New generation cardiac CT scanners (Aquilion ONE, Brilliance iCT, Discovery CT750 HD and Somatom Definition Flash) for cardiac imaging in people with suspected or known coronary artery disease in whom imaging is difficult with earlier generation CT scanners

The scope assumed that Aquilion ONE, Brilliance iCT, Discovery CT750 HD and Somatom Definition Flash provided broadly similar benefits for people in whom imaging is difficult. The assessment was carried out on this basis. The committee considered comments on the diagnostics assessment report from stakeholders about the equivalence of the 4 scanners. The committee heard that the scanners use different technologies to gain image improvement and dose reduction, and that the different approaches may provide differential benefits depending on the reasons for imaging being difficult. For the purposes of this assessment, the committee considered that it was reasonable to assume that each of the devices was capable of providing adequate imaging for all groups of people in whom imaging is difficult. Moreover, the committee concluded that many people have combinations of conditions that lead to imaging difficulties and, therefore, it would not be optimal to specify use of any 1 particular device.

Of the 24 test accuracy studies included in the diagnostics assessment report, 3 studies did not specify the model of the scanner used, 1 study used Somatom Definition Flash, 1 used Aquilion ONE and 19 used Somatom Definition (a model predating the Somatom Definition Flash). The manufacturer of the Discovery CT750 HD confirmed the CT750 HD had been used in a study in which the details of the scanner were not provided. Although test accuracy data exist for all of the systems considered, most studies did not stratify outcome reporting for the difficult-to-image subsets and thus the data could not be used in the assessment. The assessment was based on the 24 studies with stratified data. The committee concluded that the evidence base for these technologies provided it with sufficient certainty for it to formulate recommendations. However, the committee considered it important that manufacturers make available specific test accuracy data stratified by important patient subgroups and noted that the current lack of such data made it difficult to make recommendations. The committee noted that it would only make recommendations on the scanners included in the scope, although evidence from scanners other than the 4 being evaluated was included in the assessment.

The committee discussed the extent to which the usable test accuracy studies, particularly the studies of the older Somatom model, could be generalised across scanners. The committee heard from the external assessment group that statistical tests of the test accuracy results of the different scanners showed the results were not heterogeneous and thus could be combined, and the committee accepted this explanation.

The committee concluded that even though no usable data were available for Brilliance iCT, and one study each was available for Aquilion ONE, Discovery CT750 HD and Somatom Definition Flash, nevertheless it was reasonable for the committee to make recommendations that applied to all 4 of the assessed scanners.

The committee considered whether all relevant studies had been included in the assessment. In response to comments from representatives of manufacturers about the inclusion of relevant studies, the external assessment group confirmed that the literature search was designed to be very sensitive and to maximise identification of papers containing test accuracy data for the groups of patients included in the scope. In addition, all manufacturers had been asked to provide relevant data. Unfortunately, most of the data provided by manufacturers could not be included because it did not contain test accuracy results stratified for the relevant populations for the assessment.

The committee acknowledged that the reference standard, invasive coronary angiography, although not 100% accurate in clinical practice, is an accepted reference standard for assessment of anatomic coronary disease. The high accuracy of CT when compared with angiography, coupled with the known inaccuracies of angiography, imply that CT may be even more effective and cost effective than the assessment indicated because it may be more accurate, and angiography less accurate, than modelled.

The committee acknowledged that from a patient perspective, a non-invasive cardiac diagnostic test is more appealing than invasive coronary angiography because of the greater morbidity and mortality risks associated with angiography. The external assessment group informed the committee that the modelling had reflected this preference to some extent, because it included the increased morbidity and mortality resulting from invasive procedures as well as their associated costs. CT was found to be more cost effective than angiography because of the lower risk of these outcomes and the reduced costs associated with CT (reduced imaging costs and reduced downstream healthcare costs from dealing with complications), even though angiography had been assumed to be the more accurate test.

The committee considered how using new generation cardiac CT scanners for evaluating people with suspected coronary artery disease would fit in the context of the NICE guideline on recent-onset chest pain. The committee concluded that the evidence presented indicated that new generation cardiac CT was more cost effective for people in whom imaging is difficult than proceeding directly to invasive angiography. The committee noted that earlier generation CT scanners used for people in whom imaging is not difficult are even less expensive and have similar risks and benefits than the new generation scanners. CT could be more cost effective than angiography for all people presenting with chest pain and a pre-test likelihood of 10% to 29% of coronary artery disease.

The NICE guideline on recent-onset chest pain, before the 2016 update, recommended calcium scoring to assess patients with an estimated likelihood of coronary artery disease of 10% to 29%, and varied its further imaging recommendations based on the calcium level. The value of calcium scoring was outside the scope of this evaluation and, thus, not explored further by the committee.

The committee considered whether angiography after CT was needed for people with positive CT scans. The analysis showed that CT alone was more cost effective than CT with angiography for people with suspected coronary artery disease, but CT alone was not more cost effective for people with known coronary artery disease. For both groups, CT followed by invasive coronary angiography for those with positive CT scans was more cost effective than invasive coronary angiography alone.

The committee was informed that, in current practice, people who are expected to have revascularisation would usually have angiography either before or as a part of treatment, but that the rate of elective angiography was dropping. The committee also heard that with time and additional clinician experience with CT angiography, it was likely that follow-up angiography solely for diagnostic purposes would become less frequent. A negative CT scan result would be sufficient to avoid angiography and often low-risk patients with a negative scan could be discharged from specialty care immediately. The committee heard that if a CT scan result shows moderate stenosis (40% to 60%), then the patient usually proceeds to functional imaging, whereas for severe stenosis (80%), the patient would be offered invasive coronary angiography. The committee was advised that having the option to diagnose with CT before proceeding to functional imaging and/or invasive coronary angiography would not affect the throughput of a cardiology department.

The committee considered whether it was appropriate to recommend CT scanning for people with known coronary artery disease in whom imaging was not difficult. However, evidence for this population had not been examined and no analysis had been undertaken.

The committee also considered whether it was appropriate to recommend CT scanning for people with suspected coronary artery disease who had prior likelihoods of coronary artery disease higher than 10 to 29%. The sensitivity analysis had shown that the use of new generation CT scanning is within NICE's standard levels of cost effectiveness compared with angiography for people in whom imaging is difficult and who have moderately higher than 29% pre-test probabilities of coronary artery disease. The committee heard that immediate revascularisation was no longer considered the preferred practice in this population. However, because data for this population had not been examined, it was not possible to make a recommendation.

The committee concluded that, based on the results of the assessment carried out by the external assessment group, new generation cardiac CT scanners should be recommended for:

• first-line imaging of coronary arteries in people with suspected stable coronary artery disease whose estimated likelihood of coronary artery disease is 10% to 29%, and

• first-line evaluation of disease progression in people with known coronary artery disease in whom imaging is difficult.

The committee was informed that there are currently up to 40 new generation cardiac CT scanners being used in the NHS in England. The committee considered that its recommendations would help optimise the use of the scanners that are currently in operation. The committee acknowledged that the impact of its recommendations would support the further introduction of new generation cardiac CT scanners in the NHS over the next few years.

The committee considered if there were any specific equalities issues that would be relevant to this assessment, but none were raised.