Abstract

Background: Studies have demonstrated considerable accuracy of multi-slice CT coronary angiography (MSCT-CA) in comparison to invasive coronary angiography (I-CA) for evaluating coronary artery disease (CAD). The extent to which published MSCT-CA accuracy parameters are transferable to routine practice beyond high-volume tertiary centres is unknown.

Aim: To determine the accuracy of MSCT-CA for the detection of CAD in a Scottish district general hospital.

Design: Prospective study of diagnostic accuracy.

Method: One hundred patients with suspected CAD recruited from two Glasgow hospitals underwent both MSCT-CA (Philips Brilliance 40 × 0.625 collimation, 50–200 ms temporal resolution) and I-CA. Studies were reported by independent, blinded radiologists and cardiologists and compared using the AHA 15-segment model.

Results: Of 100 patients [55 male, 45 female, mean (SD) age 58.0 (10.7) years], 59 and 41% had low-intermediate and high pre-test probabilities of significant CAD, respectively. Mean (SD) heart rate during MSCT-CA was 68.8 (9.0) bpm. Fifty-seven per cent of patients had coronary artery calcification and 35% were obese. Patient prevalence of CAD was 38%. Per-patient sensitivity, specificity, positive and negative (NPV) predictive values for MSCT-CA were 92.1, 47.5, 52.2 and 90.6%, respectively. NPV was reduced to 75.0% in the high pre-test probability group. Specificity was compromised in patients with sub-optimally controlled heart rates, calcified arteries and elevated BMI.

Conclusion: Forty-Slice MSCT-CA has a high NPV for ruling out significant CAD when performed in a district hospital setting in patients with low-intermediate pre-test probability and minimal arterial calcification. Specificity is compromised by clinically appropriate strategies for dealing with unevaluable studies. Effective heart rate control during MSCT-CA is imperative. National guidelines should be utilized to govern patient selection and direct MSCT-CA reporter training to ensure quality control.

Introduction

Coronary artery disease (CAD) is the leading cause of mortality in Scotland (population 5.2 million), accounting for around 9000 deaths each year.1 Accurate diagnosis of the presence and extent of CAD is essential to guide management. Invasive coronary angiography (I-CA) is the gold standard diagnostic investigation but is associated with a small risk of significant vascular complications.2 Over the last decade, multi-slice computed tomography coronary angiography (MSCT-CA) has emerged as a non-invasive imaging modality capable of visualizing the coronary arteries. Progressive advancements in scanner technology have greatly improved the accuracy of MSCT-CA in comparison to I-CA.3,4

The primary shortcoming of existing evidence is that it derives from specialist academic centres with substantial experience in MSCT-CA. The accuracy of MSCT-CA in smaller centres with variable expertise and a more heterogeneous population is unknown. Accordingly, the European Society of Cardiology advocate MSCT-CA for assessment of symptomatic patients with an intermediate pre-test probability of CAD only if diagnostic image quality can be expected and the investigation can be expertly performed and reported.5

Recently, the National Institute for Health and Clinical Excellence (NICE) published UK guidelines for the investigation of patients with recent onset chest pain.6 MSCT-CA was considered appropriate in patients with a low estimated likelihood of underlying CAD and some evidence of coronary artery calcification. In light of these recent national recommendations, an objective assessment of MSCT-CA accuracy in routine clinical practice is vital.

We present a prospective, comparative study in a district general hospital in Scotland determining the accuracy of MSCT-CA in comparison to I-CA for the detection of significant CAD.

Method

Study population

Between January 2007 and May 2008, 100 patients from two Glasgow hospitals were recruited. Inclusion criteria were suspected cardiac ischaemia on the basis of symptoms and non-invasive stress testing with subsequent referral for elective I-CA to determine the presence or absence of CAD. Patients with previous myocardial infarction or previous CAD on I-CA were excluded. Patients with unstable symptoms, documented iodine contrast allergy, hyperthyroidism, significant renal dysfunction (defined as serum creatinine >150 µmol/l or >120 µmol/l in a diabetic patient) and possible pregnancy were excluded to minimize patient risk. Exclusion criteria based on anticipated technical difficulties with the MSCT-CA protocol were atrial fibrillation or frequent ectopic activity and inability to carry out a 12 s breath hold. The study protocol was approved by the North Glasgow Research Ethics Committee and all patients gave written informed consent.

Patient preparation

Symptoms, risk factors for CAD, medications, routine investigations and non-invasive stress test results were recorded. Treatment with oral β-blockers or rate-limiting calcium channel blockers was commenced and titrated aiming for a resting heart rate <65 bpm.7 Pre-test probability was assessed using the Duke Clinical Score.8 Patients were categorized in two groups: low-intermediate pre-test probability (Score 0–74%) and high pre-test probability (Score ≥75%).

MSCT-CA protocol

MSCT-CA was performed on the Philips Brilliance MSCT scanner with 40 simultaneous detector rows and 0.625 mm collimation. Gantry rotation time was 400 ms with a half-sector acquisition protocol and multi-sector reconstruction permitting an effective temporal resolution of between 50 and 200 ms depending on patient heart rate. Tube voltage was 120 kV or 140 kV depending on patient weight and maximum effective tube current was 600 mA/slice. Prospective ECG-dependent tube current modulation was employed to minimize radiation exposure.

