|Dear Dr Clement and ICRP authors,
The Netherlands Committee on radiation dosimetry would like to thank the authors for their commendable effort in the draft report entitled ëDiagnostic Reference Levels in Medical Imagingí. The NCS is a Dutch foundation supported by the Dutch societies for Radiology, for Nuclear Medicine and Radiation Oncology. In addition, Medical Physicists and Radiation Protection Experts and technologists are represented in the NCS. For a full description of the supporting societies and more information regarding the NCS and the NCS platform see http://radiationdosimetry.org.
First of all, I would like to state that the NCS supports the concept and implementation of diagnostic reference levels (DRLs). The NCS appreciates ICRP's efforts in the draft report to clarify the concept, supporting correct interpretation, and to provide practical guidance to implement DRLs. The NCS considers the current draft report a great improvement compared to ICRPís publications of 1996 and 2001. However, the NCS has some reservations concerning the proposed parameters for implementation monitoring and has some additional suggestions that might be helpful to improve both interpretation and implementation.
Concerning weight as most important determinant of patient dose:
That the standard-sized patient, for which DRLs are to be established, is most importantly determined by its weight, is recognized and discussed in many paragraphs of the draft report. In these discussions, possible corrections are suggested for dose estimation of non-standard sized patients. The NCS would like to comment on two correction issues:
- (Non) weight-based administered activity as DRL quantity.
The draft report is not consistent in the DRL quantity for nuclear medicine. In some paragraphs, administered activity is proposed (e.g. in 2.3.4, 'Examinations and DRL quantities'), while in other paragraphs administered activity per unit of body weight is given (e.g. in 5.4, 'DRLs in planar and SPECT nuclear medicine imaging'). The latter suggests that the DRL can be determined from the sample population by proportional correction of activity with patient weight, so according to the NCS, the use of administered activity per unit of body weight is preferred over administered activity alone and this should be applied consistently in the report. We are fully aware that both EANM and SNMMI guidelines advise fixed amounts of administered activity for many nuclear medicine procedures, predominantly for gamma cameras. However, that does not mean that there is no room for optimization. We are also aware that literature suggests more complex corrections (e.g. with quadratic weight for F18-FDG , but the use of this more simplistic linear correction is still indeed a better alternative then absolute dose figures, because it eliminates the necessity to select only standard-sized patients therefore large patient populations to establish significant large reference numbers.
- Dose estimation of non-standard sized patients in general.
If the NCS understands correctly, the discussed corrections for dose estimation, e.g. size-specific dose estimates (SSDE) for CT (discussed in 5.3.2, ëConsiderations for DRL surveys in CTí), are not meant to establish DRLs, but an additional utility to improve dose estimations for non-standard-sized patient groups to be used in the optimization process. If so, the NCS agrees with this. Merely the fact, that we have doubts about our understanding indicates that this might be clarified more through the report, so the NCS suggests to do this.
In addition, the NCS suggests the report elaborates on the limitations of using a representative sample of standard-size patients, being that this will not extrapolate with certainty into good optimization for non-standard sized patients. The example above given a preferred quadratic relation between weight and administered activity rather than a linear relation, hints at this. Also the EANM guidelines use an absolute upper limit of 530 MBq administered activity . Higher doses do not give better images and a longer scan duration is needed for an equivalent good image. Again, this is an example showing the complex relationship between dose exposure and image quality. Similar considerations hold for CT, given variations in size and CT density among patients, where automatic dose modulation accounts for this. Many parameters are at play in these cases and therefore optimization may not be specifically tailored for these non-standard sized patients.
Concerning other determinants of patient dose:
The draft report recognizes that, apart from weight, there are other dimensions to optimize patient dose at a given image quality, that therefore influence the establishment of DRLs. The NCS would like to comment on the following dimensions:
- Complexity of interventional procedures.
The draft report states that complexity of interventional procedures may be considered in setting DRLs and a multiplying factor for the DRL value (e.g. 2 or 3) may be appropriate. The NCS agrees that setting DRLs for interventional procedures is feasible, but hampered by complexity. Since the use of a multiplication factor will lead to debate, at least in The Netherlands, the NCS suggests to leave out the given multiplication factors 2 or 3 and to mention that how to deal with complexity of interventional procedures, is left to national bodies. It is valid though, to stress that a measure to correct for or incorporate complexity is essential to eliminate a source of great variations.
- Duration of nuclear medicine examinations.
