Integrating FFR_CT into Routine Clinical Practice: A Solid PLATFORM or Slippery Slope?

Coronary computed tomography angiography (CTA) is being used increasingly in patient care despite the highest standard ever set for adoption of a noninvasive imaging test for diagnosing coronary artery disease (CAD). Recent analyses from the SCOT-HEART (Scottish Computed Tomography of the HEART) trial, which randomized patients to standard of care with or without coronary CTA (1), demonstrated that following a coronary CTA, invasive angiograms are significantly more likely to show obstructive lesions, and appropriate therapy to be initiated, associated with halving of myocardial infarction occurrence (2). Combined with the recent PROMISE (Prospective Multicenter Imaging Study for Evaluation of Chest Pain) trial (3), these data supported this expanding role for coronary CTA in CAD management (4). Furthermore, coronary CTA may be used to monitor coronary plaque progression (5,6) and has an increasing role in patients with acute chest pain (7) presenting to the emergency department (8). These studies and others have fueled the proposal that coronary CTA is a highly competitive or superior modality as a gatekeeper to invasive angiography (9,10).

Complementing this role of coronary CTA, considerable effort has been devoted to assess the physiologic significance of stenosis burden by different CT-based approaches, including transluminal attenuation gradient (TAG) (11,12), contrast density difference (13), CT vasodilator-induced stress myocardial perfusion imaging (14,15), and fractional flow reserve CT (FFR_CT), the most rigorously studied parameter of functional CAD assessment (16–22). Moreover, FFR_CT has demonstrated additive value in what remains an Achilles’ heel of coronary CTA: decreased accuracy in the setting of significant calcification (23–25).

In this issue of the Journal (16), Douglas et al. present the 1-year follow-up of the previously reported 90-day quality of life (QOL) and economic impact (17) of the PLATFORM (Prospective Longitudinal Trial of FFRCT Outcome and Resource Impacts) trial (18). In an elegant editorial that accompanied the 90-day trial results (19), Hulten and Di Carli reviewed the history and evolution of CAD testing and changing clinical trends as well as pivotal FFR_CT trials (20–22). They concluded that randomized trials comparing FFR_CT, coronary CTA, and stress testing were needed to further evaluate changes in cost and outcomes with FFR_CT (19). The authors also noted that translating FFR_CT results from a clinical trial population to routine clinical practice patients might prove challenging.

In their 1-year follow-up study, Douglas and colleagues concluded that there were no meaningful differences in QOL in the planned noninvasive and invasive arms other than by 1 isolated metric; however, addition of FFR_CT led to increased cost in the planned noninvasive arm but a significant decrease in the planned invasive arm. In the planned invasive arm, regardless of the angina type experienced, an FFR_CT-guided strategy was beneficial, with a substantial 60% reduction in invasive coronary angiography (ICA), the main driver of economic impact. Additionally, FFR_CT led to a significant 61% decrease in the finding of obstructive CAD at the time of ICA. The study was underpowered to test whether this decrease in ICA led to significant differences in ‘hard’ cardiovascular events such as fatal myocardial infarction, acute coronary syndromes, or urgent revascularizations.

The following caveats should be noted: the study was conducted in European centers and reflect local practice patterns, which cannot be assumed to be identical to American centers. Moreover, as PLATFORM was not a randomized trial, the decision to medically manage or directly send patients to the cardiac catheterization laboratory must be examined more closely. In the planned invasive cohort, many patients would not likely be chosen to directly proceed to ICA in many cardiology centers under current guidelines. Conversely, American patterns of noninvasive imaging tests have had disquieting percentages of normal studies, raising questions about the appropriateness of our patient selection process (26).

The decision to proceed with ICA in PLATFORM was unclear. Only ~50% of the usual care and FFR_CT-guided arms had noninvasive testing – the results of which were not reported – before the decision to proceed to ICA, an approach at odds with current guidelines (27). As an extreme example, in some cardiology environments patients with angina are offered ICA only after failed medical therapy and/or abnormal noninvasive testing demonstrating large areas of reversible ischemia on myocardial perfusion imaging (28). Certainly there was room for additional testing using current routine modalities prior to and/or other than FFR_CT in the planned invasive arm of the PLATFORM study.

