APPENDIX C.Patient Dose Optimization and Management

C.1. Patient Dose Optimization

The following list of dose optimization feature sets illustrates the huge efforts that the CT community has developed for the performance of high quality exams. The CT clinical medical physicist benefits from an understanding of dose optimization technology in general.The following discussions are intended to highlight optimization tools available in CT – every manufacturer can and will differ in the specific implementations of these tools.

C.1.1. Automatic Exposure Control (AEC)

All CT manufacturers employ AEC (AAPM Lexicon; AEC:A scanner feature that automatically adapts the X-ray tube current to the overall patient size to achieve a specified level of image quality) whereby the projections passing through larger attenuating areas of the body intelligently increases the tube current while decreasing the tube current in areas that are less attenuating. This method recognizes that the origin of any noise in the image is dominated by those projections, which suffer high attenuation. Reducing the noise intelligently allows for overall reduction in patient exposure, depending upon patient shape and tissue properties. A perfectly round individual will not produce much tube current modulation, while imaging through the shoulders almost always will benefit from using AEC.

C.1.1.1. Angular Modulation

During the rotation around the patient the X-ray attenuation varies when the tube current intensity must be increased for thicker projections and decreased for thinner X-ray paths depending on patient anatomy. The predicted change is always determined using the CT localizer radiograph but may either be based on a size estimate of the patient, or by using attenuation values, or both.With angular modulation the system modifies the current intensity basing on CT localizer information. For some CT devices two localizers are often needed:both the anterior-posterior and lateral projection. Other systems model patient using one radiograph to estimate an ellipse. Others change the radiation intensities using the information of the 180° previously acquired values of attenuation[1].

C.1.1.2. Longitudinal Modulation

In the case of longitudinal Modulation the system adjusts the current intensity along the Z-axis of the patient for taking into account the variation of attenuation of different anatomic regions. For example, for a scan covering the thoracic region to the pelvis, the current variations can be relevant because air filled lungs attenuate much less than the denser pelvic region (40-60%tube current variation).

C.1.1.3. Angular and Longitudinal Modulation

All CT systems have the ability to combine both Angular and Longitudinal modulation during an exam. Physicists should identify those protocols that do not employ full AEC to appreciate the clinical need and indications of use.

C.1.1.4. AEC in Cardiac applications

AEC is also employed for dose reduction in ECG-gated cardiac CT examination by reducing or totally eliminating the tube current for the cardiac phases which data is not utilized in image reconstruction [2]. Prospective cardiac gaiting enables lower dose than retrospective gating, dose reductions of 20 - 70% are reported [3, 4].

C.1.1.5. Technology of AEC

Manufacturers use different technological approaches to automate tube current modulation in a context of some defined reference image quality (model) as TCM is engaged for their product [5].

  • The GE AEC controls and modulates the tube current by specifying the image quality required in terms of resultant standard deviation (SD) that is related to the tube current by inverse proportion to the square. This parameter is called Noise Index (NI). GE AEC has longitudinal and Patient Size modulation (AutomA) and angularAEC (SmartmA). Changes in mA values are based on aelliptical phantommodel, which arises from the CT localizer radiograph.
  • The Toshiba AEC (Sure Exposure) is based on the standard deviation (SD), i.e. the image noise, taking into account SD values measured in a homogenous water phantom. This system uses the CT localization images to obtain the attenuation information needed to set the tube current for each rotation. Because Toshiba AEC is controlled using image SD, the mA will depend upon the reconstruction kernel selected for primary image reconstruction.
  • The Philips AEC (DoseRight) uses a Reference Image (Case) from selected examinations that the user reviews as of appropriate quality for the particular clinical task. The system matches the image quality (noise) in the actual exam with the operator chosen (previously selected) reference case. Philips AEC Dose Right implements an angular (DOM), Patient Size (ACS) and longitudinal (Z-DOM) modulations.
  • The Siemens AEC (Care 4D) uses Quality Reference mAs to define the effective mAs (mAs/pitch) required to produce a specific image quality in an 80 kg patient for a given protocol. Using the attenuation information of the CT radiograph acquisition (at least one) the system adapts the tube current to the actual patient and adjusts the real current using an online feedback control. Siemens AEC implements angular, longitudinal and patient size modulation.

