2013 AAPM Annual Meeting
Donovan Bakalyar, PhD, Henry Ford Health System, Detroit, MI, 48202 UNITED STATES
The radiation delivered by a CT scanner is determined using a variety of parameters under the control of the operator. The multiplicity of terms used to describe these parameters can add confusion to what should be a simple task: characterizing CT radiation output in terms of the dose delivered to a cylinder.
In order to keep things as simple as possible, we start, as in TG111, with an infinite scan of an infinitely long cylindrical phantom. During such a scan the dose at any point in the cylinder will asymptotically build up to its ultimate value Deq. The spatial average of this value Deq_ave, can serve as an indicator of scanner output and along with the total number of rotations N determine Etot, the total energy absorbed by the cylinder. If a finite cylinder is long enough, a scan through it from one end to the other will yield these values in its central plane.
The asymptotic approach to Deq can conveniently be described by h(L), a robust function of the scan length (L) which depends only weakly on z axis collimation and tube potential. This function facilitates the simple calculation of the lower dose to the central plane resulting from a finite scan. If the length of the cylinder is finite, as it is for the standard CTDI phantom, the dose is reduced further in a predictable fashion resulting in good correlation between Deq_ave, and CTDIvol which in turn has been shown to be a useful tool in equating the radiation output between scanners differing in model and vendor. The extension of these considerations to stationary table configurations such as wide beam and interventional CT will also be shown.
Though CTDIvol (and Deq_ave) are useful for characterizing CT machine output, we need to include the effect of patient size in assessing patient dose. TG204 has used measurements on phantoms of varying size to determine a size specific dose estimate (SSDE) which serves to accomplish this task. The SSDE has also proven to be very helpful in designing protocols for patients of varying size, especially pediatric patients.
The Role of Physics in CT Protocol Optimization Over the Range of CT Scanner Types: Recommendations and Misconceptions - Frank N. Ranallo, Ph.D.
The understanding of how scan and reconstructions parameters affect image quality, patient dose, and total scan time is essential to the proper optimization of CT scanner protocols. This understanding is greatly complicated by the differences in user interfaces and in the effects of varying scan and reconstruction parameter with scanners from different manufacturers and even with different scanners from the same manufacturer. Certain scan parameters affect patient dose and image quality (kV, mA, rotation time, effective mAs, pitch, noise index (NI), standard deviation, target effective (TE) mAs), while a subset of these affect the total scan time - which is important for breath hold and contrast considerations. There is a logical method of approaching the modifications of these parameters to achieve the necessary diagnostic image quality at the lowest dose, within total scan time limitations. This method must take into account the model and manufacture of CT scanner, but unfortunately has often not been followed in the development of published CT protocols.
Some examples of common confusions and misconceptions involving protocol optimization:
The automatic exposure control (AEC) systems from various manufacturers have dramatically different interfaces and perform their functions very differently. Even with the use of AEC, optimal CT scans are not usually obtained with a single protocol for all size patients. In AEC mode the dose can be reduced by raising the NI or lowering the TE mAs. However raising the kV can either raise or lower the dose depending on the AEC system used. Imaging infants to large adults can utilize a range of kV from 80 to 140 kV, with large patients requiring a higher NI and lower TE mAs when using AEC.
The effects of pitch and the ways to adjust pitch to optimize image quality and patient dose have appeared to be particularly confusing. Raising the pitch above a value of 1.0 is usually not an optimal way to lower dose due to detrimental effects on image artifacts and slice thickness. Instead, the use of a pitch less than 1.0, with shorter rotation times and lower mA is preferred. Pitch values less than 1.0 do not unnecessarily over-irradiate the patient since the added radiation is effectively utilized in the image reconstruction in reducing image noise. Raising the pitch above 1.0 should be a final step used to reduce the total scan time when it is clinically too long.
