- Email: [email protected]
115: Accuracy in Quantification of the Calcifying Burden with CT: A Cork-Dog Heart Phantom Study
3rd Annual Scientific Meeting
S41
emerging as an important tool in the investigation of patients with suspected CAD. However, there are several factors that could adversely affect the adoption of CCTA in clinical p ractice, including complexity and time-consuming nature of this technique and at times, unavailability of experienced reader. We describe a new FDA approved technology, COR Analyzer II which addresse s this limitation through computer-assisted automation. Methods: In order to assess the diagnostic performance of COR Analyzer II we conducted a study comparing the results produced by the system to findings reported by two experienced level III CCTA readers. The experiment was performed on 154 CCTA s tudies, (151 studies from 64 slice MDCT and 3 studies from 16 slice MDCT). The readers independently reviewed the results and agreed on 147 out of 154 cases (95%). The readers reached a unified agreement on 7 cases (5%) after reviewing these cases jointly. Results: COR Analyzer II defines study as positive when it finds at least one significant lesion (more than 50% stenosis). Four different results are possible, including: Negative - no significant pathologies are detected; Positive – severe pathologies are detected; Warning - no severe pathologies found, but there are potential problems in automatic analysis; Failed – automatic analysis is failed. In our experiment, the system failed to process automatically 16 studies (10%). There were 63 positive (P) and 75 negative (N) cases. The measured system performance was: Hits (TP) 58, Misses (FN) 3, True negative (TN) 51, (FP) 20, Warnings on Positive (WP) 2, Warnings on Negative (WN) 4. If warnings are considered positive the derived statistics are: Sensitivity 95.2%; Specificity 68%; NPV 94.4%; PPV 71.4%. If warnings are not considered positive, Specificity is 73.3%; PPV is 74.4%. Conclusion: Based on our initial experience and the results of the conducted experiments, we believe that software-based fully automatic analysis of CCTA studies is feasible. There is pote ntial for improved triage in the ER setting with reduction of time to treatment, improved mortality and morbidity, reduction of unnecessary hospital admissions, reduction of unnecessary emergency calls to expert readers and overall significant healthcare cost savings.
Sensitivity (TP+WP)/P
Specificity TN/N
NPV TN/(TN+FN)
PPV (TP+WP)/(TP+WP +FP+WN)
95.2%
68.0%
94.4%
71.4%
115 Budoff MJ, Chow D, Gao YL, Chang WW, Schiff M, Kwan J, Kim K, Mao SS. Accuracy in Quantification of the Calcifying Burden with CT: A Cork-Dog Heart Phantom Study Division of Cardiology, Los Angeles Biomedical Rese arch Institute at Harbor-UCLA Medical Center, Torrance, CA Introduction: The quantification of calcium foci was influenced by the volume affect with CT image. The aim of this study was to test the accuracy in quantification of calcifying foci with electron beam CT (EBT, C300, GE Imatron) and multi-detector row CT ( MDCT, GE, Milwaukee, WI). Methods: This study phantom consists of a cork chest and five dog hearts. The cork chest was constructed for use as a human chest equivalent. The dog heart contained 135 calcium hydroxyapatite (CaHA) foci in coronary artery or myocardium known the volume, mass and concentration. In group 1 was consisted of 78 large (7.8-187.2 mm³) with a high density (375mg/cc or 200 mg/cc) CaHA foci in three dog hearts and group 2 was 57small (<6.6-18.9 mm³ with 200 mg/cc) or low density foci (14.8-157.8 mm³, 150mg/cc) in 2 dog hearts. Five hearts were sealed by wax and inserted into a cork chest phantom. Standard coronary calcium protocols were completed with EBT and MDCT. The volume of 135 CaHA foci was measured with Aquarius workstation (TeraRecon, San Mateo, CA). Results: In group 1, the total volume of 87 CaHA foci measured was 4165.4 and 3779.1 mm³ with EBT and MDCT respectively (P<0.001), both significantly larger than the true value (2713.9 mm³, P<0.001). In
