Statistical process control in monitoring radiotherapy quality assurance program: An institutional experience

Introduction: Statistical process control (SPC) is a handy and powerful tool for monitoring quality assurance (QA) programs in radiotherapy. The potential reasons for an uncontrolled QA state monitored via SPC can be identified using the cause-and-effect diagram. This study explains the institutional experience in monitoring weekly output constancy of medical linear accelerator (Linac) and patient-specific quality assurance (PSQA) using SPC and cause-and-effect diagram. Material and methods: Prospective monitoring of output constancy has been demonstrated by the simultaneous usage of Shewhart’s I-MR charts and time-weighted control charts. Two hundred and forty-one PSQA results were retrospectively analysed in a combined γ and dose volume histogram (DVH) based analysis using control charts and process capability indices. A PSQA analysis method has been illustrated in which the site-specific action limits (AL) and control limits (CL) for γ and DVH based analysis were obtained using SPC. Results: The simultaneous use of different control charts indicated a systematic error in the output constancy of Linac as successive measurement points fell above the CL. The reason for failure in output constancy was found using a cause-and-effect diagram due to a faulty monitor ion chamber. The obtained AL and CL for γ and DVH based analysis were used to decide pass or fail criteria in PSQA. Among the analysed treatment plans, four Head and Neck (HN), two Central Nervous System (CNS), four Gastro-Intestinal (GI), and four Gastro Urinal (GU) plans failed the PSQA analysis. Cause-and-effect analysis of these failed treatment plans in PSQA pointed out six primary potential sources of errors in the results. Conclusions: SPC tools can be adopted among institutions for consistent and comparable QA programs. If the QA process monitored using SPC falls outside the CL, cause-and-effect diagrams can be used to extract all possible contributing factors that lead to such a process state.