Radiography (1997) 3, 131 142 i THE LLEGE OF R A AP H E R S A PILOT STUDY OF THE USE OF IN VIVO DIODE DOSIMETRY FOR QUALITY ASSURANCE IN RADIOTHERAPY C. Blyth* A. S. McLeodt and D. I. Thwaitest *Department of Podiatry and Radiography, Queen Margaret College, Leith Campus, Edinburgh, Ett6 8HF," tDepartment of Oneology Physics. (Clinical Oncology and Medical Physics and Medical Engineering), University of Edinburgh, Western General Hospital, Edinburgh EH4 2XU, U.K. (Received 8 March 1996," accepted 26 October 1996) A pilot study has been carried out on the use of semiconductor diode dosimetry for quality assurance in radiotherapy. 'Entrance' and 'exit' doses were measured with diodes on patients treated for head and neck malignancies on a 6MV linear accelerator. A total number of 273 treatment set-ups (117 treatment fields) were measured, 170 entrance doses and 103 exit doses. Wedged and unwedged components of mixed plain/wedged fields were recorded separately; the total number of individual measured doses were 282 entrance and 194 exit. Measured doses were compared to expected doses. The distribution of differences were observed to be gaussian. The mean of the measured entrance doses was 1.29% lower than that of expected values, with a standard deviation of the distribution of differences of 2.97%. The proportion of patient fields where the overall entrance dose delivery deviated by more than 5% from expected was 3%. The mean of the measured exit doses was 2.84% lower than that of expected values (SD 3.54%). This is partly due to patient-related uncertainties, mainly tissue inhomogeneities, and partly due to air gaps between the patient and the shell on which the diode was positioned. Deviations of >5% in overall exit dose delivery were detected in 9% of patient fields. This study has given reliable data on head and neck treatment accuracy and has demonstrated the value of the approach, as part of a quality assurance programme; giving a basis for interventional action at 2 standard deviations. Keywords: quality control; in vivo dosimetry; head and neck cancer; dose verification; treatment accuracy. INTRODUCTION Accuracy and consistency in radiotherapy are determined by the steepness and separation of dose-effect curves, for both tumour control and normal tissue complica- tions. Information surveys have led to recommendations on required accuracy in clinical dosimetry. The lnternati0nal Commission of Radiation Units and Measure- ment (ICRU) in its report 24 [1] pointed to a need for an accuracy of at least ± 5% in the delivery of the absorbed dose to a target volume in a patient. This was supported more recently by Cunningham [2]. Mijnheer et al. [3], with regard to normal tissue complications and Brahme et al. [4], considering tumour control proposed 1078-8174/97/020131+ 12 $18.00 © 1997 The College of Radiographcrs 131