Administered activities in nuclear medicine examinations are related to the required duration of the examination, as is underpinned in the draft report. The NCS agrees that, in general, reduction of acquisition time by administration of a higher activity is not desired. However, in some cases this may be justified, for instance for the comfort of individual patients or to reduce waiting time for an examination in case of many applications. This can be a leading argument in case of limited therapeutic windows (in oncology and cardiology for the most part). The NCS suggests that the report explicitly mentions justified exceptions and elaborates on them, e.g. in paragraph 5.4 ('DRLs in planar and SPECT nuclear medicine imaging').
Concerning image quality:
To the NCSís delight, the draft report states that highest priority for any diagnostic examination is achieving image quality sufficient for the clinical purpose, and DRL quantities are not descriptors of image quality (in paragraph 2.7, 'Image quality'). However, on the issue of image quality and the information contained in the image, the NCS would like to address the following:
- Development of objective image quality criteria.
The draft report discusses image quality criteria (in paragraph 2.7), but so far, for assessing image quality only subjective measures are used in the form of evaluations by radiologists, where more objective parameters are wanted. Research in that area is carried out, but no promising relationship between physical parameters and clinical judgement by radiologists has been established yet . Although development is at an early stage, the NCS suggests to elaborate on the need of development of objective image quality criteria in the report, e.g. in paragraph 2.7. Ideally the leading parameter for DRLs is not just radiation dose, but rather image quality per dose. Since this would eliminate influencing parameters on radiation dose that are needed for the clinical task at hand. Unnecessary exposure would then be traceable per single study rather than being reliable on a sufficiently large group to eliminate these parameters.
- Anticipation on foreseen follow-up diagnostic imaging.
The draft report focuses implicitly on imaging information for a single purpose, e.g. to answer one diagnostic question. Also, available literature  suggests that for specific diagnoses (e.g. urinary tract stones) considerable dose reduction might be available, by assessing only the initial diagnostic question. However, clinical reality is that if the answer is positive, the immediate follow-up questions would be where the problem is exactly located, what the extent and severity of the problem is, etc. If the initial diagnosis is negative, the immediate follow-up question would be what then is causing the problem. Both follow-up questions are foreseen and may justify anticipation by extending the information in the first imaging of the patient. This, of course, leads to a higher patient dose due to this first image, but reduces the total patient dose, since no follow-up imaging is needed. The NCS suggests that the report explicitly discusses the justified anticipation on foreseen follow-up diagnostic questions, e.g. in paragraph 2.7.
- Emphasis on dose reduction in advertising.
Nowadays, vendors not only promote and invest in image quality, but also largely in dose reduction techniques, which is appreciated by the NCS. Nonetheless, in their advertising, they tend to emphasize the dose reduction (e.g. of ëhalf time imagingí or iterative image reconstructions in CT), and pay less attention to the risk of reduced image quality by these techniques [5, 6]. These risks should be assessed and dealt with at introduction of the technique in clinical practice, but not all institutions may have the means or expertise to do this. Although the NCS realises that risks related to emphasis on dose reduction in advertising is not a main concern in the practical use of DRLs, the NCS tentatively suggests to address this issue in the report, e.g. in paragraph 2.7. By doing this, the report would stress that optimization is not the same as using propagated protocols from the selling vendor and that commercial goals do not necessarily align with clinical goals.
On behalf of the NCS, sincerely,
Dr. Jeroen van de Kamer, chair of the NCS
1. de Groot H, et al. Optimized dose regimen for whole-body FDG-PET imaging. EJNMMI Research 2013, 3:63
2. Boellaard R, et al. FDG PET/CT: EANM procedure guidelines for tumour imaging: version 2.0. EJNMMI 2015, 42:328-354
3. De Crop A, et al. Correlation of clinical and physical-technical image quality in chest CT: a human cadaver study applied on iterative reconstruction. BMC Medical Imaging 2015, 15:32
4. Niemann T, et al. Diagnostic Performance of Low-Dose CT for the Detection of Urolithiasis: A Meta-Analysis. AJR 2008, 191:396-401
5. Ardenfors O, et al. Reduced acquisition times in whole body bone scintigraphy using a noise-reducing PixonÆ-algorithmóa qualitative evaluation study. EJNMMI Researcj 2015, 5:48
6. Guimaraes LS, et al. Appropriate patient selection at abdominal dual-energy CT using 80 kV: relationship between patient size, image noise, and image quality. Radiology 2010, 257(3):732-42