1992, The British Journal of Radiology, 65, 261-263
function taken from ICRU Report 40 (1986) can be folded into the measured dose distribution for 25 kVp shown in Fig. 3 and compared with the relative risk factor reported by Brenner and Amols (1989) for 23 kVp. If this is done a value of 1.18 is obtained, which agrees within 10% with the figure of 1.29 quoted by Brenner and Amols (1989). Acknowledgments The assistance of Mr Stuart McArdle of the FAXIL group in the Department of Medical Physics in Leeds is gratefully acknowledged. This work has been supported by the British Council under the Acciones Integradas scheme.
EIVAZI, M. T., 1990. Application of experimental microdosimetric techniques to the dosimetry of X-rays in the diagnostic energy range. PhD Thesis, University of Leeds. ICRU, 1986. The Quality Factor in Radiation Protection, Report 40 (International Commission on Radiation Units and Measurements, Bethesda). Rossi, H. H., 1984. Development of microdosimetric counters past present and future. Radiation Protection Dosimetry, 9, 161-168. WAKER, A. J., 1985. Experimental uncertainties in microdosimetric measurements and an examination of the performance of three commercially produced proportional counters. Nuclear Instruments and Methods in Physics Research, A234, 354-360. WAKER, A. J. & EIVAZI, M. T., 1989. The application of
References BRENNER, D. J. & AMOLS, H. I., 1989. Enhanced risk from low energy screen-film mammography X-rays. British Journal of Radiology, 62, 910-914. DAHMEN,
P.,
MENZEL,
H.
G.
& GRILLMAIER,
R.,
1989.
Dosimetry of photons with low pressure proportional counters. Radiation Protection Dosimetry, 29, 75-79.
microdosimetric measurements to low energy Radiation Protection Dosimetry, 29, 81-85. WAKER,
A. J. & MAYNARD, D. G.,
X-rays.
1989. The effect
of
geometrical scaling on the gas gain of proportional counters intended for microdosimetric measurements. Radiation Protection Dosimetry, 29, 37-40.
Accuracy of pelvimetry measurements on CT scanners By J . P. Wade, BSc, MSc East Anglian Regional Radiation Protection Service, Addenbrooke's Hospital, Hills Road, Cambridge {Received 17 May 1991 and in revised form 25 July 1991, accepted 4 September 1991) Keywords: Pelvimetry, Radiography digital, Pelvis measurement
There is an increasing awareness of radiation dose received by the patient and the use of low dose techniques is encouraged whenever possible. A low dose technique which is available is the use of the overview or scanogram facility on a computed tomography (CT) scanner to produce a digital radiograph for pelvimetry measurements instead of the conventional radiograph. Unfortunately ultrasound, which is the low dose technique, cannot be used for pelvimetry measurements because of the bone geometry. Several published papers discuss the dose reduction (of the order of 90%) which is received by the mother and the fetus by changing to CT pelvimetry examinations (Federle et al, 1982; Suramo et al, 1984; Dobson & Nelson, 1988), and the recent NRPB Report (1990) emphasized this dose reduction. As many district hospitals now have CT scanners, this technique is becoming more widely practised. Whilst dose reduction is a major consideration when choosing a new technique, it is also important that the results obtained are as good or better than conventional techniques, since any inaccuracies may have serious repercussions and lead to complications during labour. For pelvimetries carried out on CT scanners, a lateral Vol. 65, No. 771
view or an antero-posterior (AP) scanogram, or both, is obtained of the pelvis and the required distances are measured on the viewing console using the distance measurement facilites provided in the software. A simple test is required to assess the accuracy of these measurements, which can be incorporated into the quality control surveys. Technique
A uniform grid is an appropriate choice of object so that measurements both parallel and perpendicular to the direction of travel can be carried out. In the set of test objects produced by FAXIL (the Radiological Imaging Group, Department of Medical Physics, General Infirmary, Leeds), for the assessment of fluoroscopy equipment, there is a 26 cm diameter disc shaped object containing a regular grid of 2 cm pitch which is ideal for this purpose. The grid test object is placed in the scanning tunnel perpendicular to the direction the X-ray tube is pointing. When the X-ray tube is at the top as for an AP view then the test object should be lying horizontally, as shown in Fig. 1. The alignment lights are used to aid location of the plane where the axis of rotation occurs. A scanogram is performed with 261
Technical notes
TEST PLATFORM SUPPORTED FLAT ON COUCH - COUCH MOVEMENT BEING PERPENDICULAR TO THE PLANE OF THE PAPER
Figure 1. Diagram showing the set up of the grid phantom on the CT scanner.
the object lying in this plane, and further views can be carried out after moving the grid test object towards or away from the X-ray tube. Distances can then be measured on the scanogram images, using the software facilities provided. If a number of scanograms are taken, at different focus-object distances, then the variation of magnification with distance from the midline can be obtained. Results
This grid test object has now been used on five CT scanners (three different manufacturers). It was found that all measurements made on the image parallel to the direction of movement of the couch were accurate, but there were variations in the distances measured in the perpendicular direction. Images of the results from one
Figure 2. Scanogram of the test object positioned in the plane of the centre of rotation.
