Register for Updates | Search | Contacts | Site Map | Member Login

ICRP: Free the Annals!

View Comment

Submitted by Priscilla F. Butler, MS, American College of Radiology
   Commenting on behalf of the organisation
Document Diagnostic Reference Levels in Medical Imaging
 
ICRP Diagnostic Reference Levels in Medical Imaging (January 11, 2016) - DRAFT
ACR Comments
Page
#
First
Line
#
Type of comment
(General/
Technical/Editorial)
Comments Proposed change
0 1 General The report is reasonable and thorough.  
0 1 General This is well-researched and well-written guidance. The guidance on usage of DRLs (not using them to compare doses to individual patients, not using them as Notification Values for individual patients, not using them as regulatory limits, etc.) is very important.  It would be extremely useful if the guidance could provide concrete examples (case studies) showing how a facility used DRLs in evaluating and improving their examinations. 
0 1 General This report has been very thoughtfully written and provides a nice, detailed and nuanced discussion of many of the issues related to DRLs. The authors are to be congratulated on such an excellent report.  
0 1 General In general agreement with the document.  
0 1 General The document is well written and I agree with the assessment of their criteria.  
0 2 General Very important: The ACR Dose Index Registry (DIR) is a significant step forward because of its
1. Size
2. Continuous data collection (rather than a static survey)
3. Large number of individual exam types included
4. Detail of data presented
5. Automatic nature
6. Presentation of dose over time allowing for troubleshooting
7. Standardized reporting format
Obviously the authors are aware that the DIR is a big practical step forward, but little is discussed about the registry.
The DIR should be emphasized a bit more. 
0 2 General One of my general comments is that there are few comments on the important role of national registries. A search on “registry” returns one statement on p. 45 indicating that the ACR registry has collected data from 5 million CT examinations, and one on p. 95 stating that “An alternative to surveys is the establishment of a registry to which healthcare facilities submit dose data.” This is a valuable system that can aid in developing, monitoring and revising DRLs. Expand the discussion on national registries.
0 3 General There are a few instances in the document where Siemens-specific CT terminology is used where generic words would be preferred.  Check document to ensure that generic terminology is used rather than manufacturer-specific terms.
0 3 General There's a fair amount of redundancy (probably not a serious problem) and statements of the obvious, like, "If data cannot be recorded electronically it may be recorded on paper."   
10 3 Technical The term "standard-sized" is used relatively loosely in this report. The ICRP has defined a "standard man" as someone who is 180 cm in height and 70 kg in weight in ICRP 23 (1975). While this may be outdated and most of us feel our "average" patient is much larger than this, it should either be used and referenced in this report or this report should be clearer on the use of "standard-sized" patient. Without clarification, sites and regulatory bodies will interpret this to mean "average sized" and which may vary from region to region (even within a country) and will be another source of variability between DRLs from different regions. The report itself hints at this, but is not explicit in figure 7.1 on page 103, page 104 line 33 and on page 121, line 9 (section 34), where it suggests using a range of patients between 60 to 80 kg (and getting a distribution of 70 +/- 5 kg). Clarify what is meant by "standard-sized" patient and distinguish it from "average" or "median" sized patient. This could be done near the first mention of "standard-sized" patients on page 10 or it could be done later, preferably before page 11, line 29 (Section 5) where DRLs are defined with respect to a specific . The term "standard-sized" could be included in the Glossary that begins on page 13. When it is used, ICRP report 23 should be cited (or some other reference describing what is meant by "standard-sized").
10 20 Technical I'm left a little confused about the median vs. 75th percentile statements. Does this mean that individual sites should determine their median value of the remainder of data after top and bottom 25% are discarded, and the national, regional, etc. 75th percentile of the medians are the DRL? Then is it saying that the local median being equal to the broader median is an indication that image quality may be compromised or just that that individual median be used in some undefined way to evaluate if the site may have too low a dose? (page 10 lines 20 - 24). The presentation is confusing. The discussion needs to be rewritten for clarity.
