Quantitative assessment of inter-clinician variability of target volume delineation for medulloblastorna: quality assurance for the SIOP PNET 4 trial protocol

Background and purpose: To assess inter-clinician variability amongst specialist paediatric radiation oncologists in delineating clinical target volumes for treating medulloblastoma as a quality assurance exercise prior to the introduction of the SIOP PNET 4 trial protocol of conformal radiotherapy to the posterior fossa and tumour bed. Patients and methods: Participants from 17 UK centres attended an educational meeting and then completed a clinical planning exercise to outline: (1) the whole posterior fossa and (2) the tumour bed. Quantitative analysis of the volumes, lengths, spatial positioning and axial planes for each individual was carried out and variation between individuals analysed. Results: Outlining of the posterior fossa was reasonably consistent, although most variation was seen in defining the superior border of the tentorium. A major difference was the decision whether or not to include the post-surgical meningocoele in the clinical target volume (CTV). The CTV for the tumour bed was under treated by all participants due to lack of inclusion of pre-operative tumour extent. Conclusions: This exercise demonstrated several ambiguities in the draft protocol and highlighted particular areas of inter-clinician variation. Consequently the protocol was revised and improved to take account of these findings. We recommend that planning exercises, in conjunction with education and training, should be implemented before the start of any new radiotherapy trial. In the future, the use of image transfer will allow prospective peer review of target volumes before treatment commences. These measures are essential to ensure that alterations in clinical practice are achieved in a uniform way. (C) 2003 Elsevier Ireland Ltd. All rights reserved.