do the correction. The original version of the software had symmetric acquisition, that is, data on each end of the image acquisition was acquired. "The more efficient approach has a shorter overall asymmetric acquisition, gaining about a 24 percent reduction in exposure," he said.
They looked at 13 patients pre- and post-cone beam software upgrade and found a reduction in dose from about the 4.5 to 5 mSv range to the 3.5 to 4 mSv range. For reasons unknown, they also get better noise profiles with the asymmetric cone beam processing. He challenged the physicists in the audience to figure out why.
The Brigham group has also researched using the 320-detector row scanner to perform myocardial perfusion imaging. The radiation dose is increased because you have to repeat the CT scan after the administration of adenosine. But the defects seen are akin to those seen in an MR or nuclear scan. "This is an example of using wide-area detector CT in a novel way," he said.
In addition, the adenosine increases the heart rate so the scan is captured over two heartbeats, rather than a single beat. "We have to do a two-beat reconstruction to have the proper temporal resolution."
Whether CT myocardial perfusion imaging will overtake SPECT imaging remains to be seen. The CORE 320 trial is a multicenter study designed to test the ability of stress perfusion CT to contribute to a patient's care. CT is being compared with stress SPECT.
"There are two hypotheses. CTA with perfusion is equivalent to catheterization and SPECT, and CTA with perfusion is better than cath alone," he said.
Another point Rybicki made was about the scanner's ability to distinguish gradients in contrast opacification. With conventional CT scanners, the contrast opacifies over several heartbeats. With the wide-area detector row scanner, it happens in one beat. Because of these physics, they can measure the kinetics of the contrast in the coronary arteries.
They have shown that opacification can be measured with respect to distance from the coronary ostium, lumen cross-sectional area and lumen short-axis diameter. These measurements are reproducible over a wide variety of patients and contrast protocols. Researchers also have shown, albeit in a small number of patients, that the opacification kinetics differ between normal and diseased arteries.
The next area of research is to determine whether the contrast enhancement pattern changes correlate to pressure changes, such as those measured by fractional flow reserve (FFR). If so, lesions can be interrogated noninvasively by CT for physiologic relevance prior to going into the cath lab.
The last area Rybicki discussed was vascular profiling, meaning using the technology to better understand the role shear stress plays in the progression of noncalcified lesions. "We know that atherosclerosis and remodeling are driven by low shear stress, but most data associated with shear stress are obtained with intravascular ultrasound. We would like to do that with CT," he said. "We can look at branches and bifurcations. The only limitation is the spatial resolution of CT, which is not as good as that with IVUS."
In their work, the Brigham researchers have shown that CT imaging produces very similar patterns of low shear stress compared with IVUS. Mapping out areas in individual vessels from 13 patients, they found a sensitivity and specificity of 70 percent for CT to identify low shear stress areas. "That's not bad considering the difference in spatial resolution between CT and intravascular ultrasound," Rybicki said.
Conventional 64-slice retrospectively-gated CCTA scans the heart over the course of six to seven heartbeats and delivers on average 15 mSv of radiation exposure. In addition, the overlapping imaging produces stair-step artifacts.
The step-and-shoot technique, a prospectively-gated method that turns the x-ray tube on and off during assigned phases, has reduced the dose by half. The one disadvantage of this method, however, is it does not deliver functional information.
With wide-area detector row scanners, retrospectively-gated acquisition can occur over a single heartbeat, eliminating stair-step and motion artifacts. Additionally, cardiac function can be evaluated. "We do this in about 10 percent of patients, when we have to look at a valve or at a moving cardiac structure," said Rybicki, director of cardiac CT and vascular CT/MRI at Brigham and Women's Hospital in Boston.
When Rybicki and colleagues first acquired the 320-row scanner (Toshiba America Medical Systems), the average radiation exposure was around 12 mSv to 13 mSv. They cut the dose in half by tweaking several parameters including those associated with single and multiple segment reconstructions.
