Materials and methods

Study population
The study population was derived from an existing database, which consisted of 402 patients with chest pain suggestive of stable angina pectoris and suspected of having CAD. All patients were prospectively included in a large study evaluating 64-slice CT coronary angiography (CTCA) at our institution. All patients were referred for conventional coronary angiography (CCA) based on their presentation or functional testing that suggested the presence of ischaemia and all patients underwent multidetector CT angiography within a week before CCA. Inclusion criteria for this study were: informed consent, sinus heart rhythm and the ability to hold their breath for 15 s. Patients with a history of percutaneous coronary intervention or coronary artery bypass surgery, impaired renal function (serum creatinine >120 μmol/L) or a known intolerance to iodinated contrast medium were excluded. The Institutional Review Board approved the study and all patients signed informed consent. As this paper focuses on patients with new onset stable chest pain, we also excluded patients with acute coronary syndromes and patients with a previous myocardial infarction (Fig. 1).
Fig. 1 Flow chart of patients in the study. CTCA computed tomography coronary angiography. *Data from an existing database were used. All patients were referred for conventional coronary angiography based on their presentation or functional testing that suggested the presence of cardiac ischaemia. See Materials and methods

CT coronary calcium images
Metoprolol (100 mg, Selokeen, AstraZeneca, London, UK) was administered orally 1 h before CT in patients with heart rates >65 beats per minute. A 64-slice single source CT system (Sensation 64; Siemens, Forchheim, Germany) with a gantry rotation time of 330 ms, acquisition time of 165 ms and voxel size of 0.4 mm3 was used to acquire standard spiral low-dose and ECG-gated coronary calcium CT images. CT parameters were 32 × 2 slices per rotation, individual detector width of 0.6 mm, 3.8-mm/rotation table feed, 120-kV tube voltage, 150-mAs tube current, with activated prospective x-ray tube modulation. Overlapping slices were reconstructed at 65% of the R–R interval using the B35f convolution kernel. Reconstructed slice thickness was 3.0 mm with an increment of 1.5 mm. The radiation exposure, estimated using dedicated software (ImPACT, version 0.99x, St. George’s Hospital, Tooting, London, UK), was 1.4 mSv in men and 1.8 mSv in women. One observer (with more than 3 years’ experience), who was blinded to the CCA and clinical data, measured the coronary calcium.

Conventional coronary angiography
The CTCS and CCA were carried out within 1 week. Coronary segments were assessed on CCA following a 17-segment modified American Heart Association (AHA) classification model [14] by a single observer (with more than 10 years’ experience), who was blinded to the CT and clinical data. A mean luminal narrowing of ≥50% was considered to be a significant stenosis. Validated quantitative coronary angiography software (CAAS II®, Pie Medical, Maastricht, the Netherlands) was used.

Clinical variables and outcome
All patients were interviewed at enrollment in the prospective cohort study. Clinical parameters recorded were: age (years), sex (male/female), type of chest pain (atypical vs. typical), body mass index (BMI) (defined as weight/height2 in kg/m2), smoking status (past or current smoker, yes/no), hypertension (present/absent), dyslipidaemia (serum cholesterol >200 mg/dL or 5.18 mmol/L, present/absent), diabetes (plasma glucose ≥126 mg/dL or 7.0 mmol, present/absent) and family history of CAD (present/absent).
The CTCS was measured by the Agatston method [15] using dedicated software (syngo Calcium Scoring VE31H, Siemens, Germany). The outcome of interest was the presence of obstructive CAD defined as ≥50% stenosis in at least one vessel (present/absent) on CCA.

Sample size
As a general rule, 10 patients with the condition of interest per analysed variable are required for regression analysis. In our dataset (n = 254), 123 patients were identified as having obstructive CAD on CCA. This allowed for the analysis of 12 variables. Our sample meets the required number of cases and non-cases that has been suggested for external validation of prediction models [16].

Systematic literature search
We searched the English-language medical literature in PubMed up to October 14, 2009 for diagnostic prediction models. See the Appendix for a detailed description of the search strategy. From the included articles, clinical variables that were identified as significant predictors of CAD were extracted.

Data analysis
Age was analysed as a continuous variable. To account for the skewed distribution of the coronary calcium scoring, CTCS was transformed by taking the natural logarithm of CTCS + 1. All other variables were dichotomous. Oestrogen status was not available in our dataset. Therefore, we assumed women below the age of 50 to be oestrogen positive, women of 50 years and above to be oestrogen negative and all men to be oestrogen neutral. Obesity was considered in the model by Morise (1997) only. We defined obesity as a BMI >27 kg/m2, corresponding to their definition [17].
The extracted sets of clinical variables were analysed with multivariate logistic regression analysis, fitting new regression coefficients. No attempt was made to validate original regression coefficients, as such coefficients were often not reported. CTCS was subsequently included in each of the models. Models without CTCS were compared with corresponding models including CTCS using the likelihood ratio test. The level of significance was defined at a p value less than 0.05.
Diagnostic performance was assessed by calculating the area under the receiver operating characteristic (ROC) curve, the c-index. The c-index is a measure of discrimination and is interpreted as being the probability that a randomly chosen patient with CAD will have a higher predicted probability of disease than a randomly chosen patient without CAD [18]. An area under the ROC curve (AUC) of 0.5 corresponds to a model that provides no diagnostic information, whereas an AUC of 1.0 corresponds to a perfect diagnostic model.
STATA statistical analysis software v10.0 (StataCorp, Texas, USA) was used for logistic regression analysis.
Next, we quantified the effect of adding CTCS to the model on the classification of patients into probability categories of CAD. Four probability categories were defined: <30%, ≥30–50%, ≥50–70% and ≥70%. Reclassification tables were constructed for the Diamond & Forrester model and the Pryor model (see Tables 4 and 5) [19]. We computed the reclassification calibration statistic (RCS) [20] which is equivalent to the Hosmer–Lemeshow statistic, applied to the cross-classified cells of the reclassification table with at least 20 observations. A significant result indicates a lack of fit.
Furthermore, the following reclassification measures were calculated for each model: the overall (correct) percentage of reclassification, the net reclassification improvement (NRI) [21] and the integrated discrimination improvement (IDI) [20]. The NRI is the difference in proportions reclassifying to higher and lower probability categories among cases and non-cases. It is interpreted as the percentage reclassified, adjusted for the reclassification direction. A significant NRI indicates that classification improves when CTCS is included. The IDI compares the difference in the average regression slope of cases and non-cases among the models with and without CTCS. A significant IDI indicates that the new model performs better in discriminating cases and non-cases.
Reclassification computations were executed by using syntax made available by Cook and Ridker [20] in SAS Enterprise Guide v3 (SAS Inc, North Carolina, USA).