The coming age of coronary atherosclerosis-guided cardiovascular risk prediction and management

Coronary atherosclerosis is a slow progressing disease that cases persist completely asymptomatic until it becomes extensive and diffuse but for almost half of the patients its initial manifestation is sudden death [1]. Early work on the understanding of risk factors led to the development of cardiovascular risk scores such as the Framingham risk score in the 1960s which let to the current strategy of pharmacologic cardiovascular risk reduction focused on individual risk factor control targeted to estimated clinic. This persists as the current guideline recommended approach for blood pressure and lipid level management since the publication of the ATP III guidelines [2].

Despite the ability of cardiovascular risk scores to identify a small group of very high-risk individuals, it is known for decades that most events happen in individuals with low to intermediate cardiovascular risk as their overall population is substantially larger, a phenomenon known as prevention paradox [3]. In an attempt to address this issue, several strategies have been developed such as: 1. The development of additional risk markers to identify higher risk individuals (ankle brachial index, carotid intima media thickness, hsCRP, family history of premature CAD, etc.); 2. Improvements in risk prediction models (SCORE 2, ASCVD risk calculator, PREVENT); 3. Lowering the threshold for pharmacologic treatment, particularly since the 2013 American guidelines. The collective results of such strategies led to predictable response to the prevention paradox: a substantial increase in the proportion of the population became eligible for treatment. Since risk factors are very prevalent in the western population, this strategy led to recommendations for pharmacologic treatment in a very large, substantially lower risk population. Among other issues, this brought up the concerns of overmedicalization of healthy individuals [4].

Since then, a slow paradigm shift in cardiovascular risk prevention has been brewing, the idea that cardiovascular risk management should be guided by the actual presence and burden of disease, i. e. coronary atherosclerosis. The initial shift in this paradigm was a natural response to the overtreatment concerns mentioned above. Consistent work from hundreds of studies has demonstrated the exceptionally low yield of treating individuals with no detectable coronary artery calcification (CAC) as those individuals have negligible risk even on longer term follow-up [5]. Community based samples have demonstrated that the population of individuals eligible for statin could be substantially reduced if coronary artery calcium score were used.

Aside from identifying lower risk individuals, CAC has also been considered as a tool to identify higher risk individuals who may be considered for newer risk reduction therapies. This led to the concept of multilevel stage definition of atherosclerotic burden to define management based on the extent of coronary artery atherosclerosis measured by the proxy of coronary artery calcification (CAC). In current guidelines, CAC still is the only measurement of coronary atherosclerosis recommended by guidelines for cardiovascular risk prevention [6].

The development of contrast enhanced coronary computed tomography now allows for a more detailed assessment of the presence, extent and severity of coronary atherosclerosis. These images have allowed advanced secondary post processing analysis to quantify atherosclerotic plaque burden and plaque composition; measurement of pericoronary fat attenuation and modeling of possible coronary flow reduction. While the visual impact of such high-quality images has attracted many enthusiasts, data on how to use coronary computed tomography for risk prediction in asymptomatic individuals still is not as robust as the data supporting the predictive ability of CAC.

In the present edition of the AJPC, Parsa et al., provide insights on the potential impact of coronary computed tomography for management of cardiovascular risk [7]. Using mathematical models, the authors have demonstrated that quantification of total coronary plaque volume (TPV) can be used to stratify cardiovascular risk. As previously demonstrated with calcium score and other measurements of atherosclerotic burden by coronary computed tomography, TPV has very good correlation with cardiovascular risk. As expected from the modeling strategy, more aggressive management of higher risk individuals led to a modeled reduction in cardiovascular events. The conceptual models and results make sense and are in line with prior evidence in the field, presence and burden of coronary atherosclerosis provides incremental value for cardiovascular risk prediction and management.

A more in-depth analysis of the model strategy, however, provides aspects for consideration. While the authors provide evidence for the choice of modeled targets and adherence on both arms, other evidence suggests that the modeled difference in adherence between the two arms might be too optimistic. In the SCOT-HEART study data the difference in the use of preventative medications between the CT and usual care arm was <10 %. A second concern to translate their findings into practice is that, as in prior coronary CT studies, the data is derived from a population of symptomatic individuals were coronary CT is clinically indicated rather than the general population of candidates for primary prevention.

One final particularly key point for consideration is the recurring use of advanced post processing technologies in cardiac CT, such as the objective quantification of plaque, with the assumption that the usual clinical coronary CT information would not otherwise be available. While the authors should be commended for providing evidence that coronary atherosclerosis stages may be beneficial for incremental risk assessment, the larger question that persists is how should we measure coronary atherosclerosis to inform practice and whether objective plaque quantification provides any meaningful marginal information to predict cardiovascular risk and guide management.

This is particularly important as an extensive number of parameters can be derived to quantify plaque burden, such as CAC, segment involvement score, segment stenosis score, plaque composition and characteristics assessed in clinical coronary reads all the way to advance post processing derived features such as plaque composition, TPV, presence and degree of ischemia or pericoronary fat attenuation. The cardiovascular imaging community now needs to evaluate how can this best be translated into clinical risk prediction and management guidelines. This will require the development of clinical scores and algorithms of management incorporating plaque derived parameters. Importantly, no matter how exciting and attractive newer postprocessing techniques may be, they are secondary analysis of clinically interpretated data with substantial incremental cost. As such, it is paramount that we meaningfully assess their clinical impact to justify the added cost and complexity. Simply demonstrating that objective plaque quantification correlate with outcomes should not suffice, this has already been well established by semi quantitative parameters already reported in current clinical coronary computed tomography reports or calcium score.

Still, communally, the literature in the field is building consistent evidence that rather than relying on clinical scores purely derived from clinical parameters to predict the likelihood of disease, the actual measurement of coronary atherosclerosis, particularly if contextualized by other clinical parameters such as age and other risk factors, will become the mainstem for cardiovascular risk management. It is on us to find how best to make direct coronary atherosclerosis assessment by coronary computed tomography simple, intuitive and well supported by evidence to become this next mainstem in cardiovascular prevention.

CRediT authorship contribution statement

Marcio Sommer Bittencourt: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Resources, Methodology, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Marcio Sommer Bittencourt reports was provided by University of Pittsburgh. Marcio Sommer Bittencourt reports a relationship with Cleerly Inc that includes: speaking and lecture fees. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.