C.1.2. Automated kVp Selection

It is known that an image Contrast to Noise Ratio (CNR) can be maintained at a lower dose by lowering the tube voltage(kVp). This is particularly true when iodine contrast enhancement medium is used. At a lower kVp, iodine attenuation is increased relative to background due to the energy dependence of the attenuation coefficient of iodine; see Figures C.1 and C.2 for examples. Thus at lower kVp the conspicuity of hyper vascular (or even hypo vascular) pathology may be enhanced. The images obtained are likely of increased perceived noise even though the elevated contrast signal is visualized due to higher attenuation. To balance this effect the tube current may be increased (or de-noising applied).The reduction in dose can be particularly effective for pediatric or small patients.In general, the routine use of 100 kVp (not 120 kVp) is possible with most average size adults. Lower kVp (e.g. 70 or 80 kVp) is useful for pediatric patients and smaller adult patients [6].The use of lower kVp can reduce dose [7] but may be disruptive to workflow when iodine contrast is present in the patient,because preferred window viewing settings or the administered amount of contrast may need to be altered.The task of reducing patient dose by lowering kVp may also complicate the established TCM settings.

Figure C.1.Reduced kVp significantly improves CNR for smaller patients and can optimize image contrast [8].

120 kVp CTDIvol 12 mGy100 kVp CTDIvol 8 mGy

Figure C.2.Same 8-year-old patient imaged 1 year apart. Reduced kVp demonstrates improved enhancement with iodine with reduced patient exposure.

(Used with permission from William Pavlicek, Mayo Clinic, Scottsdale, AZ)

The CT manufacturers have developed tools that automatically suggest the optimal kVp setting for the individual patient and for the specific exam, e.g. Siemens Care kV [8]. The information for this is obtained using the patient CT localizer radiograph and the Planned Protocol scan type (contrast administered or not, soft tissue or bone, etc.). Once chosen, the optimized kVp is held constant, while the mA is modulated as with normal AEC operation. Alternative, simpler solutions can be found in developed console tools that aid the technologist in the selection of AEC parameters and suitable window settings.Optimization by kVp selection is a powerful tool in improving image quality and reducing radiation dose, particularly for examination involving administration iodine contrast.

Reductions of dose with CT Angiography and exams of the liver have been reported(18% and 25%, respectively) by lowering kVp [9].

C.1.3. Adaptive Collimation

Adaptive collimation is a dose savings tool that refers to the automatic introduction of sliding collimation blades to dynamically reduce the longitudinal exposure. The collimator blades can independently open and close at the beginning and at the end of a spiral scan. This feature should be employed, as it will aid in dose reduction as a result of over-scanning. Over-scanning(overranging) refers to the exposure of tissues beyond the boundaries of the volume (Z-axis) imaged, which occurs with multidetector spiralCT scans and results from the need for the interpolation of acquired data during image reconstruction to include sufficient projections. In fact a minimum of an extra half rotation of the X-ray tubebefore and after each scan is necessary for normal image reconstruction. Any over-scanning results in exposure to tissue that will never be part of the image data that are reconstructed.

C.1.4. Image Reconstruction and De-Noising Methods

One of the most important dose reduction tools in CT imaging is reconstruction software that lowers noise. Dose reductions by 50 to 75% are commonly reported in the literature for these commercially available products [10].In practice, de-noising does not by itself lower dose,but noise is reduced in the image thereby enabling optimization of exam protocols to reduce patient dose. Alternatively, de-noising could be motivation for reconstructing thinner images without changing the mAs or kVp and therefore not providing any dose reduction. However,less image noise permits significantly reduced mAs or kVp (or both) because noise is related to dose in quadrature.

