An Investigation of a Statistical Process Control Based Quality Assurance for Optical Stimulated Luminescence Detector (OSLD) System
K Episcopia, R Febo, D Lovelock, L Cervino, S Lim*, Memorial Sloan Kettering Cancer Center, New York, NY
Presentations
PO-GePV-T-167 (Sunday, 7/25/2021) [Eastern Time (GMT-4)]
Purpose: Optically stimulated luminescence detectors (OSLD) are widely used for in-vivo dosimetry. The current vendor recommended quality assurance (QA) process is susceptible to errors, such as calibration curve changes. In this study, we investigated a statistical process control (SPC) based quality assurance process to ensure improved consistency and accuracy within an institution's OSLD program.
Methods: All the OSLD used were nanoDotTM type OSLD (Landauer, Glenwood, IL) and read out in a MicroStarr® II system (Landauer, Glenwood, IL) calibrated with a specific set of calibration OSLDs, each exposed to doses ranging from 5 cGy to 200 cGy. A control OSLD was exposed with 100 cGy. All OSLDs were exposed in a Varian 6EX. The control OSLD was read out daily after performing the vendor suggested QA. The daily variation, V(t), on day t was calculated by taking the ratio between the readout dose at t and t-1. The upper and lower control limits, UCL and LCL, were defined as 3 standard deviations, 3σ, which were established with the initial 26 data points.
Results: The σ of V(t) was found to be 0.8% translating to a UCL and LCL of ± 2.5%. Using these limits, an additional 71 data measurements were taken. Three instances were found to breach the control limits. Two of the three failures were attributed to either calibration curve changes or chip defects that were not captured by the vendor QA. One was a false positive. The accuracy and precision of the system was estimated to be 98.5% and 66.7% respectively.
Conclusion: A quantitative QA process was established to enhance the quality control of an OSLD program and can potentially be scaled to a large hospital network. More statistics will be acquired to establish the long-term accuracy and precision of the system.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email