262
Figure 3. Scanogram of the test object moved 50 mm towards the X-ray tube.
scanner are shown in Figs 2, 3 and 4. For all three images the direction of movement is in the vertical direction of the image. The first image (Fig. 2) is in the plane of rotation; the measurements in both directions are equal. The image shown in Fig. 3 shows the test object moved 5 cm towards the X-ray tube; measurements in the x direction have been magnified while those in the y direction are the same as in Fig. 1. The image shown in Fig. 4 shows the test object moved 5 cm away from the X-ray tube; measurements in the x direction are minified whilst those in the y direction are the same as in Fig. 1. The results in Table I show how the measurement of a dimension on the phantom varies with the position of the phantom for several CT scanners. The images shown in Figs 2, 3 and 4 were obtained on scanner A; the results for three other scanners are also shown. With all the scanners, measurements taken on the centre plane were found to be within 5 % of the expected value.
Figure 4. Scanogram of the test object moved 50 mm away from the X-ray tube. The British Journal of Radiology, March 1992
Technical notes Table I. Results of measurements taken with the test phantom at different distances from the plane of rotation Distance from central plane (mm)
Measurement across 7 x 2 cm squares (mm) i.e. 140 mm
Scanner A Towards Tube—85 —80 —75 —50 —20 Centre Away from Tube—25 —50 75 —100 —115
169.60 167.28 165.71 155.00 145.49 140.01 133.83 128.01 122.17 116.73 113.62
Scanner B Towards Tube—50 Centre Away from Tube—50
145.4 134.31 127.8
Scanner C Towards Tube—50 Centre Away from Tube—50
144.4 136.5 127.8
Scanner D Towards Tube—50 Centre Away from Tube—50
144.9 135.2 125.4
Reference to figures
Figure 3, A-B Figure 2, E - F Figure 4, A-B
then the plane where measurements are required should be set to within 10 mm. In practice, it has been found that the pelvic inlet and pelvic outlet measurements taken on the lateral view are the most important dimensions and as these lie centrally in the body it is important to line up the patient carefully. On the anterior scout view, the transverse diameters of the pelvis are measured, but these will be at varying depths in the patient and so require different correction factors owing to the varying magnification. These measurements are not widely made because of the problems of defining the depth of these points in the patient (Lotz et al, 1987) and so using the right magnification. It is important to verify the accuracy of distance measurements done on the scanogram where a CT scanner is used for pelvimetry as the accuracy appears to vary between models. There is very little information available from the manufacturer regarding the accuracy of measurements taken on the scanogram. This test has proved to be simple to perform and the results easy to evaluate and it should be incorporated into regular quality control procedures. Acknowledgments I would like to thank Mrs K. E. Goldstone for her encouragement and to the radiographers at the CT scanners, at various centres around East Anglia, for their help in doing these measurements. References DOBSON, J. & NELSON, J., 1988. CT pelvimetry: replacing
conventional with digital. Radiography, 54 (613), 18-19. FEDERLE, M. P., COHEN, H. A., ROSENWEIN, M. F., BRAUT-ZAWADZKI, M. N. & CANN, L. E., 1982. Pelvimetry
Discussion
As the use of scanograms is being encouraged for pelvimetry measurements, it is important for the user to know the measurements are accurate and the limitations of the scanner in use. With conventional radiographic techniques the pelvis measurements are normally quoted to the nearest 5 mm to the radiologist or obstetrician. For this accuracy to be achieved in CT pelvimetries without introducing an error caused by magnification,
Vol. 65, No. 771
by digital radiography: a low-dose examination. Radiology, 143, 733-735. LOTZ, H., EKELUND, L., HIETALA, S. O., ERIKSSON, L., WICKLUND, D. E. & WICKMAN, G., 1987. Low dose
pelvimetry with biplane digital radiography. Ada Radiologica, 28, 577-580. NRPB, 1990. Patient Dose Reduction in Diagnostic Radiology. Documents of the NRPB 1(3) (HMSO, London). SURAMO, I., TORNIAINEN, P . , JOUPPILA, P . , KlRKINEN, P . &
LAHDE, S., 1984. A low-dose CT-pelvimetry. British Journal of Radiology, 57, 35-37.