10 36 Editorial Here it says regulatory agencies "may" use DRLs. It sounds like the ICRP is giving its approval of doing so.  Probably better wording would be "might" which is just stating that it is possible but doesn't imply approval.
11 13 Editorial Values of DRL quantities from patient examinations are collected from at several different health facilities. Delete "at"
11 19 Editorial 3-5 years: Out dated for CT because of new technology. This time frame is mentioned multiple times in the document. Newer automatic dose registry technologies such as the ACR CT Dose Index Registry allow for continuous collection of a site's data, and provide detailed semi-annual reports. Their continuous nature allows not only for assessment of new technologies and protocols, but it can also be a troubleshooting tool that can reveal when a site deviates from prescribed protocols due to training issues, equipment malfunctions, etc.
11 28 General Do the authors mean images, or data, or both? For example, iterative algorithms in CT really manipulate the data, not the images.  
12 4 General I can see how the administered radiation will vary, but this should not be the same as the dose absorbed by the patient  
13 26 Technical Constancy testing – this term is ONLY used by Siemens. I’m not sure this needs to be in the glossary at all. Perhaps a search on this term will show where it is used in the document? If a periodic test more complete than a daily QC (such as an annual CT survey) is meant, this phrasing should be used instead of ‘constancy testing.’
19 31 Editorial Delete the word "Value" at the beginning of the sentence "DRL quantities for individual patients …"
19 35 Editorial Does the Introduction section really need it's own Introduction?  
21 33 Editorial effectiveness of DLR: as documented in the article, it has been shown that DLR can be an effective tool for change. But recently, the ACR DIR has been well documented as a tool that also has clear impact. Discuss and give references such as:Lukasiewicz, A., Bhargavan-Chatfield, M., Coombs, L. et al, Radiation dose index of renal colic protocol CT studies in the United States: a report from the American College of Radiology National Radiology Data Registry. Radiology. 2014;271:445–451.
35 16 Technical I disagree that "setting DRL values by using phantom-based surveys is not appropriate." Although setting DRLs by evaluating data from standard size patients obtained from a very large population of patients is preferable, it may not be feasible due to unavailability of trained staff or resources in general in many situations. Phantoms serve as effective "standard-sized patient" surrogates to allow much more appropriate comparison of data than using data from small numbers of patients that may be of vastly different sizes. The guidance should allow for this.  
40 31 Technical Although standardization of patient size via weight restriction is one approach, a more accurate approach is to standardize patient size via body part dimensions. This is more accurate due to the variability of weight distribution within individual patients. As an example, for a chest x-ray, a tall patient of a given weight will have a thinner PA dimension than a short patient of the same weight. For cross-sectional procedures such as CT, this is an easily determined dimension.  Acknowledge the value of using a body part dimension approach in setting "standard-sized patients" for establishing DRLs.
41 3 Technical "Where this is possible [collect data for large numbers of patients], restrictions on weight can be removed."
This would be problematic in countries with well-established geographic variations in body habitus. Geographic areas with populations of large patients would be averaged in with geographic areas with smaller patients. When hospitals in these large-patient regions compare their data against the national DRLs, they may be higher that the rest of the country, but with good reason. Developing DRLs based on standard size patients (or even phantoms) would eliminate inappropriate comparisons.
Modify sentence to point out risks of inappropriate comparisons.
42 4 Technical Body part dimension is very useful. See comment for page 40. Suggest modifying to say, "patient weight or body part dimension"
42 5 Technical SSDE is available in many cases.  Suggest adding SSDE.
45 5 Technical The 2009 Spelic article is about the NEXT study, not the ACR Dose Index Registry. Remove Spelic article.
45 21 Editorial Figure 2.1 shows 1 method for data display, but the "box and whisker" plot from the ACR dose registry; and the graph of individual exam dose by day should also be presented, as these are the data the reader will be most often assessing.  
