Faculty Bibliography

Despite a better understanding of cardiovascular risk factors and attempts at optimal management, diabetes-related macrovascular events remain a significant cause of morbidity and mortality in the United States and worldwide. The trials to date have validated strict glycemic control as a method to achieve sustained reductions in the rate of nephropathy, neuropathy, and retinopathy due to diabetes. For these microvascular complications, the closer hemoglobin A1c is to normal levels, the better the outcome. Although reducing hemoglobin A1c levels to 7% has been shown to reduce macrovascular events, demonstrating an additional reduction in macrovascular events with tighter glycemic control has been more difficult to achieve. A careful review of recent trials, however, has demonstrated that treatment early in the disease course and the ability to safely maintain lower hemoglobin A1c levels might be critical factors in further reducing macrovascular events. In conclusion, with the introduction of novel antidiabetic agents, future trials using these drugs might be able to definitively establish the safety and efficacy of reducing cardiovascular events with stringent glycemic control; however, the current evidence is inconsistent.

AIM: To investigate diagnostic accuracy of high, low and mixed voltage dual energy computed tomography (DECT) for detection of prior myocardial infarction (MI). METHODS: Twenty-four consecutive patients (88% male, mean age 65 +/- 11 years old) with clinically documented prior MI (> 6 mo) were prospectively recruited to undergo late phase DECT for characterization of their MI. Computed tomography (CT) examinations were performed using a dual source CT system (64-slice Definition or 128-slice Definition FLASH, Siemens Healthcare) with initial first pass and 10 min late phase image acquisitions. Using the 17-segment model, regional systolic function was analyzed using first pass CT as normal or abnormal (hypokinetic, akinetic, dyskinetic). Regions with abnormal systolic function were identified as infarct segments. Late phase DE scans were reconstructed into: 140 kVp, 100 kVp, mixed (120 kVp) images and iodine-only datasets. Using the same 17-segment model, each dataset was evaluated for possible (grade 2) or definite (grade 3) late phase myocardial enhancement abnormalities. Logistic regression for correlated data was used to compare reconstructions in terms of the accuracy for detecting infarct segments using late myocardial hyperenhancement scores. RESULTS: All patients reported prior history of documented myocardial infarction, with most occurring more than 5 years prior (n = 18; 75% of cohort). Fifty-five of 408 (13%) segments demonstrated abnormal wall motion and were classified as infarct. The remaining 353 segments were classified as non-infarcted segments. A total of 1692 segments were analyzed for late phase enhancement abnormalities, with 91 (5.5%) segments not interpretable due to artifact. Combined grades 2 and 3 compared to grade 3 only enhancement abnormalities demonstrated significantly higher sensitivity and similar specificity for detection of infarct segments for all reconstructions evaluated. Evaluation of different voltage acquisitions demonstrated the highest diagnostic performance for the 100 kVp reconstruction which had higher diagnostic accuracy (87%; 95%CI: 80%-90%), sensitivity (86%-93%; 95%CI: 54%-78%) and specificity (90%; 95%CI: 86%-93%) compared to the other reconstructions. For sensitivity, there were significant differences noted between 100 kVp vs 140 kVp (P < 0.0005), 100 kVp vs mixed (P < 0.0001), and 100 kVp vs iodine only (P < 0.005) using combined grade 2 and grade 3 perfusion abnormalities. For specificity, there were significant differences noted between 100 kVp vs 140 kVp (P < 0.005), and 100 kVp vs mixed (P < 0.01) using combined grades 2 and 3 perfusion abnormalities. CONCLUSION: Low voltage acquisition CT, 100 kVp in this study, demonstrates superior diagnostic performance when compared to higher and mixed voltage acquisitions for detection of prior MI.