To reduce the dose even further, imagers can lower the kV and have dose in the range of 2 mSv. But unless the patients are thin, these images tend to be noisy.
So, Rybicki and colleagues looked at manipulating the phase window setting. "In prospective gating, if you narrow the window down even 1 percent so you have only one opportunity to look at the coronary windows, you will lower dose." Researchers retrospectively examined 41 patient scans that had a relatively wide window phase acquisition, between 60 to 95 percent. They reconstructed them with the window phase narrowed to 1 percent. While the dose was lowered, image quality was excellent in only two-thirds of the scans. They then increased the window phase to 10 percent, which increased the dose somewhat, but still lower than the original scans, and the rate of excellent image quality jumped to 93 percent.
They did not get a noticeable increase in the rate of excellent image quality at 20 percent. "The real power of coronary CT is in its negative predictive power. We want to exclude coronary artery disease without having to perform a catheterization. The 10 percent phase window is the right way to go," he said.
The next step was to capitalize on some efficiencies gained by cone beam processing, meaning the data acquired to do the correction. The original version of the software had symmetric acquisition, that is, data on each end of the image acquisition was acquired. "The more efficient approach has a shorter overall asymmetric acquisition, gaining about a 24 percent reduction in exposure," he said.
They looked at 13 patients pre- and post-cone beam software upgrade and found a reduction in dose from about the 4.5 to 5 mSv range to the 3.5 to 4 mSv range. For reasons unknown, they also get better noise profiles with the asymmetric cone beam processing. He challenged the physicists in the audience to figure out why.
The Brigham group has also researched using the 320-detector row scanner to perform myocardial perfusion imaging. The radiation dose is increased because you have to repeat the CT scan after the administration of adenosine. But the defects seen are akin to those seen in an MR or nuclear scan. "This is an example of using wide-area detector CT in a novel way," he said.
In addition, the adenosine increases the heart rate so the scan is captured over two heartbeats, rather than a single beat. "We have to do a two-beat reconstruction to have the proper temporal resolution."
Whether CT myocardial perfusion imaging will overtake SPECT imaging remains to be seen. The CORE 320 trial is a multicenter study designed to test the ability of stress perfusion CT to contribute to a patient's care. CT is being compared with stress SPECT.
"There are two hypotheses. CTA with perfusion is equivalent to catheterization and SPECT, and CTA with perfusion is better than cath alone," he said.
Another point Rybicki made was about the scanner's ability to distinguish gradients in contrast opacification. With conventional CT scanners, the contrast opacifies over several heartbeats. With the wide-area detector row scanner, it happens in one beat. Because of these physics, they can measure the kinetics of the contrast in the coronary arteries.
They have shown that opacification can be measured with respect to distance from the coronary ostium, lumen cross-sectional area and lumen short-axis diameter. These measurements are reproducible over a wide variety of patients and contrast protocols. Researchers also have shown, albeit in a small number of patients, that the opacification kinetics differ between normal and diseased arteries.
The next area of research is to determine whether the contrast enhancement pattern changes correlate to pressure changes, such as those measured by fractional flow reserve (FFR). If so, lesions can be interrogated noninvasively by CT for physiologic relevance prior to going into the cath lab.
The last area Rybicki discussed was vascular profiling, meaning using the technology to better understand the role shear stress plays in the progression of noncalcified lesions. "We know that atherosclerosis and remodeling are driven by low shear stress, but most data associated with shear stress are obtained with intravascular ultrasound. We would like to do that with CT," he said. "We can look at branches and bifurcations. The only limitation is the spatial resolution of CT, which is not as good as that with IVUS."
In their work, the Brigham researchers have shown that CT imaging produces very similar patterns of low shear stress compared with IVUS. Mapping out areas in individual vessels from 13 patients, they found a sensitivity and specificity of 70 percent for CT to identify low shear stress areas. "That's not bad considering the difference in spatial resolution between CT and intravascular ultrasound," Rybicki said.