A single axial CT image defined the area to be scanned, from the bifurcation of the trachea to the diaphragm. A region of interest at the origin of the descending aorta was marked to permit subsequent use of automated contrast bolus tracking. Iodinated contrast media [Omnipaque 350 (Schering AG, Berlin, Germany) in Patients 1–17 and Iomeron 400 (Bracco, Italy) thereafter] was injected via a wide bore cannula in a proximal peripheral vein. Contrast volume and rate of injection varied with patient weight from 90 to 120 ml and 5.3 to 6.9 ml/s, respectively. The contrast injection was immediately followed by a 50 ml saline ‘chaser bolus’ at a rate of 5 ml/s. Scanning was automatically triggered when contrast media in the pre-defined area of the descending aorta reached a density of 160 Hounsfield units. A single automated breath-hold command was given and helical scan acquisition commenced 3 s thereafter to minimize respiratory-related fluctuation in heart rate. Overall scan time was between 10 and 15 s depending on cardiac size.

Image reconstruction

Data were reconstructed using either a mono- or multi-segmental algorithm depending on patient heart rate. A volume acquisition approach with a pitch of 0.2 was employed, reconstructing axial images with slice thickness of 0.9 mm using a medium soft tissue reconstruction kernel. Retrospective ECG gating permitted reconstruction of images at 45, 50, 60, 75 and 80% of the RR interval to allow evaluation of the coronary arteries at the cardiac phase with least vessel motion. Reconstructed data were transferred to a dedicated offline image analysis workstation (Philips Extended Brilliance Workspace).

MSCT-CA analysis

MSCT-CAs were reported according to the 15-segment model of the American Heart Association (AHA).9 A segment with a luminal diameter reduction ≥50% was classified as a significant stenosis. MSCT-CAs were reported by two independent, experienced consultant radiologists and discrepancies involving stenoses considered by one radiologist, but not the other, to be ≥50% were resolved by an independent consultant cardiologist experienced in MSCT-CA. All reporters were blinded to each other’s MSCT-CA reports and to the I-CA reports.

For each vessel the optimal RR interval percentage reconstruction was identified. Stenoses identified in at least two independent orthogonal planes had their percentage of luminal reduction assessed on a semi-quantitative basis (from 0 to 100% in 10% increments or ‘unevaluable’). The percentage of stenosis was ascertained by use of planimetry on an axis perpendicular to the course of the segment. Visualization techniques varied between segments and included straight and curved multi-planar reformations, maximum intensity projections and volume rendering. The degree of calcification of each segment’s vessel wall was assessed subjectively as heavy (indicating high density lesions extending longitudinally along the vessel wall, resulting in beam hardening and partial volume artefact), moderate (indicating small, isolated eccentric high density lesions in the vessel wall), or none.10

I-CA protocol

I-CAs were carried out a minimum of 6 days after MSCT-CA to limit the risk of a further contrast load and a maximum of 4 weeks following MSCT-CA. I-CA was performed applying the Judkins approach via the trans-radial or trans-femoral route and acquiring standard projections.

I-CA analysis

I-CAs were reported according to the 15-segment model of the AHA9 with luminal diameter reductions ≥50% considered significant. Each ICA was reported by two of four independent consultant cardiologists experienced in performing and reporting I-CA. All reporters were blinded to each other’s I-CA reports and to the MSCT-CA reports. The degree of stenosis in each diseased coronary artery segment was assessed semi-quantitatively in two orthogonal planes. Discrepancies concerning stenoses considered by one consultant, but not the other, to be ≥50% were resolved by consensus. This method was considered to best represent current clinical practice.

Statistical analysis

Statistical analysis was performed using R version 2.9.1. Standard descriptive statistics were used. Sensitivity, specificity, positive (PPV) and negative (NPV) predictive values were calculated for MSCT-CA in comparison to I-CA for the detection of stenosis ≥50% on a per-patient, per-artery and per-segment basis with 95% confidence intervals (CI) calculated from binomial expression. Categorical variables were compared between groups using Fisher’s exact tests and continuous variables were compared between groups using t-tests or Wilcoxon tests as appropriate. Inter-observer agreement data were expressed as Cohen’s kappa statistics (κ) with bootstrap confidence intervals (1000 replicates).

Results

Patient characteristics

Over 17 months, a total of 100 patients [55 male, 45 female, mean (SD) age 58.0 (10.7) years] underwent both MSCT-CA and I-CA without complication and all were included in analysis. Patient characteristics are summarized in Table 1.

Table 1

Baseline patient characteristics (n = 100)

 % or Mean (SD) 

 
Patient characteristics  
 Male 55 
 Mean age (years) 58.0 (10.7) 
 Hypertensiona 54 
 Hypercholesterolaemiaa 87 
 Smoker (current or previous) 57 
 Family history CADa 40 
 Diabetes mellitus 11 
 Mean BMI kg/m2 28.6 (5.2) 
 BMI ≥30 kg/m2 35 
 BMI <30 kg/m2 65 
 β-blocker 66 
 Rate-limiting calcium channel  blocker 24 
Presentation and investigations  
 Chest pain 94 
 Typical angina 56 
 Median symptom duration  (months) (IQR) 5.0 (2.0–13.0) 
 Low pre-test probabilityb 19 
 Intermediate pre-test probabilityb 40 
 High pre-test probabilityb 41 
 Previous exercise tolerance test 92 
 Previous myocardial perfusion  imaging 29 
MSCT-CA characteristics  
 Mean heart rate during  MSCT-CA (bpm) 68.8 (9.0) 
 Mean radiation dose (mSv) 15.6 (3.0) 
 Moderate coronary calcification on  MSCT-CA 22 
 Heavy coronary calcification on  MSCT-CA 35 
I-CA characteristics  
 No significant CAD  (No stenosis ≥50%)c 62.9 
 Single vessel disease (one vessel with  stenosis ≥50%)c 11.3 
 Two vessel disease (two vessels with  stenoses ≥50%)c 11.3 
 Triple vessel disease (three vessels  with stenoses ≥50%)c 14.4 
 % or Mean (SD) 