group2, the total measured value of 57 foci were 592.6 with EBT and 702.9 mm³ with MDCT (P<0.001) and smaller than the true volume (1733.2 mm³, P<0.001). Conclusions: CT images magnify CaHA foci with a larger size with more dense, and underestimate the foci volume with a smaller size (<6.6 mm³) or low density. Poster Abstracts VII: CT for Atherosclerotic Plaque Imaging I Abstracts 116-124 116 Ibebuogu UN 1, Nasir K2, Ahmadi N 3, Gopal A3, Goodwin D3, Budoff MJ3. Atherosclerotic Plaque Morphology Assessment According To Burden of Coronary Artery Disease in Patients with Type 2 Diabetes Mellitus by Non Invasive CT Angiography 1 Medical college of Georgia, Augusta, GA 2 Massachusetts General Hospital, Boston, MA 3 Los Angeles Biomedical Research Institute at Harbor-UCLA, Torrance, CA Introduction: Patients with diabetes mellitus (DM) have a higher risk of coronary artery disease (CAD) and are likely to have a higher underlying atherosclerotic burden. However the atherosclerotic plaque composition in these patients is not studied. In this study we evaluated the plaque burden, morphology and distribution in type 2 DM patients using multi-detector computed tomography angiography (MDCTA). Methods: The study population consisted of 40 symptomatic diabetic subjects (63±10 years, 55% men), who underwent contrast-enhanced MDCTA. Enrolled patients had an intermediate pre-test probability of obstructive coronary artery disease. Results: Majority of individuals (n=33, 83%) had at least one segment with any plaque; 69% of diabetic had detectable coronary artery calcification (CAC) and 36% had CAC 400. Among individuals with any plaque the mean number of segments involved were 5.7±3.0 segments; the respective mean number of segments with exclusively non-calcified, calcified and mixed plaques were 1±1, 2.7±2.4, and 2±2 segments respectively. Among those with any plaque, the overall proportion of segments that had noncalcified, calcified and mixed atherosclerotic plaques were 25%, 44%, and 31% respectively (see figure). In our study, 12 patients (30%) had at least one coronary segment with significant stenosis (luminal narrowing 50%). Type 2 DM patients in our study with significant stenosis were more likely to have plaque composition, that was mixed in nature (39% vs. 28%) and less likely to be exclusively noncalcified plaque alone (17% vs. 26%) when compared to those without significant stenosis. On the other hand no difference was observed in the respective proportion of exclusively calcified plaque (44% vs. 46%). Conclusions: Our study demonstrates a high burden of CAD in patients with diabetes. Majority of atherosclerotic plaque in these patients had calcification, however among patients with significant CAD, mixed plaque composition was more commonly observed. The prognostic value of these different atherosclerotic plaque morphologies on MDCT in high risk DM patients need to be assessed in larger prospective studies.
117 Achenbach S 1, Feuchtner G2, Malhotra V3, Lesser J4, Burgstahler C5, Schroeder S5.
S41
emerging as an important tool in the investigation of patients with suspected CAD. However, there are several factors that could adversely affect the adoption of CCTA in clinical p ractice, including complexity and time-consuming nature of this technique and at times, unavailability of experienced reader. We describe a new FDA approved technology, COR Analyzer II which addresse s this limitation through computer-assisted automation. Methods: In order to assess the diagnostic performance of COR Analyzer II we conducted a study comparing the results produced by the system to findings reported by two experienced level III CCTA readers. The experiment was performed on 154 CCTA s tudies, (151 studies from 64 slice MDCT and 3 studies from 16 slice MDCT). The readers independently reviewed the results and agreed on 147 out of 154 cases (95%). The readers reached a unified agreement on 7 cases (5%) after reviewing these cases jointly. Results: COR Analyzer II defines study as positive when it finds at least one significant lesion (more than 50% stenosis). Four different results are possible, including: Negative - no significant pathologies are detected; Positive – severe pathologies are detected; Warning - no severe pathologies found, but there are potential problems in automatic analysis; Failed – automatic analysis is failed. In our experiment, the system failed to process automatically 16 studies (10%). There were 63 positive (P) and 75 negative (N) cases. The measured system performance was: Hits (TP) 58, Misses (FN) 3, True negative (TN) 51, (FP) 20, Warnings on Positive (WP) 2, Warnings on Negative (WN) 4. If warnings are considered positive the derived statistics are: Sensitivity 95.2%; Specificity 68%; NPV 94.4%; PPV 71.4%. If warnings are not considered positive, Specificity is 73.3%; PPV is 74.4%. Conclusion: Based on our initial experience and the results of the conducted experiments, we believe that software-based fully automatic analysis of CCTA studies is feasible. There is pote ntial for improved triage in the ER setting with reduction of time to treatment, improved mortality and morbidity, reduction of unnecessary hospital admissions, reduction of unnecessary emergency calls to expert readers and overall significant healthcare cost savings.
