Development and validation of ischemia risk scores

Area of Science:

• Cardiology
• Medical Diagnostics
• Predictive Analytics

Background:

• Myocardial perfusion imaging (MPI) is crucial for diagnosing ischemia.
• Physician decisions on test selection are hindered by a lack of dedicated risk scores.
• Novel risk scores are needed to aid in selecting appropriate diagnostic strategies.

Purpose of the Study:

• To derive and validate novel risk scores for predicting the probability of ischemia on MPI.
• To support physician decision-making in selecting diagnostic tests.
• To improve the accuracy of ischemia prediction.

Main Methods:

• Two novel ischemia risk scores were derived from 15,186 patients and validated in 2,995 patients.
• Logistic regression identified associations with ischemia to create point-based and calculated scores.
• Predictive performance was assessed using AUC and compared against existing models.

Main Results:

• The calculated ischemia risk score demonstrated higher predictive accuracy (AUC) than the point-based score during derivation and validation.
• The calculated score outperformed the clinical CAD model in external validation.
• Both scores showed good calibration in independent patient cohorts.

Conclusions:

• Two novel risk scores for predicting ischemia on MPI were successfully developed.
• These scores exhibit high accuracy in both derivation and external validation.
• The developed scores can assist physicians in optimizing diagnostic testing strategies.