 
Patient characteristics  
 Male 55 
 Mean age (years) 58.0 (10.7) 
 Hypertensiona 54 
 Hypercholesterolaemiaa 87 
 Smoker (current or previous) 57 
 Family history CADa 40 
 Diabetes mellitus 11 
 Mean BMI kg/m2 28.6 (5.2) 
 BMI ≥30 kg/m2 35 
 BMI <30 kg/m2 65 
 β-blocker 66 
 Rate-limiting calcium channel  blocker 24 
Presentation and investigations  
 Chest pain 94 
 Typical angina 56 
 Median symptom duration  (months) (IQR) 5.0 (2.0–13.0) 
 Low pre-test probabilityb 19 
 Intermediate pre-test probabilityb 40 
 High pre-test probabilityb 41 
 Previous exercise tolerance test 92 
 Previous myocardial perfusion  imaging 29 
MSCT-CA characteristics  
 Mean heart rate during  MSCT-CA (bpm) 68.8 (9.0) 
 Mean radiation dose (mSv) 15.6 (3.0) 
 Moderate coronary calcification on  MSCT-CA 22 
 Heavy coronary calcification on  MSCT-CA 35 
I-CA characteristics  
 No significant CAD  (No stenosis ≥50%)c 62.9 
 Single vessel disease (one vessel with  stenosis ≥50%)c 11.3 
 Two vessel disease (two vessels with  stenoses ≥50%)c 11.3 
 Triple vessel disease (three vessels  with stenoses ≥50%)c 14.4 

aHypertension, hypercholesterolaemia and family history of CAD were defined, respectively, by the Joint British Societies’ Guidelines 200511 as BP ≥140/90 (or current antihypertensive therapy), cholesterol >5 mmol/l (or current statin therapy) and angina or myocardial infarction in a male relative age <55 years or a female relative age <65 years.

bPre-test probability determined by the Duke clinical score8.

cBased on I-CA of 97 patients as in each of three patients one vessel was considered unevaluable on I-CA.

Accuracy of MSCT-CA

Of 100 patients, 99 were included in a per-patient analysis. One patient was excluded as their I-CA was not fully evaluable. The MSCT-CAs of 32 patients were considered not fully evaluable due to there being at least one unevaluable segment of a major artery. To permit a clinically relevant per-patient analysis these patients were all considered to have significant CAD, subsequently generating 8 true positives and 24 false positives. Of 38 patients identified on I-CA as having at least one stenosis ≥50%, MSCT-CA correctly identified 35 patients (92%). In the three patients where MSCT-CA failed to detect CAD diagnosed on I-CA, a 70% distal circumflex lesion, an 80% left ventricular branch lesion, a 70% obtuse marginal vessel lesion and a 60% proximal right coronary artery lesion in the presence of moderate arterial calcification were missed. Of 61 patients with no significant CAD on I-CA, 29 were correctly identified by MSCT-CA, 8 were incorrectly considered to have a stenosis ≥50% and in 24 patients the MSCT-CA was not fully evaluable. Table 2 demonstrates the accuracy parameters and 95% CIs for MSCT-CA for the detection of significant CAD in comparison to I-CA in per-patient, per-artery and per-segment analyses.

Table 2

Accuracy of MSCT-CA compared to I-CA for the detection of CAD ≥50%

Analysis Number (excluded) ICA stenosis ≥50% MSCT stenosis ≥50%
 
Sensitivity (95% CI) Specificity (95% CI) NPV (95% CI) PPV (95% CI) 
   No Yes Unevaluable 

 
Per-patienta 99 (1) No 29 24 92.1 (79.2–97.3) 47.5 (35.5–59.8) 90.6 (75.8–96.8) 52.2 (40.5–63.7) 
Yes 27 
Unevaluable 
Per-arterya 397 (3) No 239 17 56 71.8 (61.4–80.2) 76.6 (71.6–81.0) 90.9 (86.8–93.8) 45.5 (37.3–54.0) 
Yes 24 36 25 
Unevaluable 
Per-segmentb 953 (490) No 833 29 440 36.3 (27.1–46.5) 96.6 (95.2–97.6) 93.5 (91.7–94.9) 53.2 (41.0–65.1) 
Yes 58 33 50 
Unevaluable 47 
Analysis Number (excluded) ICA stenosis ≥50% MSCT stenosis ≥50%
 
Sensitivity (95% CI) Specificity (95% CI) NPV (95% CI) PPV (95% CI) 
   No Yes Unevaluable 

 
Per-patienta 99 (1) No 29 24 92.1 (79.2–97.3) 47.5 (35.5–59.8) 90.6 (75.8–96.8) 52.2 (40.5–63.7) 
Yes 27 
Unevaluable 
Per-arterya 397 (3) No 239 17 56 71.8 (61.4–80.2) 76.6 (71.6–81.0) 90.9 (86.8–93.8) 45.5 (37.3–54.0) 
Yes 24 36 25 
Unevaluable 
Per-segmentb 953 (490) No 833 29 440 36.3 (27.1–46.5) 96.6 (95.2–97.6) 93.5 (91.7–94.9) 53.2 (41.0–65.1) 
Yes 58 33 50 
Unevaluable 47 

aIn the per-patient and per-artery analyses, unevaluable segments of major arterial branches were considered to have stenoses ≥50% in order to ensure clinical relevance of the analyses. The excluded patient and the three excluded arteries were those considered not fully evaluable by I-CA.

bIn the per-segment analysis,segments unevaluable by MSCT-CA were excluded.