BACKGROUND: Obtaining diagnostic coronary CT angiography with low radiation exposure in patients with irregular heart rhythms such as atrial fibrillation (AF) remains challenging. OBJECTIVE: We evaluated image quality and inter-reader variability with the use of prospective electrocardiographic (ECG)-triggered sequential dual-source acquisition at end systole for coronary artery disease (CAD) evaluation in patients with AF. METHODS: Thirty consecutive patients with AF who underwent prospective ECG-triggered sequential dual-source acquisition were evaluated. Images were reconstructed every 50 milliseconds from 250 to 400 milliseconds after the R wave. Two independent, blinded readers evaluated the coronaries for image quality on a 5-point scale (worst to best) and stenosis on 5-point semiquantitative (none to severe) and binary scales (>50% or <50%). Diagnostic image quality was graded for each reconstruction. RESULTS: Eleven patients (37%) had significant (>/=50% stenosis) CAD. Average heart rate was 82 +/- 20 beats/min and variability range was 71 +/- 22 beats/min. Mean effective radiation dose was 6.5 +/- 2.4 mSv. Diagnostic image quality was noted in 97.9% of 304 coronary segments with median image quality of 3.0. The 300-millisecond reconstruction phase provided the highest image quality; 70% of patients showed diagnostic image quality. Combination of all phases (250-400 milliseconds) performed significantly better than single or other phase combinations (P < 0.0005 for all comparisons). Inter-reader variability for stenosis detection was excellent, with 98.4% concordance by using a binary scale (50% stenosis cutoff). CONCLUSIONS: Prospective ECG-triggered sequential dual-source CT acquisition with the use of end-systolic acquisition provides diagnostic image quality with potentially low radiation doses for evaluation of CAD in patients with AF. Use of multiple end-systolic phases over a 150-millisecond window improves diagnostic image quality.

OBJECTIVES: The purpose of this study was to assess the impact of rosiglitazone on survival in patients with diabetes mellitus (DM) and coronary artery disease (CAD). METHODS: We carried out a drug-exposure analysis in 801 patients with DM and CAD in a cardiac catheterization laboratory registry (490 patients treated with a percutaneous coronary intervention, 224 patients treated with coronary artery bypass grafting, and 87 patients treated with medication alone). RESULTS: A total of 193 patients (24.1%) were exposed to rosiglitazone. The median survival from the date of cardiac catheterization in the rosiglitazone group was 146.7 months versus 109.1 months in the unexposed group (P<0.001). At 5 years, the unadjusted survival was 82% in the rosiglitazone-exposed group versus 69% in the unexposed group (P<0.001). There was no difference in survival between rosiglitazone-exposed and rosiglitazone-unexposed patients in the groups treated with coronary artery bypass grafting or medical therapy (P=0.37 and 0.11, respectively). In a multivariable model, rosiglitazone exposure had no effect on mortality (hazard ratio=0.737; 95% confidence interval: 0.521-1.044, P=0.86). CONCLUSION: We conclude that exposure to rosiglitazone is not associated with increased mortality in diabetics who are treated for CAD. These findings support the notion that insulin sensitization with a thiazolidinedione is safe in carefully selected and treated patients with DM and CAD.

