The statistical analysis of failure of a MEVATRON77 DX67 linear accelerator over a ten year period

Purpose: A linear accelerator (linac) takes a leading role in radiation therapy. A linac consists of complicated main parts and systems and it is required that highly accurate operational procedures should be maintained. Operational failure occurs for various reasons. In this report, the failure occurrences of one linac over a ten year period were recorded and analyzed. Materials and Methods: The subject model was a MEVATRON77 DX67 (Siemens, Inc). The failure rate for each system, the form classification of the contents of failure, the operation situation at the time of failure, and the average performance life of the main parts were totaled. Moreover, the relation between the number of therapies that patients received (operating efficiency) and the failure rate within that number and the relation between environment (temperature and humidity) and the failure rate attributed to other systems were analyzed. In this report, irradiation interruption was also included with situations where treatment was unable to begin in total for the number of failure cases. Results: The cases of failure were classified into three kinds, (A): irradiation possible, (B): irradiation capacity decreased, and (C): irradiation impossible. Consequently, the total failure number of cases for ten years and eight months was 1,036, and the number of cases/rate of each kind were (A): 49/4.7%, (B): 919/88.7%, and (C): 68/6.6%. In the classification according to the system, the acceleration section accounted for 59.0% and the pulse section 23.2% of the total number of failure cases. Every year, an operating efficiency of 95% or higher was maintained. The average lives of a thyratron, a klystron, and RF driver were 4,886 hours, 17,383 hours, and 5,924 hours respectively. Moreover, although analysis of the relation between the number of therapies performed (or operating time) and the number of failures for each main machine part was observed, the tendency was not to associate them with each other. The relation between environment and the number of failures, it was observed that in the acceleration and pulse sections, failures increased as climatic temperatures fell. Conclusions: It is necessary to record equipment problems and failure in detail over a long period of time, to perform complete analyses, and to have a good working knowledge of equipment functions. Moreover, it is important to attempt to forecast future failures from observed failure patterns and, if possible, to avoid total failure in advance with prompt maintenance, and to have backup support organized in other equipment and at other institutions if equipment should break down.