Nov. 8, 2010

Stephen Vastagh
Director, International and Industry Programs
Industry Director, X-Ray Section
Medical Imaging Technology Alliance,
A Division of the National Electrical Manufacturers Association
1300 North 17th Street, Suite 1752
Arlington, Virginia 22209

Dear Stephen,

In reviewing our investigation of the anomalously high radiation exposures first reported in September 2009 in connection with CT brain perfusion studies, we think that it would be useful to summarize some of the lessons learned. The following items pertain for the most part to re-organizing or developing additional equipment-operation information that users could readily access, understand, and apply to reduce the possibility of inadvertent application of unnecessary amounts of radiation. Some points entail changes to software defaults and hardware features. These ideas surfaced during our investigation and reflect issues that appeared to contribute to the possibility of unintended high exposure. Please share these ideas with all of the CT manufacturers in MITA.

1. Brain-perfusion CT—Provide particular information and training on brain-perfusion protocols to all facilities that receiving base CT equipment, whether or not the facilities purchase the related software enabling quantitative analysis of cerebral hemodynamics. Suggested information to provide:

• Manufacturer-recommended parameter settings, i.e., tabulated or listed for each scanning-parameter set and supplemented with the corresponding values expected for the dose indices CTDI_vol and DLP. (Presumably such dose-index values would be typical of each of the associated scanning sequences comprising a complete brain-perfusion protocol.)
• A concise description for each scanning-parameter set that clearly distinguishes the particular role of its corresponding scanning sequence from the roles of the other scanning sequences comprising a complete brain-perfusion study.¹
  Such descriptions may reduce the potential for inadvertently substituting technique values associated with diagnostic-quality imaging phases for the lower-dose scanning-parameter settings actually needed for contrast-perfusion phases.
• Explanation of why a relatively poorer quality image is appropriate to reduce radiation dose in the contrast-perfusion phases in comparison to the relatively better image quality that might be needed for any diagnostic phases included in a protocol.
• Explanation of how peak skin doses relate to CTDI (because brain perfusion studies are associated with relatively high skin doses).²

2. Automatic Exposure Control (AEC)

Clarification of parameters affecting dose, along with clear instructions on how to appropriately set those parameters.

Description on how to choose AEC vs. manual modes, including examples of situations—for example, brain perfusion studies—in which AEC operation might unnecessarily complicate successful operation without additional dose savings and with no improvement in diagnostic efficacy.

Emphasis on building or modifying protocols--to check the need to change manufacturer-recommended or default AEC parameters to achieve optimal dose reduction, including the potential for an unintended dose increase (compared to manual-mode operation) if AEC parameters are not checked.

Directions on how to modify manufacturer-recommended scanning-parameter sets, if a user were to elect an AEC mode in lieu of a recommended manual-mode, the values of all dose-associated AEC-configurable parameters could default (until otherwise changed) to those that would yield a value for the sequence-maximum mAs that would match the mAs value recommended in the manual mode.

3. Pop-up Notification at Threshold for Deterministic Injury

Institute a pop-up notification so that prior to scanning, when the operating conditions associated with any protocol yield an expected value of CTDI_vol greater than or equal to 1 Gy, a notification would alert the operator that a high radiation dose, potentially leading to the development of clinically significant cataracts, skin injury, or hair loss, would be incurred by the patient were scanning to proceed.

4. User-Accessible Organization of Dose-Related Information

Organization of all dose-related information into one section of each user manual, in a dedicated dose manual, or indexed comprehensively in a concordance covering all manuals.

5. Protocol Specifications

Provide facilities with hard copies or pdf files specifying the dose-associated parameter settings recommended for particular clinical applications, including distinct “scanning-parameter sets” for for the values of each scanning sequence comprising a protocol and supplemented with the corresponding values (or, perhaps, range of values associated with AEC operation) typical for the dose indices CTDI_vol and DLP (with phantom diameter and length identified), each set tabulated or listed. In particular, the following elements could be included in the table or list:

1. Name or identifier for each manufacturer-recommended scanning-parameter set as part of the various protocols, the anatomical area covered, and a concise description of the purpose of each particular scanning-sequence within a protocol.
2. Specification of operating conditions and settings:
   • Type of scan (e.g., axial or helical)
   • Scan-acquisition time, slice thickness, kV, mA, etc.
   • Mode of operation (e.g. AEC or manual)
   • Values for all dose-related AEC parameters if an AEC mode is recommended or selected by the user.
3. Typical values of CTDI_vol and DLP with an explanation of how the typical values were determined if the protocol allows for a dose range (e.g. AEC mode).

We are interested in hearing industry’s reaction to these points, in discussing them further at the upcoming stakeholders meeting you are planning at RSNA, and in continuing a constructive dialogue on how these ideas might fit into our Initiative to Reduce Unnecessary Radiation Exposure from Medical Imaging.

---

¹ For example, in a brain-perfusion study the first sequence might entail a non-contrast phase of conventional axial scanning spanning the cranium to check for bleeding. The second sequence could be stationary-table or shuttle-mode scanning with contrast perfusion for the purpose of post-scanning quantitative analysis of hemodynamics. A third sequence might entail an additional phase of conventional axial scanning of the cranium, this phase making use of the residual contrast medium to identify potential brain lesions.

² The following article provides an explanation of the relationship between CTDI vol and peak skin dose for helical and cine-mode exams: J. A. Bauhs, T. J. Vrieze, A. N. Primak, M. R. Bruesewitz, and C. H. McCollough, 2008, "CT dosimetry: comparison of measurement techniques and devices," Radiographics Vol. 28, pp. 245-253. Available at: http://radiographics.rsna.org/content/28/1/245.full.pdf+html.