263
function taken from ICRU Report 40 (1986) can be folded into the measured dose distribution for 25 kVp shown in Fig. 3 and compared with the relative risk factor reported by Brenner and Amols (1989) for 23 kVp. If this is done a value of 1.18 is obtained, which agrees within 10% with the figure of 1.29 quoted by Brenner and Amols (1989). Acknowledgments The assistance of Mr Stuart McArdle of the FAXIL group in the Department of Medical Physics in Leeds is gratefully acknowledged. This work has been supported by the British Council under the Acciones Integradas scheme.
EIVAZI, M. T., 1990. Application of experimental microdosimetric techniques to the dosimetry of X-rays in the diagnostic energy range. PhD Thesis, University of Leeds. ICRU, 1986. The Quality Factor in Radiation Protection, Report 40 (International Commission on Radiation Units and Measurements, Bethesda). Rossi, H. H., 1984. Development of microdosimetric counters past present and future. Radiation Protection Dosimetry, 9, 161-168. WAKER, A. J., 1985. Experimental uncertainties in microdosimetric measurements and an examination of the performance of three commercially produced proportional counters. Nuclear Instruments and Methods in Physics Research, A234, 354-360. WAKER, A. J. & EIVAZI, M. T., 1989. The application of
References BRENNER, D. J. & AMOLS, H. I., 1989. Enhanced risk from low energy screen-film mammography X-rays. British Journal of Radiology, 62, 910-914. DAHMEN,
P.,
MENZEL,
H.
G.
& GRILLMAIER,
R.,
1989.
Dosimetry of photons with low pressure proportional counters. Radiation Protection Dosimetry, 29, 75-79.
microdosimetric measurements to low energy Radiation Protection Dosimetry, 29, 81-85. WAKER,
A. J. & MAYNARD, D. G.,
X-rays.
1989. The effect
of
geometrical scaling on the gas gain of proportional counters intended for microdosimetric measurements. Radiation Protection Dosimetry, 29, 37-40.
Accuracy of pelvimetry measurements on CT scanners By J . P. Wade, BSc, MSc East Anglian Regional Radiation Protection Service, Addenbrooke's Hospital, Hills Road, Cambridge {Received 17 May 1991 and in revised form 25 July 1991, accepted 4 September 1991) Keywords: Pelvimetry, Radiography digital, Pelvis measurement
There is an increasing awareness of radiation dose received by the patient and the use of low dose techniques is encouraged whenever possible. A low dose technique which is available is the use of the overview or scanogram facility on a computed tomography (CT) scanner to produce a digital radiograph for pelvimetry measurements instead of the conventional radiograph. Unfortunately ultrasound, which is the low dose technique, cannot be used for pelvimetry measurements because of the bone geometry. Several published papers discuss the dose reduction (of the order of 90%) which is received by the mother and the fetus by changing to CT pelvimetry examinations (Federle et al, 1982; Suramo et al, 1984; Dobson & Nelson, 1988), and the recent NRPB Report (1990) emphasized this dose reduction. As many district hospitals now have CT scanners, this technique is becoming more widely practised. Whilst dose reduction is a major consideration when choosing a new technique, it is also important that the results obtained are as good or better than conventional techniques, since any inaccuracies may have serious repercussions and lead to complications during labour. For pelvimetries carried out on CT scanners, a lateral Vol. 65, No. 771
view or an antero-posterior (AP) scanogram, or both, is obtained of the pelvis and the required distances are measured on the viewing console using the distance measurement facilites provided in the software. A simple test is required to assess the accuracy of these measurements, which can be incorporated into the quality control surveys. Technique
A uniform grid is an appropriate choice of object so that measurements both parallel and perpendicular to the direction of travel can be carried out. In the set of test objects produced by FAXIL (the Radiological Imaging Group, Department of Medical Physics, General Infirmary, Leeds), for the assessment of fluoroscopy equipment, there is a 26 cm diameter disc shaped object containing a regular grid of 2 cm pitch which is ideal for this purpose. The grid test object is placed in the scanning tunnel perpendicular to the direction the X-ray tube is pointing. When the X-ray tube is at the top as for an AP view then the test object should be lying horizontally, as shown in Fig. 1. The alignment lights are used to aid location of the plane where the axis of rotation occurs. A scanogram is performed with 261
Technical notes
TEST PLATFORM SUPPORTED FLAT ON COUCH - COUCH MOVEMENT BEING PERPENDICULAR TO THE PLANE OF THE PAPER
Figure 1. Diagram showing the set up of the grid phantom on the CT scanner.
the object lying in this plane, and further views can be carried out after moving the grid test object towards or away from the X-ray tube. Distances can then be measured on the scanogram images, using the software facilities provided. If a number of scanograms are taken, at different focus-object distances, then the variation of magnification with distance from the midline can be obtained. Results
This grid test object has now been used on five CT scanners (three different manufacturers). It was found that all measurements made on the image parallel to the direction of movement of the couch were accurate, but there were variations in the distances measured in the perpendicular direction. Images of the results from one
Figure 2. Scanogram of the test object positioned in the plane of the centre of rotation.