Introduction: Coronary CTA is a reliable non-invasive imaging test for evaluating coronary artery disease (CAD). Obtaining diagnostic image quality (IQ) with low radiation exposure in patients with irregular heart rates such as atrial fibrillation, however, remains challenging. This study evaluated IQ and inter-reader variability using prospective ECGtriggered sequential dual-source data acquisition at end-systole (PROS-CT) for evaluation of CAD in patients with atrial fibrillation. Methods: 30 consecutive patients (22 men, mean age 64 years, mean BMI 29) with persistent atrial fibrillation who underwent PROS-CT for evaluation of pulmonary vein anatomy were identified from our cardiac CT patient registry. Images were reconstructed using standard protocols every 50ms from 250 to 400ms after the R wave trigger. Two independent, blinded readers evaluated the coronary arteries using an 18-segment model and all available reconstructions. IQ was graded on a 5-point Likert scale (worst to best) and coronary stenosis on 5-point semi-quantitative (none to severe) and binary scales (greater or less than 50%). In addition, overall IQ and diagnostic quality were graded for each reconstruction. Results: 11 patients (37%) had significant (>50% stenosis) CAD. The average heart rate was 82+/-20 bpm and the average heart rate variability (range) was 71+/-22 bpm. The mean effective radiation dose was 6.5+/-2.4 mSv and mean CTDIvol was 23.8+/-8.0 mGy. There were 304 coronary segments available for interpretation with mean IQ of 2.9+/-0.8. Diagnostic IQ was noted in 97.9% of segments, with 2 patients demonstrating at least 1 segment with non-diagnostic IQ. The 300ms reconstruction phase provided the highest IQ with 70% of patients demonstrating diagnostic IQ, but the combination of all phases (250-400ms) performed significantly better than single or other double and triple reconstruction phase combinations (p<0.0005 for all comparisons). Inter-reader variability for detection of stenosis was excellent, with 98.4% concordance using the binary scale with 50% stenosis cutoff, and 75.3% concordance using 5-point stenosis grade. Conclusions: PROS-CT with acquisition at end-systole provides good diagnostic image quality for evaluation of significant CAD in patients with atrial fibrillation. The use of multiple end-systolic phases over a 150ms window for each study improves diagnostic IQ

Cardiac computed tomographic angiography algorithms emphasize radiation reduction while maintaining diagnostic image quality (IQ). The aim of this study was to evaluate IQ and interreader variability using prospective electrocardiographically triggered high-pitch spiral cardiac computed tomographic angiography (FLASH-CT) compared to retrospective electrocardiographic gating (RETRO-CT) for coronary artery disease evaluation in a patient population including overweight and obese individuals. MATERIALS AND METHODS: Seventy patients (24 women; mean age, 60 years) matched for gender, age, body mass index (27.4 ± 5.5 kg/m(2)), and calcium score (184 ± 328) underwent cardiac computed tomographic angiography, 35 with FLASH-CT (Definition Flash) and 35 with RETRO-CT (Somatom Definition). Images were reconstructed using standard protocols and least motion phase for RETRO-CT acquisitions. Two independent, blinded readers evaluated the coronary arteries using an 18-segment model, grading IQ on a 5-point, Likert-type scale and coronary stenosis on a 5-point semiquantitative and binary scale. RESULTS: Effective radiation dose (1.50 vs 17.3 mSv, P < .0001) and mean heart rate (58 vs 62 beats/min, P < .05) were significantly lower for FLASH-CT compared to RETRO-CT. Seven hundred forty segments (> 1.5 mm) were evaluated. There was no significant difference between FLASH-CT and RETRO-CT scans in overall per-segment IQ (3.11 ± 0.75 vs 3.10 ± 0.82, P = .94). FLASH-CT had noninferior IQ relative to RETRO-CT (95% confidence interval, -0.25 to 0.26). There was no significant difference in interreader variability in diagnosis between FLASH-CT and RETRO-CT for all coronary segments (77.5% vs 78.2%, P = .83). CONCLUSIONS: FLASH-CT is an acceptable coronary computed tomographic angiographic method for reducing radiation dose without compromising IQ for a patient population including overweight and obese individuals.

Introduction: Patients receiving radiofrequency ablation (RFA) for treatment of atrial fibrillation typically undergo pre-procedural cardiac computed tomography angiography (CCTA) to delineate pulmonary venous anatomy and transesophageal echocardiogram (TEE) to exclude left atrial and/or left atrial appendage thrombus (LAT). The addition of a late phase acquisition is theorized to aid CCTA identification and discrimination of LAT from slow left atrial appendage filling. The purpose of this study is to evaluate the diagnostic accuracy of dual-phase, ECG-gated dual-source CCTA (64-slice Definition, Siemens) compared to TEE for identification of thrombus and to assess the added value of a late phase CCTA acquisition. Methods: Fifty-three consecutive patients (37 men; mean age 63) had both dual-phase CCTA and TEE prior to RFA. Mean time between CCTA and TEE was 9 days (range 1-22). Mean early phase and late phase scan acquisition delay times were 29 sec and 30 sec, respectively. Presence of LAT was independently graded on both early phase and combined early and late phase (CP) CCTA acquisitions using a 5-point Likert scale by 2 readers blinded to the TEE results. Diagnostic accuracy for LAT was assessed for early phase and CP CCTA acquisitions using TEE results as truth. Results: CCTA identified LAT in 2 out of 3 patients with thrombi on TEE (67%). Relative to TEE, early phase and CP CCTA acquisitions demonstrated: 47% and 67% sensitivity, 84% and 100% specificity, 54% and 100% PPV, 80% and 98% NPV, respectively. Overall diagnostic accuracy was significantly improved for CP compared to early phase acquisition (98% and 77%, respectively, p<0.001). Conclusions: CCTA has excellent specificity (100%) but only modest sensitivity (66.7%) for identification of LAT in patients undergoing RFA. Addition of a late phase CCTA acquisition significantly improves overall diagnostic accuracy