Sensitivity (TP+WP)/P
Specificity TN/N
NPV TN/(TN+FN)
PPV (TP+WP)/(TP+WP +FP+WN)
95.2%
68.0%
94.4%
71.4%
115 Budoff MJ, Chow D, Gao YL, Chang WW, Schiff M, Kwan J, Kim K, Mao SS. Accuracy in Quantification of the Calcifying Burden with CT: A Cork-Dog Heart Phantom Study Division of Cardiology, Los Angeles Biomedical Rese arch Institute at Harbor-UCLA Medical Center, Torrance, CA Introduction: The quantification of calcium foci was influenced by the volume affect with CT image. The aim of this study was to test the accuracy in quantification of calcifying foci with electron beam CT (EBT, C300, GE Imatron) and multi-detector row CT ( MDCT, GE, Milwaukee, WI). Methods: This study phantom consists of a cork chest and five dog hearts. The cork chest was constructed for use as a human chest equivalent. The dog heart contained 135 calcium hydroxyapatite (CaHA) foci in coronary artery or myocardium known the volume, mass and concentration. In group 1 was consisted of 78 large (7.8-187.2 mm³) with a high density (375mg/cc or 200 mg/cc) CaHA foci in three dog hearts and group 2 was 57small (<6.6-18.9 mm³ with 200 mg/cc) or low density foci (14.8-157.8 mm³, 150mg/cc) in 2 dog hearts. Five hearts were sealed by wax and inserted into a cork chest phantom. Standard coronary calcium protocols were completed with EBT and MDCT. The volume of 135 CaHA foci was measured with Aquarius workstation (TeraRecon, San Mateo, CA). Results: In group 1, the total volume of 87 CaHA foci measured was 4165.4 and 3779.1 mm³ with EBT and MDCT respectively (P<0.001), both significantly larger than the true value (2713.9 mm³, P<0.001). In
group2, the total measured value of 57 foci were 592.6 with EBT and 702.9 mm³ with MDCT (P<0.001) and smaller than the true volume (1733.2 mm³, P<0.001). Conclusions: CT images magnify CaHA foci with a larger size with more dense, and underestimate the foci volume with a smaller size (<6.6 mm³) or low density. Poster Abstracts VII: CT for Atherosclerotic Plaque Imaging I Abstracts 116-124 116 Ibebuogu UN 1, Nasir K2, Ahmadi N 3, Gopal A3, Goodwin D3, Budoff MJ3. Atherosclerotic Plaque Morphology Assessment According To Burden of Coronary Artery Disease in Patients with Type 2 Diabetes Mellitus by Non Invasive CT Angiography 1 Medical college of Georgia, Augusta, GA 2 Massachusetts General Hospital, Boston, MA 3 Los Angeles Biomedical Research Institute at Harbor-UCLA, Torrance, CA Introduction: Patients with diabetes mellitus (DM) have a higher risk of coronary artery disease (CAD) and are likely to have a higher underlying atherosclerotic burden. However the atherosclerotic plaque composition in these patients is not studied. In this study we evaluated the plaque burden, morphology and distribution in type 2 DM patients using multi-detector computed tomography angiography (MDCTA). Methods: The study population consisted of 40 symptomatic diabetic subjects (63±10 years, 55% men), who underwent contrast-enhanced MDCTA. Enrolled patients had an intermediate pre-test probability of obstructive coronary artery disease. Results: Majority of individuals (n=33, 83%) had at least one segment with any plaque; 69% of diabetic had detectable coronary artery calcification (CAC) and 36% had CAC 400. Among individuals with any plaque the mean number of segments involved were 5.7±3.0 segments; the respective mean number of segments with exclusively non-calcified, calcified and mixed plaques were 1±1, 2.7±2.4, and 2±2 segments respectively. Among those with any plaque, the overall proportion of segments that had noncalcified, calcified and mixed atherosclerotic plaques were 25%, 44%, and 31% respectively (see figure). In our study, 12 patients (30%) had at least one coronary segment with significant stenosis (luminal narrowing 50%). Type 2 DM patients in our study with significant stenosis were more likely to have plaque composition, that was mixed in nature (39% vs. 28%) and less likely to be exclusively noncalcified plaque alone (17% vs. 26%) when compared to those without significant stenosis. On the other hand no difference was observed in the respective proportion of exclusively calcified plaque (44% vs. 46%). Conclusions: Our study demonstrates a high burden of CAD in patients with diabetes. Majority of atherosclerotic plaque in these patients had calcification, however among patients with significant CAD, mixed plaque composition was more commonly observed. The prognostic value of these different atherosclerotic plaque morphologies on MDCT in high risk DM patients need to be assessed in larger prospective studies.
117 Achenbach S 1, Feuchtner G2, Malhotra V3, Lesser J4, Burgstahler C5, Schroeder S5.