A Radiation Dose Monitoring System for the PET/CT Clinic

1914 Objectives For the purposes of quality assurance and regulatory accountability it is important to track radiation dose metrics. We designed an automated system that captures dose data from the CT scanner radiation dose report and the radiopharmaceutical administration system. The data are processed and archived as part of the permanent medical record in the nuclear medicine information system (NMIS). They are employed in the reporting system and are available for ongoing analytics and periodic review. Methods Our single PET/CT (Siemens Biograph mCT-64) scanner operates in an academic medical center and performs about 2000 F18-FDG PET/CT studies per year. Radiopharmaceutical administration is performed using an automated infusion system (Bayer/MEDRAD, Intego) setup to administer 4 MBq/kg. Greater than 95% of studies performed in our clinic employ a standard oncology protocol which involves a low-dose CT protocol (140 kV, tube current modulation with quality reference set to 50 mAs), for attenuation correction and anatomical localization, followed by PET acquisition. The radiation dose monitoring system combines data culled from a CT dose data messaging system and the NMIS. The CT dose system was locally developed using a combination of an open source DICOM server (Conquest DICOM V1.4.), database (Microsoft SQL Server), open source data conversion software (DCMTK), and locally-developed software (Powershell, Python 3, and VBscript). The NMIS (also developed in-house) provides access to radiopharmaceutical administration data, and patient height and weight. For each patient study, the software produces a size-specific dose estimate using a formula that employs the CTDIvol reported by the scanner, the injected activity and the patient’s height and weight. Effective dose from F18 was based on ICRP 106 and a size-specific correction factor that employs a blood volume estimate. Results Over a three-week trial-run in 2015, our system collected data on 118 patients (mean ± s.d: 167 ± 10 cm, 75 ± 19 kg, BMI of 27 ± 6 m2/kg). Effective dose estimates for the CT portions of the study were, mean ± s.d: 8.3 ± 2.6 mSv. Effective dose estimates for the F18 were mean ± s.d: 5.2 ± 0.9 mSv. Summed together for a total effective dose, mean ± s.d: 13.4 ± 3.2 mSv. The maximum estimated effective dose for this cohort was 23.4 mSv (17.7 mSv from CT plus 5.7 mSv from F18) and the minimum was 7.6 mSv (4.2 mSv from CT plus 3.4 mSv from F18) representing a dynamic range of 3. Effective dose from CT is strongly correlated with body mass index (r=0.89). Data tightly clustered along the regression line with only 11 estimates deviating more than 2.5 mSv from the prediction based on BMI. Conclusions We have built a system for the PET/CT clinic for real-time collection of data that facilitates individualized radiation-dose estimates taking into account both CT exposure and radiopharmaceutical administration. These data facilitate radiation dose summary reporting at the completion of the study and are collected to facilitate quality assurance and periodic review. With the protocols used in our clinic, total effective dose for PET/CT studies varies over a wide range with the majority of the dose coming from the CT scan. BMI is a good predictor of expected total effective dose.