On the stability of well-type ionization chamber source strength calibration coefficients

Purpose To investigate the variability and stability of brachytherapy source strength calibration coefficients for air-communicating and pressurized well-type ionization chambers among high-dose rate (HDR), low-dose rate (LDR), and electronic brachytherapy (EBT) source qualities. These qualities of an ionization chamber are important features to assess given their role in maintaining traceability to a primary national standards laboratory and facilitating efficacious patient care in brachytherapy. Methods The calibration records from the University of Wisconsin Accredited Dosimetry Calibration Laboratory (UWADCL) customer well-type ionization chamber database were retrospectively analyzed for calibrations performed between 1996 and 2019. A statistical analysis was performed and differentiated among calibration type to quantify the distribution of chamber calibration coefficients among several chamber models. Distributions were quantified based on their moments and by quantile analysis. For LDR calibrations, chamber response was further differentiated by seed type to study the variability in seed dependence within and across chamber models. In addition to these metrics, the calibration lineage at the UWADCL was used to assess the stability of these calibration coefficients among chamber model and source type based on the ratio of subsequent calibration coefficients. Results The distribution of brachytherapy source strength calibration coefficients for a particular chamber model is not necessarily normally distributed and is sensitive to changes in the machining tolerances or design of the chamber model. Calibration source quality also influenced the distributions of calibration coefficients for a chamber model; the air-kerma rate calibration coefficients for EBT sources were the most variable followed by LDR and then HDR source types. The stability of a chamber's source strength calibration coefficient exhibited a similar dependence on the source quality. Air-communicating and pressurized chambers exhibited an average stability between subsequent calibrations of 0.2% and 3.0%, respectively, for HDR calibrations, but could exhibit more than double this variability characteristic of their leptokurtic distributions. For LDR calibrations, the spread and stability in a model's calibration ratio toward another seed type is extensive and notable. For some seed types and chamber models exhibit variations of <0.5% while others exceeded 2.0%. Furthermore, the magnitude of this dependence is outside the variability of the source's source strength due to manufacturing, which was determined from the manufacturer and NIST source strength intercomparison records at the UWADCL. As a result, establishing and disseminating calibration conversion factors among different source and chamber models is not advised as it would substantially increase the uncertainty in a clinical user's determination of source strength. Conclusions The calibration of a well-type ionization chamber is unique to the chamber, source, and source holder. Attempting to generalize source strength calibration coefficients among chambers of the same model or source type is impractical given the variation in response observed across the calibration history of the UWADCL. Attempting to quantitate and transfer calibration coefficients in such a relative sense may significantly degrade the uncertainty relative to specifically calibrating a chamber for an individual source.