262
Figure 3. Scanogram of the test object moved 50 mm towards the X-ray tube.
scanner are shown in Figs 2, 3 and 4. For all three images the direction of movement is in the vertical direction of the image. The first image (Fig. 2) is in the plane of rotation; the measurements in both directions are equal. The image shown in Fig. 3 shows the test object moved 5 cm towards the X-ray tube; measurements in the x direction have been magnified while those in the y direction are the same as in Fig. 1. The image shown in Fig. 4 shows the test object moved 5 cm away from the X-ray tube; measurements in the x direction are minified whilst those in the y direction are the same as in Fig. 1. The results in Table I show how the measurement of a dimension on the phantom varies with the position of the phantom for several CT scanners. The images shown in Figs 2, 3 and 4 were obtained on scanner A; the results for three other scanners are also shown. With all the scanners, measurements taken on the centre plane were found to be within 5 % of the expected value.
Figure 4. Scanogram of the test object moved 50 mm away from the X-ray tube. The British Journal of Radiology, March 1992
Technical notes Table I. Results of measurements taken with the test phantom at different distances from the plane of rotation Distance from central plane (mm)
Measurement across 7 x 2 cm squares (mm) i.e. 140 mm
Scanner A Towards Tube—85 —80 —75 —50 —20 Centre Away from Tube—25 —50 75 —100 —115
169.60 167.28 165.71 155.00 145.49 140.01 133.83 128.01 122.17 116.73 113.62
Scanner B Towards Tube—50 Centre Away from Tube—50
145.4 134.31 127.8
Scanner C Towards Tube—50 Centre Away from Tube—50
144.4 136.5 127.8
Scanner D Towards Tube—50 Centre Away from Tube—50
144.9 135.2 125.4
Reference to figures
Figure 3, A-B Figure 2, E - F Figure 4, A-B
then the plane where measurements are required should be set to within 10 mm. In practice, it has been found that the pelvic inlet and pelvic outlet measurements taken on the lateral view are the most important dimensions and as these lie centrally in the body it is important to line up the patient carefully. On the anterior scout view, the transverse diameters of the pelvis are measured, but these will be at varying depths in the patient and so require different correction factors owing to the varying magnification. These measurements are not widely made because of the problems of defining the depth of these points in the patient (Lotz et al, 1987) and so using the right magnification. It is important to verify the accuracy of distance measurements done on the scanogram where a CT scanner is used for pelvimetry as the accuracy appears to vary between models. There is very little information available from the manufacturer regarding the accuracy of measurements taken on the scanogram. This test has proved to be simple to perform and the results easy to evaluate and it should be incorporated into regular quality control procedures. Acknowledgments I would like to thank Mrs K. E. Goldstone for her encouragement and to the radiographers at the CT scanners, at various centres around East Anglia, for their help in doing these measurements. References DOBSON, J. & NELSON, J., 1988. CT pelvimetry: replacing
conventional with digital. Radiography, 54 (613), 18-19. FEDERLE, M. P., COHEN, H. A., ROSENWEIN, M. F., BRAUT-ZAWADZKI, M. N. & CANN, L. E., 1982. Pelvimetry
Discussion
As the use of scanograms is being encouraged for pelvimetry measurements, it is important for the user to know the measurements are accurate and the limitations of the scanner in use. With conventional radiographic techniques the pelvis measurements are normally quoted to the nearest 5 mm to the radiologist or obstetrician. For this accuracy to be achieved in CT pelvimetries without introducing an error caused by magnification,
Vol. 65, No. 771
by digital radiography: a low-dose examination. Radiology, 143, 733-735. LOTZ, H., EKELUND, L., HIETALA, S. O., ERIKSSON, L., WICKLUND, D. E. & WICKMAN, G., 1987. Low dose
pelvimetry with biplane digital radiography. Ada Radiologica, 28, 577-580. NRPB, 1990. Patient Dose Reduction in Diagnostic Radiology. Documents of the NRPB 1(3) (HMSO, London). SURAMO, I., TORNIAINEN, P . , JOUPPILA, P . , KlRKINEN, P . &
LAHDE, S., 1984. A low-dose CT-pelvimetry. British Journal of Radiology, 57, 35-37.
263