C.1.5. Bismuth Shielding

Selective in-plane shielding using bismuth patient shields have been proposed and are commercially available. The shield is made of thin sheet of flexible latex impregnated with bismuth and shaped to cover specific superficial organs including breast, eye and thyroid. They can be used during the examination to reduce dose the specific organ, and organ dose reductions from 26% to 30% have been reported [1, 11]. However, a corresponding noise increase in the tissue just behind the shield is observed,and the presence of the shield mayintroduce image artifacts [12]. If placed over breast tissue, the shield attenuates the X-ray beam and hence decreases anterior organ dose but also attenuates X-rays coming from the posterior direction that have already contributed to organ dose and contain important image information. The AAPM recommendationon bismuth shielding (2012) is for facilities to carefully consider and implement alternatives to bismuth shielding when possible [12].

C.1.6. Dual Energy CT

Dual Energy CT offers information that is not available in single energy scans has been found useful in clinical care of patients. This technology uncovers the energy dependence of tissue to X-ray beams and permits the computation of proportional amounts of material in an image (i.e. iodine and water). Different technical solutions by manufacturer have in common the ability to simultaneously acquire data at different tube voltages for dual energy imaging, allowing the computational analysis of the various materials and tissue components within voxels of a CT image.

C.1.6.1. Dual Source CT
One technical solution currently available consists in a dual source system coupled with two different detector system (2 imaging chains); the system registers at the same time the high and low-energy X-ray (80 – 140 kVp) (Somatom Definition or Flash, Siemens Healthcare) produced by the two X-ray tubes. The advantages are the real simultaneous acquisition especially useful when a contrast medium is present.The acquisition FOV of the second system is smaller and the reconstruction for DECT is limited to this volume. The system improves energy spectral separation via use of a tin filter for one X-ray tube; advantages of this technical solution are that dose optimization can be achieved for each individual patient and AEC can also be employed.

C.1.6.2. Fast kVp Switching
In the General Electric solution a single X-ray source system switches the energy of the beam between 80 and 140 kV in a very short period of time. Normally 1/3 of the projections are obtained at 140 kV, while 2/3 of the projections are obtained with 80kV. The projections measured are nearly synchronous in time with no cross scatter, but dose optimization is hindered because the tube current is fixed for the two X-ray beam energies used.

C.1.6.3. Energy Discriminating Detectors
The Philips Medical Systems solution adopts a two-layer detector in which each layer is responsive to different energies. The advantages of this technology are that each projection has both high and low component with no cross scatter and perfect synchronization of the information. The spectral information is always on and can be used at a later time if the raw projection data is available.

C.1.6.4. Multi-Spin
Toshiba uses a wide detector array configured to perform two rotations in serial at low and high kVp. The mA is automatically adjusted for the two rotations to obtain similar noise for each. This solution necessitates only one kVp switch per DECT acquisition with no cross-scatter. Individual parameter adaptation is permitted but a delay for kVp switching is necessary and as a consequence there can be motion between the two scans.

C.2. Patient Dose Management

In the US, NCRP Report 160 estimates that the average American receives an annual effective dose of 3.0 mSv from medical radiation, half of which now originates from CT. As concluded by Smith-Bindman[13]the radiation doses from commonly performed diagnostic CT examinations are higher and more variable than generally quoted. Having variability in CT radiation dose for a patient undergoing a CT exam is inevitable and in fact absolutely necessary, because radiation dose is highly dependent upon the patient size. Other sources of variability with CT exposure exist and are also normal and appropriate, including the need to repeat an exam due to patient movement, addition of extended anatomy and added delayed series as part of an enhanced understanding of a disease process. However, radiation dose that may not be necessary and can be avoided should be identified and steps taken to reduce the exposure if possible. Examples of this might be due to settings on the scanner, differences between scanners, procedural differences among technologists, patient centering, etc. Management of dose is a team effort to be addressed by the CT staff, a CT protocol committee or with an organized Quality Assurance task group. Having improved estimates of patient dose, especially when coupled with RDSR information using standardized (agreed upon) metrics of patient dose is highly enabling; it permits the application of best practice imaging with radiation dose included as a considered component in the exam protocol. It allows inter-scanner and institutional comparisons of CT use, and is fundamental in supporting efforts to identify and remove unnecessary variability to lower overall radiation use with CT exams. A CT medical physicist is an essential and key member of this team effort by contributing their understanding of the technology of dose reduction, the capture of CT dose metrics and scanner use and the required technical analysis of variability as part of the overall Quality Assurance program for patient dose management.