Introduction: Patients receiving radiofrequency ablation (RFA) for the treatment of atrial fibrillation frequently undergo pre-procedural cardiac CT for evaluation of the left atrium and pulmonary veins. Left atrial outpouchings (LAO), including diverticula and accessory appendages, can be mistaken for an ostium of a pulmonary vein, which are important to identify as there is a potential risk of complications during RFA. The prevalence of these outpouchings has been described to be as high as 27 percent in the population of patients undergoing routine cardiac CT.1 The purpose of this study is to describe the prevalence, morphology, and size of LAO in patients undergoing RFA for treatment of atrial fibrillation. Methods: Fifty consecutive patients referred for RFA were identified from our registry of patients undergoing gated cardiac CT. Data was independently analyzed by two blinded readers for LAO. Images were evaluated using multiplanar reformatted and 3D reconstruction. The presence of LAO was defined as any abnormality that had a discernable ostium stemming from the left atrial wall. The number and size of LAO were recorded. Comparison of prevalence was evaluated using the Fisher's exact test. Results: There were a total of 29 LAO found in 24 of the 50 patients for a calculated prevalence of 48% (95% CI: 33.6 to 62.6). The prevalence in our population was significantly higher than reported in the general cohort of patients undergoing routine cardiac CT (p=0.003). The average size (length, width, and depth) of the LAO were 0.54 +/- 0.28 by 0.39 +/- 0.20 by 0.56 +/- 0.26 cm. Conclusions: Patients undergoing RFA for atrial fibrillation have a high prevalence of L

Diabetic patients with coronary artery disease (CAD) are more likely to develop heart failure (HF) than nondiabetic patients, but the mechanism responsible is unclear. Evidence suggests that infarct size and accompanying remodeling may not explain this difference. We used cardiac magnetic resonance (CMR) imaging to compare degree of left ventricular (LV) myocardial scar and remodeling in diabetic and nondiabetic patients with CAD. We evaluated 85 patients (39 diabetic, 46 nondiabetic) who underwent coronary angiography showing obstructive CAD and CMR imaging within 6 months of each other. Myocardial scar was measured by late gadolinium enhancement on CMR imaging and was graded according to spatial and transmural extents on a semiquantitative scale. More diabetic than nondiabetic patients had HF (69% vs 43%, p <0.03); however, groups did not differ in total scar burden (0.94 +/- 0.60 vs 1.17 +/- 0.74, p = NS), spatial extent of scar, or extent of transmural scar. Diabetes remained an independent predictor of HF after adjustment for CAD and other variables. LV ejection fraction (36 +/- 12% vs 37 +/- 14%, p = NS) and end-diastolic volume (215 +/- 56 vs 217 +/- 76 ml, p = NS) were similar for diabetic and nondiabetic patients, respectively. In conclusion, although diabetic patients with CAD had a higher prevalence of HF than nondiabetic patients, there was no difference in myocardial scar, LV volume, or LV ejection fraction. These findings support the theory that mechanisms other than extent of myocardial injury and negative remodeling play a significant role in the development of HF in diabetic patients with CAD