Inter-Observer Reliability in Myocardial Blood Flow Measurement Using Rb-82-PET/CT

Background/Objectives: Myocardial perfusion PET/CT, e.g., using Rb-82, enables the measurement of absolute myocardial blood flow (MBF) and, hence, the blood flow reserve, which plays a crucial role in the diagnosis of coronary artery disease. However, the values might depend on the observer's analysis and, thus, depend on the observer's experience. We investigated the inter-observer reliability of MBF measurements in Rb-82-PET/CT. Methods: Thirty consecutive patients who underwent Rb-82-PET/CT for suspected myocardial ischemia were analyzed. Stress conditions were simulated using 400 µg Regadenoson. MBF was measured using Corridor 4DM software. Myocardial flow reserve (MFR) is the ratio of absolute flow measurements under stress and rest conditions. Four different physicians with different levels of experience (two senior physicians and two residents) measured the MBF. The experience of the two residents was one year of experience and no previous experience, respectively. Intraclass correlation coefficients (ICC) were calculated for MBF and MFR in the three main vascular territories as well as in the whole left ventricle. Additionally, the left ventricular ejection fraction (EF) was determined. Results: The evaluation of MBF showed an excellent correlation between all observers, indicated by a low inter-observer variability. Comparing the two different levels of experience, senior physicians showed a slightly higher accordance in MBF assessment (ICC ranging from 0.9 to 0.98) than the residents (ICC ranging from 0.86 to 0.94). The MFR calculations showed a slightly lower reliability due to the accumulation of variability in rest and stress measurements (ICC between 0.78 and 0.81), still indicating a good reliability. The EF calculations also showed a good reliability among the different observers (ICC 0.83). Conclusions: Measurement of MBF, MFR and EF in Rb-82-PET/CT is highly robust with a very low inter-observer variability. A slightly higher inter-observer reliability was found for more experienced physicians.