Vascular Remodeling Radiomic Risk Model (FRP) Calculator

Clinical calculation support · Verify results before clinical use

Input parameters

Enter the output from a validated CCTA radiomic workflow and record HRP status.

0.63FRP cutoff
Enter the FRP probability (0.00–1.00) calculated by your validated imaging pipeline. This page does not calculate FRP from CCTA images.
Use the HRP definition and reading protocol applied in the source study/local CCTA reporting workflow. Common features include low-attenuation plaque, positive remodeling, napkin-ring sign, and spotty calcification.
Context only; FAI is not used in the FRP hazard calculations on this page.
Context only; CAC is not used in the calculations below because the published HR estimates are not a patient-specific absolute-risk model.
Reset calculator

Clinical context and limitations

The fat radiomic profile (FRP) is a CCTA-derived radiomic signature intended to capture adverse structural remodeling in pericoronary adipose tissue. In the original work, a machine-learning model using radiomic features was developed and evaluated against tissue biology and cardiovascular outcomes.

This calculator accepts an FRP value already produced by a suitable pipeline. It does not reproduce the image segmentation, radiomic feature extraction, model training, or external validation steps needed to generate a clinically valid FRP.

Published estimates used here

  • FRP cut-off: 0.63, as reported in the SCOT-HEART analysis.
  • Continuous association: adjusted HR 1.12 per 0.01 FRP increment (95% CI 1.08–1.15).
  • Grouped association: adjusted HR 10.84 for FRP ≥0.63 versus <0.63 (95% CI 5.06–23.22).
  • FRP−/HRP+ versus FRP−/HRP−: HR 5.97 (95% CI 1.24–28.69); FRP+/HRP−: HR 32.44 (95% CI 7.00–150.38); FRP+/HRP+: HR 43.33 (95% CI 9.14–205.48).

Use limitations

These are observational, cohort-specific associations with substantial uncertainty in some strata. They should not be interpreted as multipliers of an individual's absolute probability of an event, used alone to direct treatment, or extrapolated to other populations without validation. FAI and CAC are displayed as contextual inputs only.

References

Oikonomou EK, Williams MC, Kotanidis CP, et al. A novel machine learning-derived radiotranscriptomic signature of perivascular fat improves cardiac risk prediction using coronary CT angiography. European Heart Journal. 2019;40(43):3529–3543. DOI: 10.1093/eurheartj/ehz592.
Clinical disclaimer: This tool is an educational and research decision-support aid. It is not a substitute for validated image analysis, local validation, guideline-directed prevention, or clinician judgment. Do not use the displayed hazard ratios as absolute-risk estimates or as the sole basis for clinical decisions.