Effect of heart rate, arterial calcification and body mass index on MSCT-CA accuracy

Rate-limiting medication was prescribed to 90% of patients (66% β-blockers, 24% calcium channel blockers). Mean heart rate during MSCT-CA was 68.8 (9.0) bpm. Average heart rate during MSCT-CA was <70 bpm in 54 patients and ≥70 bpm in 46 patients. The mean number of MSCT-CA evaluable segments per patient was 10.3 (3.1) in the <70 bpm group and 8.8 (3.6) in the ≥70 bpm group (P = 0.019). Correspondingly, the percentage of MSCT-CAs deemed overall not fully evaluable was 28.3 and 37.0% in the <70 bpm and ≥70 bpm groups, respectively. The accuracy parameters of MSCT-CA in the different heart rate groups demonstrated similar sensitivity and NPV but specificity and PPV were higher in patients with lower heart rates (Table 3).

Table 3

Accuracy of MSCT-CA in comparison to I-CA for the detection of CAD ≥50%: subgroup analyses

Sub-group Sn% (95% CI) Sp% (95% CI) NPV% (95% CI) PPV% (95% CI) 

 
Males 93.5 (79.3–98.2) 50.0 (31.4–68.6) 85.7 (60.1–96.0) 70.7 (55.5–82.4) 
Females 85.7 (48.7–99.3) 45.9 (31.0–61.6) 94.4 (74.2–99.7) 23.1 (11.0–42.1) 
Heart rate ≥70 bpm 92.3 (66.7–99.6) 39.4 (24.7–56.3) 92.9 (68.5–99.6) 37.5 (22.9–54.7) 
Heart rate <70 bpm 92.0 (75.0–97.8) 57.1 (39.1–73.5) 88.9 (67.2–96.9) 65.7 (49.2–79.2) 
No calcification 100.0 (5.1–100.0) 48.8 (34.3–63.5) 100.0 (83.9–100.0) 4.5 (0.2–21.8) 
Moderate/Heavy calcification 91.9 (78.7–97.2) 45.0 (25.8–65.8) 75.0 (46.8–91.1) 75.6 (61.3–85.8) 
BMI ≥25 kg/m2 92.3 (75.9–97.9) 45.8 (32.6–59.7) 91.7 (74.2–97.7) 48.0 (34.8–61.5) 
BMI <25 kg/m2 91.7 (64.6–99.6) 53.8 (29.1–76.8) 87.5 (52.9–99.4) 64.7 (41.3–82.7) 
Low/Intermediate PTP 75.0 (40.9–92.9) 52.0 (38.5–65.2) 92.9 (77.4–98.0) 20.0 (9.5–37.3) 
High PTP 96.7 (83.3–99.8) 27.3 (9.7–56.6) 75.0 (30.1–98.7) 78.4 (62.8–88.6) 
Sub-group Sn% (95% CI) Sp% (95% CI) NPV% (95% CI) PPV% (95% CI) 

 
Males 93.5 (79.3–98.2) 50.0 (31.4–68.6) 85.7 (60.1–96.0) 70.7 (55.5–82.4) 
Females 85.7 (48.7–99.3) 45.9 (31.0–61.6) 94.4 (74.2–99.7) 23.1 (11.0–42.1) 
Heart rate ≥70 bpm 92.3 (66.7–99.6) 39.4 (24.7–56.3) 92.9 (68.5–99.6) 37.5 (22.9–54.7) 
Heart rate <70 bpm 92.0 (75.0–97.8) 57.1 (39.1–73.5) 88.9 (67.2–96.9) 65.7 (49.2–79.2) 
No calcification 100.0 (5.1–100.0) 48.8 (34.3–63.5) 100.0 (83.9–100.0) 4.5 (0.2–21.8) 
Moderate/Heavy calcification 91.9 (78.7–97.2) 45.0 (25.8–65.8) 75.0 (46.8–91.1) 75.6 (61.3–85.8) 
BMI ≥25 kg/m2 92.3 (75.9–97.9) 45.8 (32.6–59.7) 91.7 (74.2–97.7) 48.0 (34.8–61.5) 
BMI <25 kg/m2 91.7 (64.6–99.6) 53.8 (29.1–76.8) 87.5 (52.9–99.4) 64.7 (41.3–82.7) 
Low/Intermediate PTP 75.0 (40.9–92.9) 52.0 (38.5–65.2) 92.9 (77.4–98.0) 20.0 (9.5–37.3) 
High PTP 96.7 (83.3–99.8) 27.3 (9.7–56.6) 75.0 (30.1–98.7) 78.4 (62.8–88.6) 

PTP: Pre-test probability by Duke clinical score;8 Sn: Sensitivity; Sp: Specificity.

