The document constitutes a nice review of quality assurance (QA) issues with state-of-the-art radiation therapy equipment. Its publication follows the recent trend by the ICRP to consider QA within their remit. The draft requires further editing before it is published. While the description of events which did result or could have resulted in an accidental exposure are very interesting, and the recommendations based on them are very good, some of these recommendations are repeated three times throughout the document: In the beginning (Ten Main Points), in one of the chapters (where the subject is discussed in detail) and summarized at the end (Conclusions and Recommendations). Such repetition is not necessary, either the last chapter needs to be trimmed down or the Ten Main Points, eliminated all together. There is also repetition of the events, some of which are described in the text and then repeated in Appendix A. It would be useful if the events in the Appendix which have already been discussed throughout the text are identified, as it is not clear whether they are included or simply are of the same type. Another repetition that can also be eliminated is to put the references all the end of the document, not at the end of each chapter, as many of the references are the same throughout. Regarding references, care should be taken that the ones listed belong to the scientific literature, and not given orally at congresses; in all cases they should be accessible. Furthermore, all the references need cross-checking. The one from Huq et al in the Int. J. Radiat. Oncol. Biol. Phys. is from Volume 71, not 70. Other issues are also repeated, as if different chapters had been written by different authors and not properly edited. An example is the lack of radiation field visualization in tomotherapy. In fact, issues associated with tomotherapy are mentioned so often in the document that they could be interpreted by some readers as limitations to such system. Nowhere in the document is it stated that difficulties in checking equipment performance parameters do not make the system less useful clinically. Regarding the ‘main’ recommendations, networking should be given more emphasis. It is recognized throughout the text (paragraphs 1 and 49 for example) that networking is important, and even paragraph 79 states the requirement of “assistance from computer specialists”. In fact, modern radiotherapy departments should have information technology (IT) personnel (with basic knowledge in radiation therapy procedures) as part of their staff. Another issue with staff deals with medical physicists. Although their need is recognized throughout the document, it is not sufficiently emphasized that the hiring of additional physicists and/or the re-training of existing ones should occur before a new technology is installed. The issue of in vivo-dosimetry should also be revised. In paragraphs 15 and 58, it seems to consider it an option; paragraphs 19 and 31 state otherwise. In fact, footnote 5 says “For more complex technologies, such as IMRT or tomotherapy in vivo dosimetry is not used”. None of these statements have a reference; it should be given. A quick bibliographic search has provided many references validating the use for in vivo dosimetry in IMRT. Three of these references are listed here: • P.D. Higgins et al. In vivo diode dosimetry for routine quality assurance in IMRT. (2003). Med. Phys. Volume 30, Issue 12, pp. 3118-3123 • R Ramaseshan et al. Performance characteristics of a microMOSFET as an in vivo dosimeter in radiation therapy (2004) Phys. Med. Biol. 49 4031-4048 • Investigation of the use of MOSFET for clinical IMRT dosimetric verification. Cynthia F. Chuang et al. (2002) Med. Phys. Volume 29, Issue 6, pp. 1109-1115 The chapter that needs more work is Chapter 5. The description of process flow diagrams is confusing and could be improved. The diagram mentioned in paragraph 165 is missing, unless it is Figure 3a. If so, the reference to Figure 3 in 166 should be changed to Figure 3a and the one given as fig xx to Figure 3b. Of the three prospective approaches given in 5.5, only the second one is well described. The failure mode and effect analysis being conducted by the AAPM TG 100 is poorly written and the third approach, the risk matrix, is not explained at all! Finally, the whole document seems to emphasize that the risk with the most severe consequences lies within the treatment planning phase, mainly with the treatment planning system itself. Unfortunately, whether the technology is new or not, the main risk is the miscalibration of the treatment unit in terms of absorbed dose rate.