49 10 Technical On page 47, line 1, the report seems to define the "DRL value" as the 75th centile of the distribution of the DRL quantity being measured. Therefore the statement on page 49, line 10 that states "where the local median value of the DRL quantity is greater than the national or regional DRL value" then practice is not optimized. Is this meant to be interpreted that the median of a local value should be compared to the 75th centile of a national or regional value?  Clarify the terms DRL quantity, DRL value and the median value of the DRL quantity
50 12 Technical Similarly, on page 50, line 12 states "Since the national values typically represent the 75th percentile values for the national distributions of DRL quantities, the median of the available national DRLs should approximate the 75th percentile value to be expected from a regional patient survey.". Perhaps this can be illustrated how the median of available national DRLs approximate the 75th centile value. Clarify the terms DRL quantity, DRL value and the median value of the DRL quantity
71 26 Technical For CT-guided interventional procedures, DLP may not be a suitable DRL quantity because the CT imaging required typically takes place over a narrow range of scan lengths (L) and may result in a DLP value that is unusually low in comparison with "standard" CT acquisitions.  "the Commission recommends that DRLs be established for CTDIvol, the number CT sequences obtained, and CT fluoroscopy time." 
80 8 General The entire paragraph from lines 8 to 16 discusses PKA and Ka,e but not Ka,i. Note that the "2014 American College of Radiology ACR-AAPM Practice Parameter for Diagnostic Reference Levels and Achievable Doses in Medical X-ray Imaging" recommends using Ka,i as the suitable DRL for (projection) radiography. See: http://www.acr.org/~/media/ACR/Documents/PGTS/
guidelines/Reference_Levels_Diagnostic_Xray.pdf 
A discussion of Ka,i is warranted, specifically it's relevance as a DRL quantity in projection radiography.
81 10 General ICRP suggests DRLs should be technology specific, siting the use of iterative reconstruction in CT. (1.5.4) If this were carried to the extreme for other technologies, then sites using a high speed screen technology (say 800) should not be compared to sites using a lower speed (400) because the difference is technology dependent. But this ignores one of the major values of making these comparisons, namely to point out to sites that they may be using exceptionally outdated technology.   
81 24 Technical As mentioned above, patient dimension is more useful.  
83 10 Technical There is a statement that says if CTDIvol is not displayed, then it will have to be calculated from CTDIw and pitch. Pitch is always displayed, but if CTDIvol is not displayed then it is VERY unlikely that CTDIw will be displayed or be otherwise readily available. I’m not sure that this guidance is very helpful without explaining how one would be expected to find CTDIw.   
83 13 General The paragraph makes reference to a "standard-size" patient, but the section does not appropriately describe or define what "standard-size" is for setting CT DRLs. Recommend modifying the sentence that begins in line 16 as follows: "DRL values are designed to help determine whether the amount of ionising radiation applied for a medical imaging procedure in a representative sample of patient size ranges (note: bold font added only for emphasis of recommended change.), for a defined clinical task, is too high or too low."
83 13 General The "2014 American College of Radiology ACR-AAPM Practice Parameter for Diagnostic Reference Levels and Achievable Doses in Medical X-ray Imaging" recommends the use if CTDIvol as the suitable DRL for CT, paired with the use of "Patient Lateral Dimension" to accommodate the impact of patient size on developing a procedure-specific DRL. Recommend modifying the sentence that begins in line 20 as follows: "It is therefore necessary to ensure that the survey data reflect values for appropriate patient size ranges." (note: bold font added only for emphasis of recommended change.)
83 21 Technical Although it would be ideal if only ‘standard-size’ patients were used for a DRL survey, the idea that any patient could be identified as ‘standard-size’ is difficult to understand. ‘Standard-size’ patients would be represented by a range of sizes (however that is defined), and identifying the cut-off points for that range would be fairly arbitrary. If there is a standard way to do this, can it be explained? In CT especially, automatic exposure control is heavily used which is completely dependent on patient habitus. DRLs for CT are heavily influenced by patient population characteristics.  
