Core principle that serves as the foundation of our product:
Everyone is unique: multifactorial nature of CAD
The multifactorial nature of Coronary Artery Disease means that CAD usually does not come from one single cause. It develops because many factors act together over time, including the combined effect of cholesterol, blood pressure, diabetes, smoking, inflammation, genetics, and lifestyle, age/sex, rather than by one single factor.
Thus, CAD is not only about a big blockage in one artery. A person can have: plaque buildup in large coronary arteries, abnormal vessel wall biology, endothelial dysfunction, thrombosis risk and top of that the microvascular dysfunction.
BIOME Science is building the next generation of AI diagnostic software bringing precision medicine in cardiac imaging to resolve the grey zone driven by the multifactorial nature of CAD. Thus, precision medicine becomes accessible before the catheterization lab, leading to less unnecessary invasive procedures and more personalized care. Moreover, the software can provide a follow-up monitoring of intermediate lesions over time with earlier and more accurate diagnostics of CAD.
Clinical decision making
Why “multifactorial” matters clinically ? First, the anatomy alone is insufficient. Two patients with similar angiographic stenosis severity may have different ischemic burden and outcomes because CAD expression depends on: plaque characteristics, diffuse vs focal disease, vessel remodeling, microvascular function, collateralization, myocardial demand, blood rheology. Second, the Ischemia is multifactorial too. Myocardial ischemia in CAD is not determined only by epicardial lumen loss. It can reflect the combined effects of: focal stenosis, diffuse atherosclerotic burden, endothelial dysfunction, vasospasm, microvascular dysfunction, elevated microvascular resistance, altered hyperemic response This is especially relevant in patients with INOCA, microvascular angina, or mixed epicardial–microvascular disease. Moreover, any acute events are not perfectly predicted by stenosis severity. A non-severe lesion may still be high risk if it is biologically vulnerable.
Implications for modeling and precision medicine in view of multifactorial nature of CAD:
Impact of Individual Patient Factors on Decision to Place a Stent
Because CAD is multifactorial, a robust disease model should not rely on a single axis such as percent stenosis. It should ideally integrate: anatomical data, coronary physiology, plaque phenotype, microvascular state, clinical risk factors, biomarkers, patient-specific modifiers. That is exactly why modern CAD assessment increasingly moves from a purely anatomical view toward multimodal, patient-specific risk and physiology modeling. Let’s examine scenarios that exhibit a strong multifactor influence:
Male, 57 years old
BMI* = 34, Patient with type 2 diabetes and hypertension, presenting with anginal symptoms.
Patient has only 45% LAD stenosis with a low CAD RADS score of 2 - low risk.
Typical AI statistical models trained on CT data only will likely predict NO ISCHEMIA. However, Coronary Flow software & invasive FFR demonstrate presence of ischemia with FFR = 0,78.
Conclusion: Although, present intermediate lesion is of relatively low risk, this patient experiences complex multifactorial disease patterns, leading to a decision that he would likely benefit from stenting.
Male, 64 years old
BMI* = 24. Patient with no significant comorbidities apart from tobacco use, presenting with a severe calcified stenosis of 60% in RCA with CAD-RADS score of 3 - high risk.
Typical AI models will most likely predict ISCHEMIA in this scenario. However, Coronary Flow software & invasive FFR demonstrate that there is NO ischemia with FFR = 0,91.
Conclusion: Although, present lesion is of high risk in terms of volume and diameter, this patient experiences complex multifactorial patterns in the presence of a rigid calcified plaque, leading to a decision that he has NO need to be stented at the moment. However, a follow-up monitoring of this intermediate lesion is recommended.
