Ischemia-guided Management Using Cardiac SPECT: Reconciling Real-World Evidence in a post-ISCHEMIA Trial World

Results from adequately powered, multicenter, prospective outcomes trials provide the highest-level evidence to guide patient management. Historically, the widespread utilization of cardiovascular functional tests for the evaluation of symptomatic patients with suspected coronary artery disease (CAD) was based on large-scale, retrospective, non-randomized observational studies that demonstrated a positive relationship between a patients’ burden of ischemia and future risk for myocardial infarction and death. Such studies utilizing single photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) also suggested that routine coronary revascularization improved event-free survival in patients with ischemia involving >10% of the left ventricular myocardium.¹ Taken together, these data served to promote the broad clinical adoption of ischemia-guided management in patients with established or suspected stable CAD, to include routine coronary revascularization in patients with significant ischemia on stress imaging.² Many cardiologists and previous guideline-writers may have assumed that the practice of ischemia-driven revascularization for chronic CAD would be proven effective if a prospective outcomes trial testing this practice were ever performed, and some likely wondered if such a trial was even necessary.

The landmark, National Heart Lung and Blood Institute-funded ISCHEMIA (International Study of Comparative Health Effectiveness With Medical and Invasive Approaches) trial, a large, multicenter, prospective randomized comparative effectiveness study that compared an initial invasive strategy versus contemporary medical therapy in stable patients with significant ischemia on functional testing (50% having undergone SPECT), was just such a trial.^{3, 4} The results, which have been broadly discussed, showed no overall benefit of an invasive strategy with goal for coronary revascularization on all-cause death, death from cardiovascular causes, myocardial infarction (MI) or the combination primary endpoint in 5179 randomized patients followed for a median of 3.2 years. Interestingly, core-lab interpretations of SPECT studies demonstrated that 15% of site-read studies interpreted as showing at least moderate ischemia demonstrated mild or no ischemia per core lab interpretation. As most patients in the ISCHEMIA trial underwent coronary computed tomographic angiography (CCTA) to ensure the presence of angiographically significant CAD, investigators compared the prognostic value of CAD severity on CCTA versus ischemia severity on SPECT core lab interpretation (12% none/mild, 33% moderate, 55% severe ischemia) and other functional tests.⁵ Both increasing CAD severity on CCTA, using the Duke Prognostic Index which incorporates stenosis severity and location, and the severity of ischemia were independently associated with increased mortality. However, when adjusted for CCTA CAD severity, ischemia severity did not identify a subgroup with treatment benefit from revascularization on mortality, MI, the trial combined primary end point, or cardiovascular death or MI. These results challenged the age-old assumption that ischemia should be the arbiter for revascularization decision-making. As a result, many argued for increased utilization of CCTA in patients with suspected CAD so that revascularization decisions might be optimally made based on symptom severity plus anatomic information, without ischemia testing. Additionally, due to increasing recognition that non-obstructive atherosclerosis, plaque burden, and high-risk plaque features uniquely identified on CCTA may better inform decisions regarding preventive treatments in patients without known CAD, and recent guidelines elevating CCTA to a Class I first-line test in symptomatic patients, CCTA has seen increasing worldwide clinical utilization.^{2, 6-8}

In this issue of Circulation: Cardiovascular Imaging, Patel and colleagues present real-world evidence describing the prognostic value of ischemia on SPECT MPI and the impact of subsequent coronary revascularization on outcomes in a single-center observational study.⁹ The authors evaluated 1764 consecutive patients (76% men) from 2010-2021 at a single Italian center with known (29.3%) or suspected CAD who underwent rest/stress Tc-99 SPECT MPI and had either CCTA or invasive coronary angiography (ICA) within the following 1 month. Importantly, patients who underwent SPECT MPI and did not undergo subsequent early CCTA or ICA (N=10,198) were excluded, highlighting the selected nature of this cohort. Post-SPECT patient management was left to the discretion of treating physicians, with institutional practice patterns generally such that patients with normal perfusion (summed stress score [SSS] <4) were referred for CCTA and those with SSS ≥4 referred for ICA. Stenosis severity was assessed combining both CCTA and ICA angiographic findings to calculate the CAD Prognostic Index (CADPI). Importantly, this index only considers angiographic stenosis severity, number of vessels, and location, and does not assign any risk or consider the extent or severity of non-obstructive disease (all <50% stenoses = CADPI score of 0), plaque burden, or high-risk plaque features on CCTA.⁹ Among the cohort, 28.4% had >10% ischemia on SPECT MPI and the median follow-up was 4.6 years for the outcome of cardiac death or MI (8.4% with events). A total of 32.8% of subjects underwent early coronary revascularization (<90 days following index MPI).

