This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nicholson, R Articles by Fetherston, S Search for Related Content PubMed PubMed Citation Articles by Nicholson, R Articles by Fetherston, S

Primary radiation outside the imaged volume of a multislice helical CT scan

Department of Radiology, St Mary's Hospital, London W2 1PG, UK

	Abstract

Top Abstract Introduction Materials and method Results Discussion and conclusion References

 
Multislice helical CT scanning has advantages of speed and X-ray tube loading, making it possible to image larger volumes in a single exposure. Our aim is to investigate dose implications for short scans from the additional X-ray tube rotations required to reconstruct a given volume in helical scanning. To this end a multislice scanner was compared with a single slice scanner. Two independent methods were used. The first was based on optical density measurements taken from a film exposed free-in-air as it moves with the CT bed along the scan axis. The second used measurements from a pencil ionization chamber supported free-in-air at the centre of the CT aperture for the duration of both a long scan and a short scan. This method assumes the same excess primary radiation at the extremes of both scans and the measurements are incorporated into two simultaneous equations. The dose–length product outside the imaged volume has been compared with the dose–length product inside the imaged volume using both methods. For 4 x 5 mm multislice collimation with a 360° interpolation and a pitch of 0.875, the film and simultaneous equations methods show an excess dose–length product at the extremes of the scan volume equivalent to 3.3 cm and 3.5 cm extra scan length, respectively. This represents a large percentage of a short scan and is substantially greater than for a helical scan using the single slice scanner with 5 mm collimation, a 360° interpolation and a pitch of 1. The latter showed an excess dose–length product at the extremes which was equivalent to 0.35 cm scan length by the film method and 0.25 cm using simultaneous equations. Taking the abdominal protocols recommended by the respective manufacturers, however, the multislice scanner could cover a 45 cm scan length in a single exposure, while the single slice scanner needed six exposures to image the same volume. With the multislice scanner set at 4 x 2.5 mm collimation, 360° interpolation and a pitch of 0.875, the dose–length product outside the volume of interest was equivalent to 1.9 cm scan length by the first method and 1.8 cm using the second method. With 4 x 1 mm collimation it was equivalent to 1.0 cm using both methods. Changing the interpolation algorithm from 360° to 180° had no effect on the additional equivalent scan length while doubling the pitch resulted in a 25% increase. We conclude from this study that with the multislice scanner, the axial mode is to be preferred for short CT scans such as those used in patient biopsies and drainage. For paediatric helical scans shorter than 13 cm, dose length product is reduced by using 4 x 2.5 mm rather than 4 x 5 mm collimation. For longer scans, however, the increased CT dose index associated with narrower collimation in the multislice mode offsets the dose reduction at the extremes.

	Introduction

Top Abstract Introduction Materials and method Results Discussion and conclusion References

 
Multislice CT scanners use a fan of X-rays, broadened in the direction of the scanner axis, to allow coverage of a number of parallel rows of solid state detectors. These can be combined to produce a detection system for four simultaneous axial CT images up to 8 mm wide [1–3]. This makes it possible to image a given volume in one quarter of the time using one quarter of the tube loading of a single slice scanner. To acquire sufficient data for helical reconstruction, however, additional rotations beyond the imaged volume are required and the increased table movement per rotation in multislice as opposed to single slice scanners leads to an increased dose–length product (DLP) outside the imaged volume relative to dose within the imaged volume. Only the X-rays passing through the volume of interest carry direct diagnostic information, but towards the edges of the volume these data must be combined with data from X-rays passing outside the volume if the image is to be reconstructed and the diagnostic information displayed. Z-axis interpolation is required for image reconstruction in all helical scanning, and in multislice scanners z-axis filter interpolation is used to obtain smooth transitions between detectors in the z direction [4]. This may require the tube to rotate through a larger additional angle at the extremes of the scan. Since CT contributes a high proportion of the total diagnostic radiology collective dose in the UK (4500 man Sv per year compared with 16 000 man Sv per year from conventional diagnostic radiology) [5], it is particularly important to minimize any redundant radiation.

Two methods have been used to estimate the additional equivalent scan length outside the volume of interest. The first, a film method, is straightforward if the CT scanner delivers a constant output, as is the case with the scanners used in this study. Some scanners, however, modulate the X-ray beam to compensate for differences in patient thickness. The second method compares the cumulative in-air ionization chamber (IC) readings from a short and a long scan when the chamber is fixed at the centre of the gantry aperture.

	Materials and method

Top Abstract Introduction Materials and method Results Discussion and conclusion References

 
Film method
With headrest removed, a Kodak X-Omat V non-screen therapy verification film (Eastman, Kodak, Rochester, NY) was supported from the head end of the CT imaging table so that it projected beyond the table towards the CT gantry. The film was fixed with a small amount of lateral curvature to keep it horizontal, and aligned with the axis of the scanner (Figure 1). It was then exposed free-in-air to a helical scan of approximately 5 cm length, as the film moves ahead of the table through the scanner. The absence of scattering material within the aperture ensures that the blackening of the film is almost entirely owing to primary radiation.

