Overview
The 2024 European Society of Cardiology (ESC) guidelines recommend a more structured way to estimate the likelihood of obstructive coronary artery disease (CAD) in people with suspected chronic coronary syndromes. A major feature of this approach is the risk factor-weighted clinical likelihood model, or RF-CL model. It is designed to improve the initial pre-test probability estimate, helping clinicians decide who should undergo further testing, who may need invasive coronary angiography, and who may safely avoid unnecessary investigations.
This Great Debate article examines both sides of the argument: why the RF-CL model is useful in real-world practice, and why some experts remain cautious about its limitations. The discussion matters because estimating disease likelihood is the first step in the modern diagnostic pathway for stable chest pain and related symptoms.
What the RF-CL model is meant to do
The RF-CL model estimates the probability that a patient has obstructive CAD before advanced testing is performed. It builds on traditional pre-test probability approaches, but adds more information to improve accuracy. The model combines:
Age
Sex
Nature of symptoms
Five clinical risk factors
Coronary calcium data, if available
The idea is not simply to say whether coronary disease is present. Rather, it tries to estimate how likely it is that symptoms are caused by flow-limiting, obstructive atherosclerotic disease in the coronary arteries.
This is important because not all chest pain is due to blocked arteries. Some patients have non-obstructive coronary disease, microvascular dysfunction, vasospasm, or even non-cardiac causes such as musculoskeletal pain, gastrointestinal disease, or anxiety-related symptoms.
Why this model was developed
Older pre-test probability models tended to rely mainly on age, sex, and symptom type. While useful, those models often overestimated disease likelihood in many modern populations, especially where preventive therapy, smoking rates, and diagnostic patterns have changed.
As a result, many patients were sent for testing despite having a very low probability of obstructive CAD. That led to unnecessary imaging, invasive procedures, anxiety, cost, and delay in identifying the true cause of symptoms.
The RF-CL model aims to correct this by incorporating clinical risk factors and, when available, coronary artery calcium information. This makes the estimate more individualized and better aligned with contemporary patient populations.
How the model is used in practice
In the ESC pathway, the RF-CL estimate is the first formal step after the medical history is taken. Clinicians assess the patient’s symptoms, age, sex, and risk profile, then use the model to determine whether the likelihood of obstructive CAD is low, intermediate, or high.
In general:
Very low likelihood patients may not need immediate cardiac testing.
Intermediate likelihood patients may benefit from non-invasive testing such as coronary CT angiography, stress imaging, or functional assessment.
Very high likelihood patients, particularly those with an estimated probability above 85%, may be considered for invasive coronary angiography.
The model also supports “de-risking,” meaning it can identify patients whose probability of obstructive CAD is sufficiently low that further tests can reasonably be deferred, provided there are no red flags or alternative urgent concerns.
Strengths of the RF-CL model
Supporters of the RF-CL model point to several important advantages.
First, it improves calibration. In other words, the estimated probabilities are closer to what is actually observed in real patients than many older models.
Second, it is practical. The score is relatively quick to use, does not require complex calculations at the bedside, and has been externally validated in multiple contemporary cohorts from Europe, North America, and Asia.
Third, it helps reduce unnecessary testing. If a patient is very unlikely to have obstructive CAD, there may be no value in exposing them to radiation, contrast agents, procedural risk, or the cascade of follow-up tests that can follow a false-positive result.
Fourth, it supports better triage. In busy clinics and emergency or outpatient chest pain pathways, prioritizing the right patients for the right tests can improve efficiency and reduce waiting times.
Why some experts remain skeptical
Despite these advantages, the RF-CL model has sparked debate.
One concern is symptom interpretation. Real-world symptoms are often vague, mixed, or difficult to categorize. Chest discomfort may be typical in some respects but atypical in others, and the “specificity” of symptoms can be hard to judge reliably between clinicians.
Another issue is risk factor definition and weighting. The model relies on how certain risk factors are defined and measured. If blood pressure, diabetes, smoking history, lipid abnormalities, or family history are documented inconsistently, the score may not be as robust as it appears.
There is also concern about the so-called enrichment factors that can increase disease likelihood. These may include coronary calcium, ECG findings, and other clinical clues, but their use can be subjective or variable across settings.