84 34 Technical Although it is somewhat implicit based on reading the preceding paragraphs, DRLs for planar nuclear medicine does not explicitly state that it should also take into account the radiopharmaceutical used. In the PET section of the document (page 89, lines 14-16) it clearly states that "DRL values have to be set for each radiopharmaceutical". Recommend modifying the sentence that begins in line 37 as follows: "The Commission recommends the establishment of weight-based administered activities (MBq/kg), set for each radiopharmaceutical (note: bold font added only for emphasis of recommended change) for all types of nuclear medicine investigations except for those where the radiopharmaceutical is concentrated predominantly in a single organ ..."
85 4 Technical Similar to the comment above for page 84 line 219, DRLs for SPECT imaging procedures should also explicitly take into account the radiopharmaceutical used. Recommend modifying the sentence that begins in line 4 as follows: "The Commission recommends the establishment of weight-based administered activities (MBq/kg), set for each radiopharmaceutical (note: bold font added only for emphasis of recommended change)."
85 11 Technical Incomplete references. Should CRCPD 2006 be CRCPD 2003? (I could not find CRCPD 2006 in the references.)
Also, suggest adding ACR guidance on nuclear medicine procedures - http://www.acr.org/Quality-Safety/Standards-Guidelines/Practice-Guidelines-by-Modality/Nuclear-Medicine
85 16 Editorial “Since the majority of hospitals and clinics use recommended administered activity levels or lower levels, there is less inter-departmental variation in patient dose than in diagnostic radiology.”
How does ICRP know that “the majority of hospitals and clinics use recommended administered activity levels or lower levels.”
A reference is needed to support this statement.
85 33 Editorial “Some institutions where a large volume of nuclear medicine procedures is performed administer more activity to patients in order to achieve higher patient throughput. From a radiological protection point of view, this is not desirable.”
This statement appears to contradict the statement on page 85 that “majority of hospitals and clinics use recommended administered activity levels or lower levels”.
Also clarification is needed on what is meant by "not desirable."
 
86 16 Editorial “Since more than 90% of current PET examinations use 18FDG, this section discusses only 18FDG PET and PET/CT”. Need reference for 90%.
86 34 Editorial “The patient should be encouraged to drink water and then void prior to scanning, in order to limit the radiation dose to bladder.” Would recommend rewriting this to “The patient should be encouraged to drink water and then void prior to scanning to decrease the whole body radiation dose, especially to the bladder.”
86 36 Editorial For PET/CT (p 87) the authors mention some practical ways to decrease dose; they may consider adding a few similar comments for general nuclear medicine such as optimizing camera technique and other maneuvers such as hydration, etc. Suggest referring to publically available resources, such as those on the Image Wisely website for general nuclear medicine. See http://www.imagewisely.org/imaging-modalities/nuclear-medicine/articles/clinical-aspects 
88 18 Technical ‘topogram’ or ‘topogram/scout’ – both of these are manufacturer’s terms (Siemens topogram and GE scout).  The generic term ‘localizer radiograph’ should be used instead of topogram or scout image.
88 25 Editorial "For a diagnostic contrast-enhanced CT, standard protocols should be used. It is preferable to perform a diagnostic CT only for limited portions of the body. For the rest of the body, a low-dose CT is sufficient for attenuation correction and anatomic localization."
This is not usually done. If a diagnostic CT is ordered usually the entire scanned CT portion of the body is done at standard diagnostic setting.
 
92 2 General The pediatric section is somewhat rambling. It discusses DRLs but says that many of these are unclear in many respects, particularly such important information as the diameter of the phantom used in the calculation. A summary paragraph putting all the discussed information together would be very helpful. The very significant data already present in the ACR registry should factor into these discussions.
92 16 Technical Collimation is also very important for pediatric fluoroscopic imaging, especially for interventional fluoroscopic procedures. Recommend modifying the sentence that begins in line 16 as follows: "For projection radiography, fluoroscopy, and interventional fluoroscopy, (note: bold font added only for emphasis of recommended change) the relevance of appropriate collimation is higher in pediatrics."