The authors demonstrate that compared to clinical risk factors, both % ischemia and stenosis provided independent and incremental prognostic value. In a multivariable model containing both ischemia (5% increment) and stenosis severity (CADPI score increments of 10), ischemia severity was a stronger predictor of adverse events. Among subjects who underwent early revascularization, there was an association with reduced cardiac death or MI in those with >10% ischemia, with the authors recreating the classic figure from Hachamovitch and colleagues who analyzed SPECT MPI and outcomes data from a large, retrospective, observational cohort.¹ Patel and colleagues conclude that ischemic burden provides independent and incremental prognostic value beyond angiographic CAD severity and can identify patients most likely to benefit from early coronary revascularization.

So how are we to interpret these findings which appear, at first glance, to challenge the results of the prospective randomized ISCHEMIA trial? First, we congratulate the authors for publishing this important, beautifully presented study. However, it is important to remind ourselves of the inherent limitations of non-randomized, observational evidence, specifically the significant risk of unmeasured confounding. In addition to the highly selected nature of the study population, whereby all underwent SPECT followed by early invasive or non-invasive coronary angiography, the decision to perform ICA and revascularization in a non-randomized fashion typically selects for healthier patients as compared to similar patients managed conservatively. Such invasively managed patients who underwent revascularization were likely less frail, more compliant with medications, at lower risk for bleeding, and potentially different from those who underwent medical management due to a host of other unmeasured variables that can only be mitigated through the power of randomization. Hence, one should not conclude that ischemia-guided revascularization improves outcomes in patient with chronic CAD, but perhaps that the degree of ischemia, along with many other variables such as anatomic severity, suitability for revascularization, symptom severity, among others, may be a part of patient-level coronary revascularization decision-making.

There are several additional limitations to this analysis that warrant discussion. First, this study should not be viewed as a comparison of anatomy versus physiology for assessing clinical outcomes, due to the significant risk of confounding and strategy used for patient selection. Nor was it a comparison of non-invasive testing strategies, such as initial CCTA (or ICA) versus SPECT, as all patients underwent SPECT-first followed by anatomic testing. Patients who underwent initial CCTA and those who underwent SPECT without subsequent angiography were not studied. The assessment of anatomic CAD in this analysis was also relatively simplistic as it used a hierarchical CAD score based solely on lumen stenosis ≥50%. The assessment of CAD in this study did not include the established independent and incremental prognostic value of non-obstructive CAD (<50%), coronary plaque burden, or high-risk plaque on CCTA.¹⁰ Indeed, variables such as coronary artery calcium score, coronary segmental involvement score, plaque volumes (including non-calcified and low-attenuation plaque), and the presence and extent of high-risk plaque likely would have enhanced the prognostic value of anatomic CAD tests, particularly CCTA. We were also not presented data on inter- and intra-reader agreement for CADPI scoring in patients who underwent CCTA and ICA. As mentioned, post-SPECT adherence to guideline-directed medical treatment, risk factor control, and measures of compliance with a healthy lifestyle were not ascertained or controlled for in event prediction models. Finally, as the study population spanned more than a decade, guideline directed medical treatments and invasive revascularization techniques, to include the use of invasive fractional flow reserve and intravascular imaging for revascularization decision-making, likely varied during the study period and were not included in this analysis, making the results less than contemporary.

So, what does the future hold for utilizing ischemia in patient decision making? There is certainly much more to learn and ischemia information will likely continue to play a significant role in patient management. Future randomized, prospective comparative effectiveness studies assessing the role of ischemia in guiding patient management, including decisions about revascularization, and outcomes are still needed, despite the informative results from the ISCHEMIA trial. Such studies should include more contemporary, lower radiation tests for ischemia, such a stress/rest cardiac positron emission tomography and stress cardiac MRI, tests which provide more accurate detection of CAD and additional measures of coronary flow and flow reserve as compared to SPECT MPI. In the post-ISCHEMIA trial world, change in routine clinical practice to utilizing these more contemporary tests in patients with suspected CAD is perhaps overdue.