View larger version (131K): [in this window] [in a new window]   Figure 1. Kodak X-Omat V therapy verification film positioned within the aperture of the CT gantry.

A film was taken for the single slice scanner (PQ-CT; Philips Medical Systems, Best, The Netherlands) using 5 mm collimation, 120 kV, 200 mAs per slice, pitch 1 and 360° interpolation.

Optical density measurements of the film were taken at 5 mm intervals (or less) along the scan axis, using a 1 mm diameter densitometer aperture. Care must be taken to ensure that the frequency of sampling does not synchronize with peaks or troughs from any modulation in optical density that might occur within the reconstruction volume. While a microdensitometer is preferable, if it is available, a 1 mm aperture should not contribute more than 10% to the overall error for any of the scans we investigated and the error will be much less than this for scans with larger overshoots. Optical density readings were converted to percentage air kerma values using a calibration graph taken from a film similarly exposed to a series of axial slices (5 cm apart) with the same kV but differing mAs values. Pencil IC measurements at the isocentre were used to confirm that output is indeed proportional to mAs for the scanner.

Graphs were plotted of percentage air kerma against distance along the scan axis. The areas under the graph, inside and outside the imaged volume, were taken as measures of DLP from the primary radiation. As expected, the change in dose was approximately linear at the extremes of the scan, so areas were calculated by breaking the graph into combinations of right-angled triangles and rectangles. The total area beneath the graph, outside the imaged volume, is denoted by DLP_E. The area inside the imaged volume is denoted by DLP_V, and the reconstruction length is L. The amount of additional radiation from the excess region can be calculated in terms of L_equiv. This is the length of imaged region that would give the same free-air DLP as the excess region. L_equiv is then given by

Simultaneous equations method
A 10 cm pencil IC (10 x 5-10.3CT; Radcal Corporation, Monrovia, CA) was supported from the floor at the centre of the aperture of the scanner. Free-in-air measurements were made for scans of approximately 5 cm length and were repeated using the longest available scan length. The longer scan length was at least three times that of the shorter, and the scan length was confirmed by checking the number of images reconstructed when the reconstruction increment is equal to the slice width. Measurements were made on the multislice scanner using 44.8 cm and 5.2 cm scan lengths, with 4 x 5 mm collimations at 50 mAs per slice, a 360° interpolation and a pitch of 0.875. The simultaneous equation method was similarly used to calculate L_equiv on the multislice scanner for 4 x 2.5 mm and 4 x 1 mm collimations, followed by the 4 x 5 mm collimation with a 180° interpolation then a change in pitch to 1.75. Measurements were taken on the single slice scanner using 5 mm collimation, 360° interpolation and a pitch of 1.

The low mAs values allow repeated readings to be taken without undue tube loading. In each case the table was allowed to glide about 15 cm below the IC, intercepting the beam throughout. Readings were repeated three times (or until the standard deviation of the mean was less than 1%).

An equation was constructed for each scan length using the unknown constant R_E to represent the contribution to the IC reading from outside the imaged volume, and R_V(L) to represent the contribution to the IC reading from within an imaged volume of length L. The total IC reading for a scan of length L is denoted by R_TOT(L) so that:

For scans with the same exposure settings but differing lengths R_v(L)/L is constant.

From the definition

Measurements of R_TOT(L) from a long and a short scan, of length L₁ and L₂, respectively, can be combined to give

	Results

Top Abstract Introduction Materials and method Results Discussion and conclusion References

 
Film method
Either side of the scan axis the films showed z-axis modulation, as the distance of the X-ray source from these positions on the film varied during X-ray tube rotation. These sections of the film were used to confirm the distance moved by the table during a single rotation. Along the scan axis the optical density of the film within the imaged volume remained uniform, since the X-ray focus remained at a fixed distance throughout the rotation and there was no absorbing material within the gantry aperture to modulate the signal (Figure 2). At the start of the scan, before the volume of interest was reached, dose increased approximately linearly until the trailing edge of the beam reached the initial position of the leading edge and the dose reached a plateau. The reverse effect was seen at the end of the scan (Figure 3).

View larger version (89K): [in this window] [in a new window]   Figure 2. The pattern of film blackening from a helical scan of 5.2 cm length using the multislice scanner with 4 x 5 mm data acquisition (a) is contrasted with the pattern from a helical scan of 5.0 cm length and 5 mm data acquisition using the single slice scanner (b). The unmodulated region of each film lies along the CT scan axis.

View larger version (21K): [in this window] [in a new window]   Figure 3. (a) Multislice scanner and (b) single slice scanner show the primary radiation energy deposited per unit length along the long axis of the patient as a percentage of the average primary radiation energy deposited per unit length within the reconstructed volume. Values are taken from optical density measurements along the scan axes of the films shown in Figure 1 for the multislice and single slice scanners, respectively.

The results for both methods are shown in Tables 1 and 2. The tables show the scan length L_equiv that would give a free-air DLP equivalent to the total free-air DLP outside the imaged volume at a dose equal to that within the imaged volume. Table 1 compares the multislice scanner with the single slice scanner. Table 2 explores the way in which L_equiv at the extremes varies with collimation, interpolation algorithm and pitch in the multislice scanner.

