Author: Max Chen, Sean Marquardt, Ben Yang, Kai Yang 👨🔬
Affiliation: Cary Academy, Massachusetts General Hospital, Winchester High School 🌍
Purpose: To analyze the impact of scanner model variation on the effective dose conversion factor (“k-factor”), which is most commonly used for CT effective dose calculation.
Methods: The standard examinations function of CT-Expo v2.8 was used to calculate the k-factor of each unique combination of Gender, Scanner Model, kVp, and Exam Type. Four typical adult exam types: Routine Chest, Routine Abdomen and Pelvis, Whole Trunk, and Brain (lenses included), and all the scanner models listed in CT-Expo standard examinations function were included. Using the default scan parameters for each exam with a fixed CTDIvol of 10mGy, CT-Expo provided organ dose and effective dose values. The k-factor was calculated using the ratio between effective dose and Dose Length Product with a unit of mSv/mGycm. The data was analyzed with t-tests to compare the difference between gender and exam type. Multi-variable linear regression was performed to investigate the relative impact of each variable.
Results: With 868 scanner models, the overall mean (standard deviation) of k-factor was: 0.0157 (0.0019) for Routine Abdomen & Pelvis, 0.0021 (0.0011) for Brain (lenses included), 0.0172 (0.0040) for Routine Chest, 0.0159 (0.0024) for Whole Trunk, respectively. There was no difference in k-factor for different kVp from CT-Expo v2.8. There was a statistically significant difference between gender and exam type from the t-tests (p<0.01). From the multivariable linear regression model, there was a significant impact from gender and exam type on the k-factor, while there was no significant impact from scanner model, with a coefficient of -0.0045 for exam type, -0.0034 for gender, and 2.17E-07 for scanner model, respectively.
Conclusion: Factors including gender, exam type, scanner model, and kVp were analyzed according to their impact on the k-factor using the CT-Expo software. Varying scanner models had the least impact on k-factor compared to exam type and gender.