0 Localization,
Inhomogeneities
and Treatment
COMPUTERIZED-TOMOGRAPHY:
Planning for Particle
Therapy
ISOTOPE SOURCES
DOUGLAS P. BOYD, Ph.D. University of California, San Francisco, Department of Radiology, School of Medicine, San Francisco, CA 94143, U.S.A. Recently pure rotary fan beam computerized tomographic (CT) body summers have been suuxssfuUy demonstrated by Artronfcs, AS&E, GE and Varian. These scanners have the large solid angle detection efUciency required by proposed practical isotope source-based CT scanners. Current scanners employ broad-energy bremsstrahlung sources at 10&140 kVp. The nonlinearity of polychromatk X-ray attenuation introduces signfficant absolute accuracy errors in the CT image even wben sophistkated software corrections are employed. Accuracy is important in heavy ion treatment planning and quantitative physWo&al measurements. A 100 Ci ‘-Gd source would enable accurate, industry competitive 20~ scans, but would requin a specfh activit&3-10 times tbe current state of tbe art.’ The dual monochromatic energy emission at “43” keV and “100” keV of ’ Gd can be utbed to sfmultaneously recomstruct electron density and atomic number CT images. These techniques may be useful for the development of a low-cost dedicatedCT scanner for therapy planning. Isotope sources,
Computer
assisted tomography
(CAT) scanners,
INTRODUCTION
OF RADIOACTIVE X-RAY
The
stable
planning.
trum, and simplicity that can lead to cost savings and reduction in scanner complexity. A stable emission rate and stable energy spectrum are extremely important in CT scanning because the image reconstruction process demands a high degree of consistency in the data. These requirements are usually met using highly stabilized, constant-potential X-ray generators as well as monitor detectors that record variations in source intensity during the scan. This equipment could in principle be replaced by a radioactive source. In addition, gantry design could be simplified and improved because the need for high voltage cables and cooling hose wind-up mechanisms or alternatively, high voltage slip rings would be eliminated. Finally, the more compact size and reduced weight of a radioactive source would suggest further savings in gantry design. The use of a monoenergetic photon source as available from many isopopes would have a number of important advantages in CT. As in conventional diagnostic radiology more monochromatic sources permit dose reductions in images of comparable contrast resolution. This is because the contrast mechanism usually has a strong energy dependence, weighted toward lower energies while the transmission function increases with energy. These two functions balance at a particular optimum energy depending upon the specific material involved, i.e. iodine, calcium or soft tissue. Wickizer et al.”
Computerized-tomographic (CT) scanning of the head and body has been a recent and revolutionary development in the field of diagnostic X-ray imaging. CT scanning offers a sensitive, non-invasive, soft tissue imaging modality that is useful for the diagnosis and treatment planning of an increasing variety of diseases.’ All current commercial CT scanners use conventional diagnostic X-ray sources consisting of a 10-100 mA electron beam directed onto a tungsten target which produces bremsstrahlung X-radiation as well as peaks of characteristic radiation from the tungsten K-shell. These sources are typically operated at the maximum intensity allowed by target heating and cooling considerations. Recently, the introduction of fan beam scanning technology in combination with position-sensitive detector arrays has permitted larger solid angles and more efficient source utilization.’ This raises the possibility of using a radioactive isotopic source of X-rays in CT in place of conventional sources. This paper will discuss the advantages of isotopic sources, evaluate feasibility in terms of existing technology, and suggest alternative source types. ADVANTAGES
Treatment
ISOTOPE
SOURCES
advantages of isotopic sources lies in their emission rate. monochromatic photon spec53
54
Radiation Oncology 0 Biology 0 Physics
showed a substantial dose reduction by adding copper filtration to a CT scanner. A second important advantage of monoenergetic sources is the resulting improvement in the accuracy of the CT image and the elimination of artifacts. Errors in the values of the CT-numbers representing linear attenuation coefficients are introduced by the non-linearity of the polychromatic X-ray attenuation process. This is due to the progressive hardening of the beam as it passes through the body. These errors appear as subtle shifts in the true values of various tissues of up to 3% or as artifactual bands of elevated density near bone or lower values near air. Manufacturers provide software corrections to this problem which have varying degrees of success. Accuracy is important for a wide variety of diagnostic purposes ranging from the distinction between cystic and solid masses, to the problem of bone mineral quantitation. Another application for highly accurate CT images would be in the development of treatment plans for heavy ion therapy. A further use of monoenergetic sources would be to use scans at two energies to determine separately atomic number and electron density images. This method has been used for bone mineral densitometry using projection radiography and carries over to CT with the advantage of CT’s 3-dimensional image.’ Alvarez and Macovski’ have published one energydependent reconstruction scheme using scans with two polychromatic beams with differing mean energies. This method requires an accurate knowledge of the spectral shapes as well as the exact energy dependence of the Compton and photoelectric attenuation processes. Dual monochromatic scans would simplify the analysis substantially and improve the accuracy and precision. TECHNICAL FEASIBILITY A number of prototype CT scanners have been constructed using isotopic sources. A low intensity, low resolution scanner using 13’1is in use in Zurich’ for scanning the wrist and forearm for bone mineral determinations. A lT3Gd source was used at UCLA’ and a 241Amsource at Stanford’ for research into the principles of pencil beam scanning and fan beam scanning, respectively. None of these scanners have sufficient intensity for general purpose clinical use. In order to estimate the required source intensity it is useful to extrapolate from the design parameters of existing successful scanners. Referring to Table 1, one can see that the use of larger detector arrays has reduced scan time and reduced source intensity. In the last column of Table 1 the equivalent source strength in Curies is given. This is based on a calculation of equivalent dose rate output at 70 beV of the source and may be an overestimate of the true
equivalent source strength because of the greater information per unit dose of a monoenergetic source. Thus the values in Table 1 may be conservative estimates of sources intensity by a factor of 2-3 or more. In a scanner optimized for an isotopic source, solid angles and scan time could be extended over the values typical of current commerical scanners. Hypothetical scanner 1 would relax the scan time of a GE or Varian design to 20 set and would require a 200 Ci (or less) isotopic source. Hypothetical scanner 2 would extend the solid angle by using an 80” fan with 600 detectors and achieve another factor of two reduction in source strength. In a scanner specialized for treatment planning for radiation therapy it could be argued that a 60 set scan time would be acceptable, providing that the design avoids streaks due to motion. One way to avoid motion streaking is to provide rapid gantry rotation so that the artifact produced by a movement is smeared over a reasonably large range of angles3 This could be accomplished by using several complete rotations during one 60 set scan. Thus hypothetical scanner 3 would provide blurring simulating conditions during a 60 set therapy session, but avoid streaks, and require 33 Ci (or less) source strength. Table 2 indicates two available high intensity sources with photon energies in the range typically employed in CT scanning. The short half-lives of ‘641b and lS3Gd help insure that the specific activity (Cilg) is adequate to get a high yield in a small area of 0.5 cm’ or less as required for target size in CT scanning. In addition self absorption in the target limits the maximum useable thickness to less than 1 mm so that effective quantities of isotope are typically substantially less than one gram. ‘@Yb has a number of contaminating high energy X-rays that become even more significant after the differential attenuation of the 52 keV line following passage through a body. One solution might be to use a detector such as Xenon that has a low detection efficiency above 50 keV. lT3Gd is a very promising candidate and offers two lines, one at a mean of 41.5 keV and the other at a mean of 100 keV. Theoretically lT3Gd could be produced with a specific activity of 3500 Cilg but in practice is limited to 6OCilg or less because of production difficulties leading to a large admixture of “*Gd along with the ‘“Gd. The maximum effective activity per cm* of area using a 60 Ci/g if “‘Gd would be 12 Ci. Higher activities would require an isotope separation step to enrich the fraction of lS3Gd relative to 15*Gd. Perhaps one of the new Uranium isotope separation technologies based on laser separation or plasma techniques could provide an economical solution to this problem.
Computerized-tomography: isotope sources 0 D. P. BOYD
Table 1. Representative
Type Delta body EMI body GElVarian Hypothetical Hypothetical Hypothetical
Table 2. Photon
Photon energy WV)
40 50 75 100 200 500
1 2 3
transmission
Scan time (set)
Number detectors
Ave. tube current (mA)
150 20 516 20 20 60
3lslice 30/slice 3OO/slice 3OO/slice 600 (80” fan) 600 (80” fan)
30 30 lo-20 3 1.5 0.5
Equiv. source strength 0) 2000 60?%0 200 100 33
Table 3. Diagnostic
vs energy
Transmission through 27 cm of tissue
requirement
(cm*/&
(%)
0)
0.249 0.214 0.180 0.166 0.135 0.096
0.12 0.31 0.78 1.13 2.61 7.49
470 212 100 74 40 20
Tissue attenuation coefficient
CT scanners
energy source candidates
Relative
source
Another solution to the specific activity problem would be to use a higher energy source. This would have the advantage of less self-absorption in the target, and more transmission through the body. Table 3 gives attenuation coefficients and per cent transmission through 27 cm of tissue as a function of energy. The last column gives the relative source strengths required to give equivalent contrast resolution in the CT image at each energy assuming the Compton process dominates the contrast mechanism. CT scans at high energies will have relatively less sensitivity to high 2 materials such as iodine and calcium than scans at low energy because of the rapid variation with energy of the photoelectric contrast mechanism. However, high energy CT images would be useful in particle therapy treatment planning where knowledge of the electron density distribution is required. Table 4 lists some possibilities for a higher energy source. ‘% is commercially available in 100 Ci strengths and emits photons at 308, 468 and 608 keV. At these energies there are substantial problems with detector arrays. Xenon ionization chambers would not be suitable because of low efficiency. NaI or BGO scintillators also have relatively low efficiencies in practical sizes. Crosstalk between detector elements in a closely packed array is likely to be a problem due to Compton scattering in the detector material. For these reasons a source in the 100-200 keV range such as “Co or “‘Eu may be better choices.