Of 100 patients, 22 patients had moderate calcification and 35 patients had heavy calcification on MSCT-CA. Calcification was significantly more common in males than females (74.5 vs. 35.6%, P < 0.001) and with increasing age (P < 0.001). The mean number of MSCT-CA evaluable segments per patient with no or only moderate arterial calcification was 10.0 (3.1), whereas the mean number of evaluable segments in the presence of heavy calcification was 8.9 (3.9). This difference was not statistically significant. However, arterial calcification lowered NPV from 100.0 to 75.0% and increased PPV from 4.5% to 75.6% (Table 3).

Mean body mass index (BMI) was 28.6 (5.2) kg/m2 with no significant difference between males and females. Seventy-five per cent of the patients had a BMI ≥25 kg/m2 and 35% had a BMI ≥30 kg/m2. The percentage of unevaluable segments in the four BMI groups, <25 kg/m2, ≥25 kg/m2, <30 kg/m2 and ≥30 kg/m2 was 30, 32, 35 and 37%, respectively. This perceived difference was not, however, statistically significant and did not significantly affect accuracy parameters across the groups. There was no significant difference in presence or absence of arterial calcification or in mean heart rate during MSCT-CA between the BMI groups.

Effect of pre-test probability on MSCT-CA

Fifty-nine per cent of the patients had a low-intermediate pre-test probability and 41% a high pre-test probability of CAD. Significantly more males than females had a high pre-test probability (39 and 2, respectively, P < 0.001). There were no significant differences in BMI or mean heart rate between the pre-test probability groups but arterial calcification was present more frequently in the high pre-test probability group (87.8 vs. 35.6%, P < 0.001). The prevalence of significant CAD in the high pre-test probability group and the low-intermediate pre-test probability group was 73.2 and 13.8%, respectively. Correspondingly, the sensitivity and PPV of MSCT-CA were higher in the high pre-test probability group while specificity and NPV were lower (Table 3).

Inter-observer agreement

Inter-observer agreement κ between I-CA reporters and MSCT-CA reporters in the detection of significant CAD was 0.74 (95% CI 0.58–0.87) and 0.61 (95% CI 0.38–0.85), respectively, for stenoses ≥50% and 0.84 (95% CI 0.60–0.89) and 0.83 (95% CI 0.45–1.00), respectively, for stenoses ≥70%. For this analysis of agreement unevaluable segments were excluded.

Discussion

This is the first UK study comparing MSCT-CA to I-CA in a district general hospital. Only two previous studies (both Scandinavian) have reported 40–64-Slice MSCT-CA accuracy explicitly in a district hospital setting.12,13 In the per-patient analysis, sensitivity, specificity, PPV and NPV for 40-Slice MSCT-CA were 92, 48, 52 and 91%, respectively. While our results are at variance with a recently published meta-analysis of 64-Slice MSCT-CA accuracy where sensitivity, specificity, PPV and NPV were 99, 89, 93 and 100%, respectively,4 they are comparable with those of CATSCAN, the first multi-centre study of MSCT-CA accuracy,14 and to a previous Norwegian study of 16-Slice MSCT-CA in a community hospital setting,15 where specificity and PPV were also low (54 and 50% and 29 and 57%, respectively).

MSCT-CA accuracy in this district general hospital study was compromised by a high number of unevaluable segments. Based on the AHA 15-segment model, 34% of all coronary artery segments were deemed unevaluable by MSCT-CA. This is in contrast to previous studies where the number of unevaluable segments is typically <10%.3,4 Specificity and PPV in this study were lowered by our strategy for dealing with those unevaluable segments. In the per-patient analysis, all MSCT-CAs that were not fully evaluable were considered positive for significant CAD which in effect increased the number of false positive scans. We consider this strategy clinically relevant, however, as in practice a patient with an unevaluable MSCT-CA would proceed to I-CA. This analytical approach was not adopted by all authors of previous studies with many discounting unevaluable segments in the per-patient analysis.

Sub-optimal heart rate control during MSCT-CA undoubtedly contributed to the high percentage of unevaluable segments in this study. Despite, 90% of the patients in our study taking rate controlling medication, just over half had heart rates <70 bpm and only a third had heart rates <65 bpm. This clearly influenced MSCT-CA image quality with significantly more unevaluable segments in the higher heart rate group and a noticeable deterioration in specificity and PPV. We utilized a multi-segment reconstruction algorithm at higher heart rates in an attempt to overcome motion artefact. While previous studies have demonstrated better image quality with this strategy,16,17 others have indicated difficulties with the technique due to variations in coronary artery position with consecutive cardiac contractions.18,19 Recent work has confirmed improved accuracy of MSCT-CA with lower heart rates achieved by intravenous β blockade.20 This may have improved accuracy in our study.

Segment evaluability may also have been adversely affected by characteristics of the patient population studied. More than one third of the patients were obese with BMI ≥30 kg/m2 while over half had arterial calcification on MSCT-CA. Increasing BMI correlates with increasing image noise21 and has a significant and independent impact on image quality.22 Arterial calcification degrades image quality due to partial volume effects and bloom artefacts20,23 (Figure 1). Consequently, studies of MSCT-CA accuracy commonly exclude patients with high Agatston calcium scores.14,24 In our study, the small noted reductions in the number of evaluable segments in the context of elevated BMI or arterial calcification were not statistically significant. However, incorrect reporting of ‘evaluable’ segments due to artefact may have adversely influenced accuracy.

Figure 1.