C.2.1. Oversight and CT Scanning Paradigm

Various national and international governmental agencies and accrediting bodies have developed standards for the safe usage of radiation dose in CT. An increased effort to monitor and to educate users on appropriate use of CT improves safety and quality. CT is an exceptionally powerful and unique diagnostic tool in any clinical setting, but the presumed risks from ionizing radiation must also be considered. The paradigm in all cases would be to obtain an image that satisfies the clinical objective with a minimal amount of radiation dose to the patient.

C.2.2. Protocols and Management of Patient Dose

A protocol is a collection of machine settings and operator instructions that fully describes a CT examination. It specifies image data collection and reconstruction formats, patient positioning, contrast administration and conditional patient care related instructions (age, pregnancy, etc.). Factors affecting patient dose should be explicitly specified, including kVp, use ofAEC, extent of coverage in the FOV, number of irradiation events and anatomy (organ systems) to be imaged. Newly published DICOM Supplement 121 recognizes CT imaging modalities as having two representations; Defined and Performed Protocols.

C.2.2.1. Defined Protocols

A complete Defined Protocol only exists in full by combining information from two totally separate locations – CT scanner settings and parameters from data files stored on the scanner and separately, the facilitytechnologist instructions and textual content that are detailed (along with kVp machine settings, tables, graphics etc.) in a separate document. Defined Protocols are created to meet the wide range of clinical goals (diagnostic indications) as commonly encountered at a location. Consequently, the library of Defined Protocols may be few or large in number depending upon the practice.The AAPM website provides excellent examples of CT Defined Protocols that are offered as examples for common medical indications [14]. The medical physicist can play an important role in review of CT scanner settings for Defined Protocols.

In the US, several ACR publications make note of medical physicist participation with reviewing and advising on protocols. For accredited facilities, the ACR requires the medical physicist to perform an annual review of CT protocols [15]. While a regular review process of all protocols is to be instituted, this document also makes note of at least six protocols that must be reviewed(more if required by state or local regulatory bodies). Separately noted in this documentation is the review of the appropriate use of dose reduction methods.

C.2.2.2. Performed Protocols

Performed Protocols depict exposures to the patient that have actually occurred.Numerous instances of Defined Protocols will be altered as necessary to provide individual patient specific optimal exams as determined by the radiologist. Patient status issues,e.g. renal insufficiency, venous access, obesity, initial findings,or movement may constitute a need for departure from what may have been defined.

Though CTDIvol and DLP as depicted in the RDSR from a Performed Protocol (what was recorded as a result of an actual episode of care) do not represent the absorbed dose to a patient, these readings do offer baseline representation of the actual technical scanning parameters affecting radiation output for that patient’s care. In the US, referring to the ACR Table below (C.1), the CTDIvol for each DefinedProtocol should be listed in the site’s Definedexam protocol library (textual protocol) and must be observable by the technologist at the scanner for a patient exam. A match must exist between the Defined Protocol settings listed in the library and on the CT device. For CT facilities that are ACR accredited, four protocols are reviewed for dose that represent protocols having a single irradiation event, not counting the CT localizerradiographs.The CTDIvol values for these protocols should be near or below the reference values.