Smartma Theory - GE Revolution CT User Manual

consistent image noise in spite of the wide range of patient size. Since image noise variability is
substantially reduced, a significant overall patient dose reduction is possible with proper scan
parameter selection.

2.6.1.2 SmartmA theory

SmartmA, available on all GE multi-slice CT scanners and first introduced on GE single slice
scanners in 1994
many years in conventional X-ray systems. The goal of SmartmA is to make reconstructed
images contain X-ray quantum noise at a level desired by the user, independent of patient size
and/or anatomy. In addition to adjusting for patient attenuation along the z-axis, SmartmA
modulates patient exposure as a function of X-ray tube angle as the tube rotates around the
patient anatomy, leading to concurrent mA modulation in the x, y and z-axes.
SmartmA tube current modulation is determined from several sources of information. The
primary source of information is the overall attenuation and shape of the patient anatomy. Using
the information contained in a single scout scan of the patient scout, the CT system computes
the required mA to be used based on the selected Noise Index setting, the calculated
attenuation of the patient anatomy being scanned, and ASiR-V selection.
SmartmA enables substantial dose reduction by adjusting tube current to minimize X-rays over
angles that have less importance in determining the overall image noise content. In anatomy
that is highly asymmetric, such as the shoulders, X-rays are significantly less attenuated in the
anterior-posterior direction than in the lateral direction. Thus, the overwhelming abundance of X-
rays can be substantially reduced closer to the anterior-posterior direction, without a significant
effect on overall image noise. For each scanned section in the z-axis direction, the system
calculates an mA for the lateral and anterior-posterior patient axes from an estimation of the
patient attenuation through the long and short axes, respectively.
The system determines the tube current using the patient's scout projection data and a set of
empirically determined noise prediction coefficients. The scout projections contain density, size,
and shape information as a function of the patient habitus in the z-direction. The total projection
attenuation (projection area) contains the patient density and size information, while patient
shape information, i.e. oval ratio, an estimate of the eccentricity of the patient habitus as a
function of z, is contained within amplitude and width information derived from the scout
projection data. These patient characteristics determine how much X-ray will reach the detector
for a specified technique and hence predict the image standard deviation due to X-ray noise for
the standard reconstruction algorithm.
To predict the image noise at a given z-position, the projection area and oval ratio are obtained
from the patient's scout. The expected X-ray noise for the reference technique (reference noise)
is then derived as a function of the projection area and oval ratio from the scout using
coefficients determined by a least squares fit of the noise measurements from a set of
phantoms representing a clinical range of patient sizes and shapes.
Knowing the reference noise and the difference between the reference technique and the user
selected prescribed technique, the mA required to obtain the prescribed Noise Index is
calculated using well known X-ray physics relationships. It is known that the noise is inversely
related to the number of photons and the number of photons is related to the slice thickness,
image acquisition time, and mA. In the GE SmartmA design, adjustment factors for helical pitch
are also incorporated in the calculation to account for noise differences that scale between
helical selections and the axial reference technique.
Chapter 11 Scan
, is the CT equivalent of the auto exposure control systems employed for
1,2
Revolution CT User Manual
Direction 5480385-1EN, Revision 1
233