(a) I-CA LAO Caudal view demonstrating significant left main stem, circumflex and left anterior descending artery stenoses in an elderly gentleman, (b) MSCT-CA volume rendered image of the elderly gentleman in (a) demonstrating heavily calcified arteries which compromised evaluation of luminal stenoses, (c) I-CA of RCA in LAO 30° demonstrating a significant proximal RCA stenosis >50% and (d) MSCT-CA curved multi-planar reconstruction demonstrating the same proximal RCA stenosis as (c) with severity of stenosis easily assessed due to the presence of soft rather than calcified plaque.

Figure 1.

(a) I-CA LAO Caudal view demonstrating significant left main stem, circumflex and left anterior descending artery stenoses in an elderly gentleman, (b) MSCT-CA volume rendered image of the elderly gentleman in (a) demonstrating heavily calcified arteries which compromised evaluation of luminal stenoses, (c) I-CA of RCA in LAO 30° demonstrating a significant proximal RCA stenosis >50% and (d) MSCT-CA curved multi-planar reconstruction demonstrating the same proximal RCA stenosis as (c) with severity of stenosis easily assessed due to the presence of soft rather than calcified plaque.

MSCT-CA accuracy evidently varies with CAD prevalence. The NPV of MSCT-CA for ruling out significant CAD on a per-patient basis in our study was 91%. As expected, the NPV was higher (93%) in the low-intermediate pre-test probability group where the prevalence of CAD was lower. Conversely, NPV in the high pre-test probability group was 75%. This is consistent with previous studies where CAD prevalence was high. In the multicentre study CorE-64, the prevalence of CAD was 56% and NPV in the per-patient analysis was 83%.24 Similarly, a recent study of 40-Slice MSCT-CA where the prevalence of CAD was 77% reported NPV in the per-patient analysis to be 55%.25

Appropriate patient selection is imperative in order to emulate the high accuracy of MSCT-CA demonstrated in studies from large academic centres in routine practice. The training and experience of those involved in performing and reporting the scans is equally important. The need for radiologists and cardiologists to undergo appropriate training and to achieve accreditation in MSCT-CA is recognized in international guidelines.5,26 Level 2 competence is defined as the minimum experience required for independent performing and interpretation of MSCT-CA. In our study, one of the radiologists had achieved Level 2 competence prior to commencing the study and the other completed this during the study. The cardiologist involved in resolving discrepancies between the two radiologists’ reports was Level 2 competent and the most experienced of the three reporters. Notably the less experienced radiologist’s reported MSCT-CA accuracy was equivalent to the Level 2 competent radiologist’s accuracy when results were separated out per reporter. While the relative inexperience of our centre may have adversely affected diagnostic accuracy, we feel our reporters were representative of those likely to be performing and reporting MSCT-CA in district hospitals. Our findings are consistent with the multi-centre studies of MSCT-CA where accuracy was reduced even despite reporting being standardized in core laboratories.14,24,27

MSCT-CA is becoming increasingly accessible in both regional and district hospitals and appears attractive in comparison to I-CA as it permits a non-invasive evaluation of CAD. However, MSCT-CA and I-CA have a number of common limitations. Both MSCT-CA and I-CA in isolation provide anatomical information rather than functional data and, therefore, do not permit assessment of myocardial perfusion or ischaemia. Furthermore, both techniques require the administration of iodinated contrast and a significant radiation dose. MSCT-CA is further limited by its inferior temporal and spatial resolution and, as this study has demonstrated, image quality is adversely affected by common patient characteristics such as coronary artery calcification, elevated BMI and poorly controlled heart rates. Indeed, the potential requirement for intravenous β-blockade during MSCT-CA presents a further complication. One limitation of our study was the use of a 40-Slice MSCT scanner rather than the now widely distributed 64-Slice scanner. However, 40-Slice scanners are not considered significantly inferior to their 64-Slice counterparts with published accuracy parameters comparable with those from 64-Slice studies.28–30 The development of the now commercially available dual source and 320-Slice scanners may overcome many of the technical difficulties of 40–64-Slice scanning and will certainly reduce the requisite radiation exposure.31,32 However, until NHS funding is sufficient to install these scanners in multiple sites, there will be continued use of 40–64-Slice scanners and even 16-Slice scanners and so the findings of our study retain clinical relevance.

Conclusion

This study demonstrates the high negative predictive value of 40-Slice MSCT-CA for ruling out significant CAD when performed in a district hospital setting in Scottish patients with a low-intermediate pre-test probability. The Duke Clinical Score appeared a reliable method for identifying those patients. Specificity and positive predictive value of MSCT-CA are compromised by clinically appropriate strategies for dealing with unevaluable studies. Future MSCT-CA in our population should be limited to patients at low-intermediate risk with minimal arterial calcification. Effective heart rate control during MSCT-CA is imperative. National guidelines should be utilized to govern patient selection and direct appropriate MSCT-CA reporter training and accreditation to ensure quality control.

Funding

The Chief Scientist Office of the Scottish Executive who also approved the study design (CZG/2/266).

Conflict of interest: None declared.

Acknowledgements

We would like to acknowledge the radiographers and nursing staff of the Department of Radiology at Stobhill Hospital without whom this work would not have been possible.