Some experts also point out that the RF-CL model becomes only semi-quantitative or qualitative above 45%. That means the precise numerical estimate becomes less exact in higher-risk ranges, which may reduce its usefulness when decision-making is most difficult.
Finally, the very high-likelihood threshold of more than 85% for recommending invasive angiography raises questions about how often such patients are truly encountered and how that estimate is reached in daily practice.
Why obstructive CAD and ischemia are not the same thing
A central concept in this debate is that the RF-CL model is designed to estimate the likelihood of obstructive CAD, not ischemia.
This distinction matters. Obstructive CAD refers to anatomically significant narrowing in the coronary arteries. Ischemia refers to insufficient blood flow to the heart muscle, usually during stress or exertion.
A patient may have ischemic symptoms without visible obstructive coronary lesions, for example because of microvascular angina, vasospastic angina, or diffuse non-obstructive atherosclerosis. Conversely, some people with obstructive lesions may not demonstrate ischemia on testing, depending on the lesion’s functional significance and the method used.
Therefore, a good RF-CL score does not replace clinical judgment or functional assessment when ischemia is suspected. It is one part of a broader diagnostic strategy.
How coronary calcium can help
When available, coronary artery calcium information can improve the model’s estimate. Coronary calcium is a marker of atherosclerotic burden and often correlates with the likelihood of coronary disease.
A high calcium score can increase suspicion for CAD, while a calcium score of zero may lower the likelihood of obstructive disease in some patients, especially when symptoms are not strongly suggestive of angina. However, a zero score does not absolutely exclude CAD, particularly in younger individuals, those with non-calcified plaque, or those with high-risk clinical features.
The RF-CL model uses this information as an additional enrichment factor, not as a standalone diagnosis.
Clinical impact: reducing overtesting without missing disease
One of the main goals of the RF-CL model is to reduce overtesting while preserving safety.
In contemporary practice, many patients with stable chest symptoms have a low probability of obstructive CAD. If these individuals are identified early, they can often be reassured, treated for other causes of symptoms, and managed with lifestyle advice and risk factor control rather than extensive cardiac testing.
At the same time, the model is intended to avoid undertesting in those with higher likelihood. Patients with more convincing symptoms or multiple risk factors should not be falsely reassured by a low-value estimate. The challenge is achieving the right balance.
This is where careful history-taking remains essential. No model can substitute for a skilled clinician who can recognize unstable features, exercise tolerance changes, associated symptoms, or clues suggesting another urgent diagnosis.
Practical limitations in everyday care
Even a well-designed model may be difficult to apply perfectly in routine practice.
Time pressure, incomplete records, and variability in documentation can all affect accuracy. Different health systems also vary in the availability of coronary calcium scoring, coronary CT angiography, or stress imaging. In some settings, clinicians may not have easy access to the data needed to fully apply the score.
In addition, patient populations differ. A model validated in one region may perform differently in another depending on baseline CAD prevalence, referral patterns, and preventive therapy use.
For these reasons, the RF-CL model should be viewed as a decision support tool rather than an automatic rule. It assists clinical reasoning; it does not replace it.
What the debate ultimately shows
The great strength of the RF-CL model is that it represents a more modern and individualized way to estimate the probability of obstructive CAD. It is more nuanced than older approaches and can safely reduce unnecessary testing in a meaningful number of symptomatic patients.
The main criticism is not that the model is useless, but that its real-world value depends on how accurately clinicians can define symptoms, capture risk factors, and interpret the output in context. A probability estimate is only as good as the data entered into it and the clinical judgment used to interpret it.
In other words, the RF-CL model is a helpful improvement, but not a perfect solution.
Bottom line
For patients with suspected chronic coronary syndromes, the risk factor-weighted clinical likelihood model is an important step forward in estimating the initial probability of obstructive CAD. It improves risk stratification, supports smarter testing, and may help avoid unnecessary procedures.
At the same time, it should be used with awareness of its limitations, especially around symptom classification, risk factor definitions, and the fact that it estimates obstructive disease rather than ischemia itself.
The debate is therefore not whether the RF-CL model has value, but how best to use it in practice. When applied thoughtfully, it can improve care. When used uncritically, it may oversimplify a complex clinical problem.