92 30 Technical Although it is true that CTDIvol and DLP are based on a 16 cm phantom for head examinations for both adult and pediatric procedures, CTDIvol and DLP are not always based on a 32 cm phantom for body examinations, particularly for pediatric body examinations. Some vendors utilize a 16 cm phantom for pediatric body examinations. In addition, the American College of Radiology CT Accreditation Program as well as the "2014 American College of Radiology ACR-AAPM Practice Parameter for Diagnostic Reference Levels and Achievable Doses in Medical X-ray Imaging" requires the use of a 16 cm phantom for pediatric body examinations. Recommend modifying the sentence that begins in line 30 as follows: "For CT, the DRL quantities are CTDIvol and DLP. These DRL quantities are based on calibration with a 16 cm phantom for head examinations, for both adult and pediatric examinations. These DRL quantities are based on calibration with a 32 cm phantom for adult body examinations, but may be based on either a 32 cm or 16 cm phantom for pediatric body examinations." (note: bold font added only for emphasis of recommended change)
93 5 Editorial The authors may choose to mention that the simplest method to lower dose is to use alternative modalities such as US or MR. For pediatrics, for example, appendicitis is a common CT indication and US has always had a role, but MR has in increasing role. Also, US has an important role in pediatric intussusception, and can help avoid fluoroscopy.  
95 10 Editorial The word "overlies" should be "overlays" since the verb is transitive, i.e., has a direct object (data points). Lie, as in to lie down, is an intransitive verb. overlays
110 22 Technical ‘topogram’ or ‘topogram/scout’ – both of these are manufacturer’s terms (Siemens topogram and GE scout).  The generic term ‘localizer radiograph’ should be used instead of topogram or scout image.
112 18 General Section 7.5.2 should cite the AAPM's CT protocol practice guideline which provides detailed advice on how to review and manage protocols for CT. : AAPM medical physics practice guideline 1.a: CT protocol management and review practice guideline. Cody DD, Fisher TS, Gress DA, Layman RR Jr, McNitt-Gray MF, Pizzutiello RJ Jr, Fairobent LA. J Appl Clin Med Phys. 2013 Sep 6;14(5):3-12. doi: 0.1120/jacmp.v14i5.4462.  Please cite the AAPM medical physics practice guideline 1.a: CT protocol and review practice guideline in this section. This practice guideline gives an even more detailed suggestions as to what protocol parameters to examine than are provided here. 
112 38 Technical "Geometric efficiency" is usually not something that the user can select and few systems provide information regarding this parameter. It is primarily linked to beam collimation, so these two should not be listed separately.  Page 112, line 37 be changed to list beam collimation and geometric efficiency as one item: (b) Beam Collimation (geometric efficiency)
113 7 Editorial This sentence does not seem to be complete: “When CT protocols are set up, the process should take into account how the interaction of the different variables that can be set.”  Perhaps a better version is: “When CT protocols are set up, the process should take into account of the interactions of the different variables that can be set.”
118 11 Technical The statement that a "A DRL value is considered to be exceeded when the local median value of a DRL quantity for a representative sample of standard-sized patients is greater than the local, national, or regional DRL value." is confusing as this again seems to imply that the median value of the quantity is what is being compared here.  Clarify the terms DRL quantity, DRL value and the median value of the DRL quantity
118 20 Editorial The meaning of this statement is not clear. Could an example or two be provided to help convey the intended meaning?  
124 5 General The term ‘clinical audit’ is used but has not been described or defined previously. Is this meant to be the same as a DRL survey?  
124 8 General The term ‘clinical audit’ is used but has not been described or defined previously. Is this meant to be the same as a DRL survey?  
124 9 General The term ‘annual survey’ is used without a clear definition.  












Learn to solve a Rubix Cube easily with the simple layer-by-layer method. Please subscirbe for a premium HTML editor membership to stop adding promotional messages to the edited documents.