References

1
British Heart Foundation
Scotland Coronary Heart Disease Statistics 2009–2010
 , 
2009
 
[www.heartstats.org] Accessed 15 September 2010
2
de Bono
D
Complications of diagnostic cardiac catheterisation: results from 34,041 patients in the United Kingdom confidential enquiry into cardiac catheter complications. The Joint Audit Committee of the British Cardiac Society and Royal College of Physicians of London
Br Heart J
 , 
1993
, vol. 
70
 (pg. 
297
-
300
)
3
Abdulla
J
Abildstrom
SZ
Gotzsche
O
Christensen
E
Kober
L
Torp-Pedersen
C
64-multislice detector computed tomography coronary angiography as potential alternative to conventional coronary angiography: a systematic review and meta-analysis
Eur Heart J
 , 
2007
, vol. 
28
 (pg. 
3042
-
50
)
4
Mowatt
G
Cook
JA
Hillis
GS
Walker
S
Fraser
C
Jia
X
, et al.  . 
64-Slice computed tomography angiography in the diagnosis and assessment of coronary artery disease: systematic review and meta-analysis
Heart
 , 
2008
, vol. 
94
 (pg. 
1386
-
93
)
5
Schroeder
S
Achenbach
S
Bengel
F
Burgstahler
C
Cademartiri
F
de Feyter
P
, et al.  . 
Cardiac computed tomography: indications, applications, limitations, and training requirements: report of a Writing Group deployed by the Working Group Nuclear Cardiology and Cardiac CT of the European Society of Cardiology and the European Council of Nuclear Cardiology
Eur Heart J
 , 
2008
, vol. 
29
 (pg. 
531
-
56
)
6
National Institute for Health and Clinical Excellence
Chest pain of recent onset: assessment and diagnosis of recent onset chest pain or discomfort of suspected cardiac origin (clinical guideline 95)
 , 
2010
 
[www.nice.org.uk/guidance/CG95] Accessed 15 September 2010
7
Budoff
MJ
Achenbach
S
Blumenthal
RS
Carr
JJ
Goldin
JG
Greenland
P
, et al.  . 
Assessment of coronary artery disease by cardiac computed tomography: a scientific statement from the American Heart Association Committee on Cardiovascular Imaging and Intervention, Council on Cardiovascular Radiology and Intervention, and Committee on Cardiac Imaging, Council on Clinical Cardiology
Circulation
 , 
2006
, vol. 
114
 (pg. 
1761
-
91
)
8
Pryor
DB
Shaw
L
McCants
CB
Lee
KL
Mark
DB
Harrell
FE
, et al.  . 
Value of the history and physical in identifying patients at increased risk for coronary artery disease
Ann Intern Med
 , 
1993
, vol. 
118
 (pg. 
81
-
90
)
9
Austen
WG
Edwards
JE
Frye
RL
Gensini
GG
Gott
VL
Griffith
LS
, et al.  . 
A reporting system on patients evaluated for coronary artery disease. Report of the Ad Hoc Committee for Grading of Coronary Artery Disease, Council on Cardiovascular Surgery, American Heart Association
Circulation
 , 
1975
, vol. 
51
 
Supp 4
(pg. 
5
-
40
)
10
Mollet
NR
Cademartiri
F
Nieman
K
Saia
F
Lemos
PA
McFadden
EP
, et al.  . 
Multislice spiral computed tomography coronary angiography in patients with stable angina pectoris
J Am Coll Cardiol
 , 
2004
, vol. 
43
 (pg. 
2265
-
70
)
11
JBS 2: Joint British Societies’ guidelines on prevention of cardiovascular disease in clinical practice
Heart
 , 
2005
, vol. 
91
 