Source
Energies WV) 50 63 110 131 177 198 41.5 97 103
‘@Yb fl/2 = 30 days
“‘Gd 1I/2= 241 days
Table 4. Potential
Source
Energies (keV)
Yield @)
specific activity
185 45 18 11 22 35 60 30 22
(Cilg) - 25
-60
high energy sources Yield (%I
Activity
57co t llZ = 270 days
122 138
89 9
?
“‘Eu t 112 = 1.8 years
140 158 187 247
43 32 10 15
?
‘% t II2= 14 days
308 468 608
140 50 13
1OOCi in 3x4mm2
ALTERNATIVE SOURCES Several alternative techniques are available that offer many of the advantages of radioactive isotopic sources. The simplest alternative would be to use a highly filtered conventional bremmstrahlung source. However, filtering is achieved at a great penalty in intensity in most schemes. A more efficient method may be to use a secondary emitter near the tungsten target and produce characteristic radiation by means of X-ray fluorescence. Such a source may have a fluorescence efficiency in the range of 50%’ but suffers from poor solid angle coverage of the primary target if a small secondary source area is needed. A prototype secondary emission source has been described by Drexler et ~1.~This source uses a
56
Radiation
Oncology
0
Biology
0
Physics
well geometry to gain high geometrical efficiency. According to the authors, their source will produce 24 mR/min/mA at a distance of 50 cm and with a 4 mm spot size. This output would be in range of the requirement for hypothetical scanners 1 and 2 if the current 4 mA prototype can be scaled up to 100 mA. A third alternative would be to employ the energy dependent reconstruction scheme proposed by Alvarez and Macovski’ by acquiring projection data in two broad overlapping energy intervals. This can be accomplished by changing tube kVp between two values during the scan, or by switching between two filters, or by using an energy sensitive detection scheme such as by using alternatively filtered detectors in the array.
CONCLUSIONS The use of more monochromatic sources in CT scanning promises more accurate reconstructions and more specific information about selected materials. These benefits are expected to be particularly important in radiation therapy treatment planning. In addition, an isotopic source could reduce the cost and complexity of current commercial scanners and potentially offer an alternative approach to specific problems. Unfortunately, current source technology still falls short of the specific activity requirements for CT by a factor of 3-10. It is problematic whether the economic incentives of the CT industry will stimulate the necessary development in the isotopic source industry sufficiently to solve this problem.
REFERENCES
I. Alvarez, R.E., Macovski, A.: Energy selective reconPhys. struction in X-ray computerized-tomography. Med. Biol. 21: 733-734, 1976. 2. Boyd, D.: Instrumentation considerations for fan beam computerized tomography of the body. In Image Processing Project&s,
For 2-D and 3-D Reconstruction Stanford, 4-7 Aug. 1975 (Opt. Sot.
6.
7.
From
Am.), pp. TuA6-1 to 4. 3. Boyd, D., Korobkin, M., Moss, A.: Engineering status of computerized-tomographic scanning. Opt. Engng 16: 374, 1977 and Proc. SPIE 96: 1976. 4. Brooks, R.A., Di Chiro, G.: Principles of computer assisted tomography (CAT) in radiographic and radioisotopic imaging. Phys. Med. Biol. 21: 689-732, 1976. 5. Cho, Z.H., Ahn, IS., Isai, C.M.: IEEE Trans. Nucl. Sci. NS-21(l): 218. 1974.
8. 9.
10.
Drexler, G., Gossran, M., Panzer, W.: A new Auorescence X-ray source. Biomedical Dosimetry: Proc. Symp. in Vienna, IO-14 Mar. 1975, Vienna, IAEA, 1975, p. 499. Genant, H.K., Boyd, D., Korobkin, M., Norman, D.: Quantitative bone mineral determination using dual energy scanning (Abstract). Invest. Rad. 11 (6): 1976. Hoffman, E.J., Phelps, M.E.: Production of monoenergetic X-rays from 8-87 keV. Phys. Med. Biol. 19: 19-35, 1974. Ruegsegger, P., Elsasser, U., Anliker, M.: Quantification of bone mineralization using computed-tomography. Wickizer, C.R., Zacher. R., Krippner, K.: Analysis of X-ray beam hardening in CT. In Applications of Optical Instrumentation in Medicine V, Proceedings SPIE, Vol. 1976, %.
Inhomogeneities
and Treatment
COMPUTERIZED-TOMOGRAPHY:
Planning for Particle
Therapy
ISOTOPE SOURCES
DOUGLAS P. BOYD, Ph.D. University of California, San Francisco, Department of Radiology, School of Medicine, San Francisco, CA 94143, U.S.A. Recently pure rotary fan beam computerized tomographic (CT) body summers have been suuxssfuUy demonstrated by Artronfcs, AS&E, GE and Varian. These scanners have the large solid angle detection efUciency required by proposed practical isotope source-based CT scanners. Current scanners employ broad-energy bremsstrahlung sources at 10&140 kVp. The nonlinearity of polychromatk X-ray attenuation introduces signfficant absolute accuracy errors in the CT image even wben sophistkated software corrections are employed. Accuracy is important in heavy ion treatment planning and quantitative physWo&al measurements. A 100 Ci ