General comments: The report reads well and generally has a logical structure that covers most of the important issues. However, there still appears to be an emphasis on diagnostic radiology with insufficient attention paid to the important radiation protection issues associated with recent developments in nuclear medicine and radiotherapy. Specific comments: P = page, L = line P23, L16-18: What is the basis for the first level of justification? Who accepts that radiation in medicine is invariably benign (like motherhood and apple pie)? Why does ICRP endorse this apparently totally unsubstantiated claim? P25, L19: It is ‘patients’ not ‘patient doses’ that are being protected so the heading would read better as – “Optimisation of Protection for Patients in Medical Exposures”. P26, L17-19: Should it not be pointed out that management of patient dose in radiotherapy cannot be achieved by use of a diagnostic reference level? P26, L19-20: When setting values for diagnostic reference levels, no direct account is taken of the suitability of the images for the diagnostic task, so it is inappropriate to claim that the use of DRLs ‘is a method of evaluating whether the patient dose is commensurate with the medical task”. DRLs simply provide a method for identifying when patient doses are unusually high, as stated on page 33 lines 19-20. The further investigation that is triggered by exceeding a DRL might well establish what dose is appropriate for the medical task, but that dose could be lower or higher than the DRL, depending on the particular local circumstances. P28, L14: Are there not many other types of ‘medical intervention’ that exist in medicine? Is not any active treatment of a patient for any condition a type of ‘medical intervention’? P29, L13: Why does this chapter on ‘Practical Methods of Protection’ not contain (or even start with) a section dealing with ‘Exposure of Patients’? P33, L9-10: How do DRLs provide a mechanism to manage patient doses to be commensurate with the medical purpose? (see comment on P26, L19-20) P34, L25: Now that DRLs have been established and used in a number of countries for many common types of x-ray examination and most common nuclear medicine procedures, is it still valid to say that “a multiplicity of levels will reduce their usefulness”? P35, L8-9: Surely it will be immediately obvious whether the observed value for any sample is higher or lower than the DRL? What is not obvious is whether the observed value for an unselected sample is an appropriate one to be comparing with the DRL. P36, L5: Do you mean ‘more than one quantity’ or ‘more than one level’? P36, L23: See comment on P26, L19-20. P38, L14: Insert ‘safe’ before ‘disposal’. P39, L5: It is not only radiologists who undertake radiological procedures (e.g. radiographers, cardiologists, dentists, etc. also do so) P39, L19: Now that there are legal requirements in many countries for regular reviews of patient doses to check compliance with DRLs and as part of quality assurance programmes, is it correct to say that ‘diagnostic exposures rarely need to be measured’? P40, L11: In view of rising concerns about the high levels of dose to healthy tissues surrounding the treatment volume associated with the new intensity modulated and image guided radiotherapy techniques and the significant risk of radiation-induced cancer, is it wise for ICRP to maintain that ‘the avoidance of accidents is the predominate (predominant?) issue’?