Another significant drawback to the routine use of FFR_CT is the lack of advantage in the noninvasive arm. Costs were higher in the planned noninvasive cohort, even when FFR_CT was assigned a cost weight equal to coronary CTA, emphasizing that FFR_CT should not be universally proposed to patients during CAD work-up but considered in patients designated to ICA. Importantly, 297 patients in both study arms underwent coronary CTA. Of those, 201 studies showed stenoses ≥30% with subsequent submission for FFR_CT analyses. Of those, 177 were actually analyzed and FFR_CT values provided to clinicians. In other words, FFR_CT values were available in only 58% of noninvasive and 60% of invasive arm cases, respectively, for clinical decision making. More than likely, a proportion of the decisions not to proceed with ICA might have been done on the basis of coronary CTA alone. In fact, observational studies suggest coronary CTA alone can dramatically reduce by almost half the incidence of ICA in controlled clinical environments (9).

The major advantage of FFR_CT above other methods of physiologic assessment resides in its direct reference to invasive FFR. An invasive FFR-guided method of artery-specific revascularization of stable CAD patients was addressed in the FAME (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation) trials (29,30) and led American (31) and European (32) guidelines to recommend routine use of FFR with a cutoff of 0.80 to guide the decision to medically manage or revascularize patients. Furthermore, FFR_CT has now undergone testing and validation against invasive FFR in multiple clinical trials (20–22), leading to its approval by the U.S. Food and Drug Administration (FDA). In the PLATFORM cases where invasive FFR was performed (50 vessels in 29 patients), FFR_CT had an overall accuracy of 84%. An additional advantage of FFR_CT is that it can be added onto coronary CTA for artery-specific physiological measurement when deemed necessary, with no need for repeat acquisition and, hence, no additional radiation or contrast administration.

As perceived by Hulten and Di Carli (19), introducing FFR_CT into the real world was challenging. We performed the first 100 U.S. studies after FDA release and likely have one of the largest volumes in clinical practice currently (25,33). In our environment, when not enrolled in a clinical trial, patients had to pay for the test out of pocket. This required a detailed, understandable explanation with supplemental written and video material. It also meant many patients could not get the test. Hospital and office administrators needed to be convinced of the value proposition, especially with no insurance reimbursement. In the early stages, sole reliance on FFR_CT was not routinely accepted; therefore, multimodality noninvasive testing was requested by other health care team members, patients, or both. There was concern about the ‘black box’ nature of the measurement, given the unclear contributions of the CT angiogram versus the segmentation and computational fluid dynamics modeling leading to the final FFR_CT value, the skill of the personnel performing the interpretation/segmentation/modeling, the effects of noncalcified plaque on the modeling, concern about a sole commercial vendor and concern about nonconcordance with clinical features and other noninvasive testing. FFR_CT provided a convenient ‘second opinion’ on the coronary CTA interpretation with additive functional assessment, which was confirming in some cases, challenging in others. Clinical care had to adjust to the changing paradigm of FFR_CT with values provided not only at the distal tip of the arteries, but also the capability of measurements at any level of the artery and across a narrowing. Despite these limitations, we found that FFR_CT implemented in a routine clinical practice predicted artery-specific revascularization behavior better than coronary CTA alone (33), particularly in cases with heavy calcification (25).

For FFR_CT to gain broader use, outcomes trials similar to the FAME trials should be performed to unequivocally demonstrate the central role this technology can play in managing patients with CAD; currently, FFR_CT is only being used as a ‘surrogate’ to invasive FFR. We should not assume this surrogate measure has the same outcome associations as invasive FFR until these measures meet the same standards in similarly designed rigorous trials. Nonetheless, the enhanced role FFR_CT provided in patients with planned ICA in PLATFORM is noteworthy.

This trial further expanded the “PLATFORM” of coronary CTA by providing an additional tool to enhance our ability to refer the right patients for ICA. Obviously, it would have been more powerful if PLATFORM was a randomized trial, if the workup of the population agreed with current guidelines, and if FFR_CT results were compared to coronary CTA alone and provocative stress testing. This is particularly the case given previous single-center observational studies as well as registry data proposed CTA – without FFR_CT – as a gatekeeper to ICA (9,10). However, while the exact role of FFR_CT in managing patients with CAD continues to be explored, either complementary to or in replacement of other noninvasive strategies, the present effort by Douglas et al. analyzing the 1-year follow-up of the PLATFORM trial provides 1 more step in our understanding of the growing role of coronary CTA in patient care.