	Discussion and conclusion

Top Abstract Introduction Materials and method Results Discussion and conclusion References

Since the excess dose at the extremes of the scan is independent of the chosen scan length it constitutes a higher proportion of the total DLP of a short scan than a long scan. This has serious implications for helical scans of short length such as those used for patient biopsies and drainage under CT guidance. It also has implications for paediatric examinations where fewer slices are required due to patient size.

For a given pitch, table movement during tube revolution is proportionately smaller for narrower collimations, and this results in less overshoot at the extremes. If the Mx8000 collimation is reduced from 4 x 5 mm to 4 x 2.5 mm, keeping pitch constant, L_equiv of the overshoot is reduced by 45%. Reduction in collimation from 4 x 5 mm to 4 x 1 mm reduces L_equiv by 70%. The fact that L_equiv is not reduced in proportion to the collimation implies that a greater excess rotation angle is required for narrower collimations with the z-axis filter interpolation used. Doubling the pitch increases L_equiv by only 25%, suggesting that the larger pitch requires a smaller excess rotation angle while changing the interpolation algorithm from 360° to 180° has no effect on the overshoot. These overshoots are fixed by the manufacturers.

Speed and tube loading are less important for short scans, and on the Mx8000 scanner there is the option of reducing the collimation from 4 x 5 mm to 4 x 2.5 mm to give the same imaged slice width. This gives a 45% reduction in L_equiv. The reduction in DLP_E, however, is to some extent offset by a 10% increase in CT dose index (CTDI) for the narrower collimation. This is because the X-ray beam penumbra must lie outside the active detector array in multislice scanning so that an unattenuated beam irradiates the detectors equally in the z direction. Since the penumbra remains the same for the different collimations, it constitutes a higher proportion of the dose profile with the narrower collimations. For a scan length of 13 cm the saving in dose from reduced overshoot becomes equal to the increase in dose from increased CTDI in the Mx8000 scanner, so overall dose reduction is only possible for scans shorter than 13 cm. Alternatively, axial scans may be considered.

The issue of primary dose outside the imaged volume does not necessarily imply that the effective dose for a multislice helical CT examination is more than for a similar single slice helical CT scan with the same image quality. It does, however, highlight additional dose that is particularly pronounced for short scans and an awareness of this could lead to an optimization of protocols to give the lowest dose. The necessary excess X-ray tube rotation depends on the form of the z-axis filter interpolation and our results show that a doubling of pitch does not double the value of L_equiv as expected, and that halving the interpolation angle produces no reduction in L_equiv. This suggests there may be scope for further development of the z-axis filter interpolation to reduce dose from the overshoot without compromising image quality. Another way forward might be to provide the option of collimating the beam and combining detectors in the multislice array to produce the equivalent of a single slice scanner for short helical scans, combining the advantages of multislice and single slice in the same scanner.

	Acknowledgments

 
We would like to thank Jim Weir from the Radiology Department at Queen Mary's, Roehampton for his willing help and cooperation in this project.

Received for publication March 5, 2001. Accepted for publication December 20, 2001.

	References

Top Abstract Introduction Materials and method Results Discussion and conclusion References

 

1. Medical Devices Agency. ImPACT technology update No. 1 Multi-slice CT scanners, MDA/00/23. London: MDA, 2000.
2. Kalender WA. Computed tomography fundamentals, system technology, image quality application. Munich, Germany: Publicis MCD Verlag, 2000:14–131.
3. Kopecky KK, Buckwalter KA, Sokiranski R. Multi-slice CT spirals past single-slice CT in diagnostic efficacy. Diagn Imaging (San Franc) 1999;21:36–42.
4. Taguchi K, Aradate H. Algorithm for image reconstruction in multi-slice helical CT. Med Phys 1998;25:550–61.[Medline]
5. Hughes JS, editor. Ionising radiation exposure of the UK population: 1999 review. NRPB-R311. Chilton, UK: National Radiological Protection Board, 1999.

This article has been cited by other articles:

				K. Rothkamm, S. Balroop, J. Shekhdar, P. Fernie, and V. Goh Leukocyte DNA Damage after Multi-Detector Row CT: A Quantitative Biomarker of Low-Level Radiation Exposure Radiology, January 1, 2007; 242(1): 244 - 251. [Abstract] [Full Text] [PDF]

				A. J. van der Molen and J. Geleijns Overranging in Multisection CT: Quantification and Relative Contribution to Dose--Comparison of Four 16-Section CT Scanners Radiology, December 1, 2006; 242(1): 208 - 216. [Abstract] [Full Text] [PDF]

				M A Lewis and S Edyvean Patient dose reduction in CT Br. J. Radiol., October 1, 2005; 78(934): 880 - 883. [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nicholson, R Articles by Fetherston, S Search for Related Content PubMed PubMed Citation Articles by Nicholson, R Articles by Fetherston, S