Researchers test CTA post-processing techniques to quantify non-culprit lesions
Researchers have found that 64-slice CT angiography provides an accurate grade of stenosis for non-culprit coronary lesions in proximal coronary segments. They also determined that a cross-sectional area post-processing technique correlated best with quantitative coronary angiography to quantify non-culprit lesions.
Lead author Jonathan D. Dodd, MD, from the department of radiology at St. Vincent’s University Hospital in Dublin, Ireland, and colleagues from Harvard University, Germany and the Netherlands commented that the current understanding of the atherosclerotic process as a diffuse one necessitates a focus on non-culprit lesions when managing acute coronary syndrome (ACS) patients.
“Methods to grade non-culprit lesions are important because they allow optimal therapeutic stratification of patients, emphasize the necessity of staged procedures, and determine suitability of patients for coronary artery bypass surgery,” according to the study published in the August issue of the American Journal of Roentgenology.
Researchers evaluated 29 patients who presented with ACS and had non-culprit coronary lesions of ≥ 30 percent stenosis quantified on quantitative coronary angiography software (QCAPlus, Sanders Data Systems).
All patients received a CTA exam (Somatom Sensation 64, Siemens Medical Solutions) and researchers tested five post-processing techniques to determine which one correlated best with quantitative coronary angiography:
The best overall correlation among all CT techniques with quantitative coronary angiography was cross-sectional area (R = 0.56, p < 0.01, Spearman’s rank correlation). The finding corroborates previous work, according to the authors.
Researchers found that MIP yielded the smallest interobserver variability (kappa = 0.7, p < 0.001), which also corroborates previous findings. Interobserver agreement for MPR, coronary software diameter and area were 0.6, 0.62 and 0.57, respectively.
Dodd and colleagues commented favorably on the ability of the software to rapidly produce curved MPR images and quantify stenosis. However, limitations of the semi-automated software included an inability to correctly grade several partially calcified lesions, as well as ostial lesions.
They also found that the semi-automated software tended to underestimate and overestimated vessel and area diameters, respectively.
“Of all the CT techniques, the semi-automated software showed the overall weakest accuracy and performance and cannot be recommended based on the data of this study at this time,” the authors wrote.
Overall, CT correlated strongly with angiography to quantifying non-culprit lesions greater than 3 mm in diameter, but poorly with lesions smaller than 3 mm.
CT also had moderate to strong correlations for mixed and non-calcified plaque, but poor correlations for calcified plaque.
“The most important findings from our study are that cardiac 64-MDCT shows a strong correlation with quantitative catheter angiography for grade of stenosis of non-culprit lesions in the proximal segment coronary arteries and for mixed or non-calcified plaques,” the authors wrote. “Large, purely calcified plaques and lesions in more distal coronary segments [≤ 3 mm diameter] remained difficult to accurately quantify.”
In an accompanying editorial, Andre J. Duerinckx, MD, said the study is very promising preliminary work in the field of coronary CTA, and it opens the door for future research.
“Until the advent of coronary CTA, we had no easy approach to coronary plaque detection or characterization, except for invasive coronary angiography in combination with intravascular ultrasound,” Duerinckx wrote.
But positive vessel remodeling, indicative of early disease, seen by CTA or intravascular ultrasound often appears normal on catheter angiography, he said. Therefore, CTA holds great promise to detect and help manage non-culprit plaques.
He added that the limitations shown by Dodd et al of CTA in this respect indicate more research and development need to be done.
Lead author Jonathan D. Dodd, MD, from the department of radiology at St. Vincent’s University Hospital in Dublin, Ireland, and colleagues from Harvard University, Germany and the Netherlands commented that the current understanding of the atherosclerotic process as a diffuse one necessitates a focus on non-culprit lesions when managing acute coronary syndrome (ACS) patients.
 | |
 | |
 | |
 | |
 | |
| A. Quantitative coronary angiography shows 38% stenosis of non-culprit lesion in right coronary artery of 53-year-old man with acute coronary syndrome. B. Corresponding maximum-intensity-projection (MIP) CT image shows 45% stenosis. C. Multiplanar reformat (MPR) image shows 46% stenosis. Note that calcified plaque (arrow) was excluded from luminal diameter measurements. D. Cross-sectional luminal area analysis image shows 56% stenosis. Note that plaque (arrow) was excluded from luminal diameter measurements. E. Coronary artery software area analysis show 40% stenosis. Note software was unable to exclude some mixed plaque (arrow) from analysis despite manual window adjustments, resulting in some falsely high luminal measurements. Source: American Roentgen Ray Society |
Researchers evaluated 29 patients who presented with ACS and had non-culprit coronary lesions of ≥ 30 percent stenosis quantified on quantitative coronary angiography software (QCAPlus, Sanders Data Systems).
All patients received a CTA exam (Somatom Sensation 64, Siemens Medical Solutions) and researchers tested five post-processing techniques to determine which one correlated best with quantitative coronary angiography:
- maximum intensity projection (MIP)
- multiplanar reformat (MPR)
- cross-sectional area
- diameter derived from semi-quantitative coronary software, and
- area derived from semi-quantitative coronary software (Circulation, Siemens).
The best overall correlation among all CT techniques with quantitative coronary angiography was cross-sectional area (R = 0.56, p < 0.01, Spearman’s rank correlation). The finding corroborates previous work, according to the authors.
Researchers found that MIP yielded the smallest interobserver variability (kappa = 0.7, p < 0.001), which also corroborates previous findings. Interobserver agreement for MPR, coronary software diameter and area were 0.6, 0.62 and 0.57, respectively.
Dodd and colleagues commented favorably on the ability of the software to rapidly produce curved MPR images and quantify stenosis. However, limitations of the semi-automated software included an inability to correctly grade several partially calcified lesions, as well as ostial lesions.
They also found that the semi-automated software tended to underestimate and overestimated vessel and area diameters, respectively.
“Of all the CT techniques, the semi-automated software showed the overall weakest accuracy and performance and cannot be recommended based on the data of this study at this time,” the authors wrote.
Overall, CT correlated strongly with angiography to quantifying non-culprit lesions greater than 3 mm in diameter, but poorly with lesions smaller than 3 mm.
CT also had moderate to strong correlations for mixed and non-calcified plaque, but poor correlations for calcified plaque.
“The most important findings from our study are that cardiac 64-MDCT shows a strong correlation with quantitative catheter angiography for grade of stenosis of non-culprit lesions in the proximal segment coronary arteries and for mixed or non-calcified plaques,” the authors wrote. “Large, purely calcified plaques and lesions in more distal coronary segments [≤ 3 mm diameter] remained difficult to accurately quantify.”
In an accompanying editorial, Andre J. Duerinckx, MD, said the study is very promising preliminary work in the field of coronary CTA, and it opens the door for future research.
“Until the advent of coronary CTA, we had no easy approach to coronary plaque detection or characterization, except for invasive coronary angiography in combination with intravascular ultrasound,” Duerinckx wrote.
But positive vessel remodeling, indicative of early disease, seen by CTA or intravascular ultrasound often appears normal on catheter angiography, he said. Therefore, CTA holds great promise to detect and help manage non-culprit plaques.
He added that the limitations shown by Dodd et al of CTA in this respect indicate more research and development need to be done.