Suppl 5
(pg. 
v1
-
52
)
12
Halvorsen
BA
Rodevand
O
Hagen
G
Herud
E
Mielczarek
W
Molstad
P
[Angiography with 64-channel CT upon suspicion of stable coronary disease]
Tidsskr Nor Laegeforen
 , 
2008
, vol. 
128
 (pg. 
2172
-
6
)
13
Ovrehus
KA
Munkholm
H
Bottcher
M
Botker
HE
Norgaard
BL
Coronary computed tomographic angiography in patients suspected of coronary artery disease: impact of observer experience on diagnostic performance and interobserver reproducibility
J Cardiovasc Comput Tomogr
 , 
2010
, vol. 
4
 (pg. 
186
-
94
)
14
Garcia
MJ
Lessick
J
Hoffmann
MH
Accuracy of 16-row multidetector computed tomography for the assessment of coronary artery stenosis
JAMA
 , 
2006
, vol. 
296
 (pg. 
403
-
11
)
15
Rodevand
O
Hogalmen
G
Gudim
LP
Indrebo
T
Molstad
P
Vandik
PO
Limited usefulness of non-invasive coronary angiography with 16-detector multislice computer tomography at a community hospital
Scand Cardiovasc J
 , 
2006
, vol. 
40
 (pg. 
76
-
82
)
16
Dewey
M
Laule
M
Krug
L
Schnapauff
D
Rogalla
P
Rutsch
W
, et al.  . 
Multisegment and halfscan reconstruction of 16-slice computed tomography for detection of coronary artery stenoses
Invest Radiol
 , 
2004
, vol. 
39
 (pg. 
223
-
9
)
17
Greuter
MJ
Flohr
T
van Ooijen
PM
Oudkerk
M
A model for temporal resolution of multidetector computed tomography of coronary arteries in relation to rotation time, heart rate and reconstruction algorithm
Eur Radiol
 , 
2007
, vol. 
17
 (pg. 
784
-
812
)
18
Halliburton
SS
Stillman
AE
Flohr
T
Ohnesorge
B
Obuchowski
N
Lieber
M
, et al.  . 
Do segmented reconstruction algorithms for cardiac multi-slice computed tomography improve image quality?
Herz
 , 
2003
, vol. 
28
 (pg. 
20
-
31
)
19
Herzog
C
Nguyen
SA
Savino
G
Zwerner
PL
Doll
J
Nielsen
CD
, et al.  . 
Does two-segment image reconstruction at 64-section CT coronary angiography improve image quality and diagnostic accuracy?
Radiology
 , 
2007
, vol. 
244
 (pg. 
121
-
9
)
20
Brodoefel
H
Reimann
A
Burgstahler
C
Schumacher
F
Herberts
T
Tsiflikas
I
, et al.  . 
Noninvasive coronary angiography using 64-slice spiral computed tomography in an unselected patient collective: effect of heart rate, heart rate variability and coronary calcifications on image quality and diagnostic accuracy
Eur J Radiol
 , 
2008
, vol. 
66
 (pg. 
134
-
41
)
21
Yoshimura
N
Sabir
A
Kubo
T
Lin
PJ
Clouse
ME
Hatabu
H
Correlation between image noise and body weight in coronary CTA with 16-row MDCT
Acad Radiol
 , 
2006
, vol. 
13
 (pg. 
324
-
8
)
22
Brodoefel
H
Tsiflikas
I
Burgstahler
C
Reimann
A
Thomas
C
Schroeder
S
, et al.  . 
Cardiac dual-source computed tomography: effect of body mass index on image quality and diagnostic accuracy
Invest Radiol
 , 
2008
, vol. 
43
 (pg. 
712
-
8
)
23
Ong
TK
Chin
SP
Liew
CK
Chan
WL
Seyfarth
MT
Liew
HB
, et al.  . 
Accuracy of 64-row multidetector computed tomography in detecting coronary artery disease in 134 symptomatic patients: influence of calcification
Am Heart J
 , 
2006
, vol. 
151
 (pg. 
1323
-
6
)
24
Miller
JM
Rochitte
CE
Dewey
M
Arbab-Zadeh
A
Niinuma
H
Gottlieb
I
, et al.  . 
Diagnostic performance of coronary angiography by 64-row CT
N Engl J Med
 , 
2008
, vol. 
359
 (pg. 
2324
-
36
)
25
Halon
DA
Gaspar
T
Adawi
S
Rubinshtein
R
Schliamser
JE
Peled
N
, et al.  . 
Uses and limitations of 40 slice multi-detector row spiral computed tomography for diagnosing coronary lesions in unselected patients referred for routine invasive coronary angiography
Cardiology
 , 
2007
, vol. 
108
 (pg. 
200
-
9
)
26
Budoff
MJ
Cohen
MC
Garcia
MJ
Hodgson
JM
Hundley
WG
Lima
JA
, et al.  . 
ACCF/AHA clinical competence statement on cardiac imaging with computed tomography and magnetic resonance: a report of the American College of Cardiology Foundation/American Heart Association/American College of Physicians Task Force on Clinical Competence and Training
J Am Coll Cardiol
 , 
2005
, vol. 
46
 (pg. 
383
-
402
)
27
Budoff
MJ
Dowe
D
Jollis
JG
Gitter
M
Sutherland
J
Halamert
E
, et al.  . 
Diagnostic performance of 64-multidetector row coronary computed tomographic angiography for evaluation of coronary artery stenosis in individuals without known coronary artery disease: results from the prospective multicenter ACCURACY (Assessment by Coronary Computed Tomographic Angiography of Individuals Undergoing Invasive Coronary Angiography) trial
J Am Coll Cardiol
 , 
2008
, vol. 
52
 (pg. 
1724
-
32
)
28
Lim
MC
Wong
TW
Yaneza
LO
De Larrazabal
C
Lau
JK
Boey
HK
Non-invasive detection of significant coronary artery disease with multi-section computed tomography angiography in patients with suspected coronary artery disease
Clin Radiol
 , 
2006
, vol. 
61
 (pg. 
174
-
80
)
29
Pouleur
AC
le Polain de Waroux
JB
Kefer
J
Pasquet
A
Coche
E
Vanoverschelde
JL
, et al.  . 
Usefulness of 40-slice multidetector row computed tomography to detect coronary disease in patients prior to cardiac valve surgery
Eur Radiol
 , 
2007
, vol. 
17
 (pg. 
3199
-
207
)
30
Watkins
MW
Hesse
B
Green
CE
Greenberg
NL
Manning
M
Chaudry
E
, et al.  . 
Detection of coronary artery stenosis using 40-channel computed tomography with multi-segment reconstruction
Am J Cardiol
 , 
2007
, vol. 
99
 (pg. 
175
-
81
)
31
Achenbach
S
Ropers
U
Kuettner
A
Anders
K
Pflederer
T
Komatsu
S
, et al.  . 
Randomized comparison of 64-slice single- and dual-source computed tomography coronary angiography for the detection of coronary artery disease
JACC Cardiovasc Imaging
 , 
2008
, vol. 
1
 (pg. 
177
-
86
)
32
Rybicki
FJ
Otero
HJ
Steigner
ML
Vorobiof
G
Nallamshetty
L
Mitsouras
D
, et al.  . 
Initial evaluation of coronary images from 320-detector row computed tomography
Int J Cardiovasc Imaging
 , 
2008
, vol. 
24
 (pg. 
535
-
46
)