Room: Room 205
Purpose: Automated Tube Current Modulation (ATCM) systems are implemented in modern CT scanners adjusting X-ray flux for patient size to save patient dose and achieve uniform image quality. Different vendors developed different ATCM strategies that offer different relationships between dose, noise, and patient size. Therefore, there is a need to assess image quality features as a function of patient habitus. This workâ€™s goal is to show that it is possible to predict noise for real patient images using Mercury Phantom (MP) and, thus, to inform protocol design.
Methods: This IRB-exempt study included 1519 examinations performed by two scanners from two vendors and two clinical protocols (abdominopelvic with and chest without contrast). The adult MP design includes four different sizes (18.5cm, 23cm, 30cm, 37cm) representing different patient size ranges. Noise magnitude (average pixel standard deviation in uniform image) was automatically estimated in patients and in MP using a previously validated algorithm. Noise values were calibrated to match patient noise values for a 30cm size. A quadratic fit of noise as a function of size was applied to patients and MP data. Lastly, the MP quadratic equation was used to fit the patient data. In each patient distribution fit, the normalized root mean square error (nRMSE) values were calculated as a metric that indicates how well MP model can predict the noise in clinical operations as a function of size.
Results: nRMSE in patient fit ranges between 8.1% and 11.8%. nRMSE using shifted MP equation to fit patient data ranges between 10.2% and 16.4%. The average difference in nRMSE is 3.0%.
Conclusion: MP can predict image noise in clinical patient populations. By assessing image quality in a phantom with multiple sizes, protocol parameters can be designed and optimized in a highly constrained setup to predict clinical scanner and ATCM system performance.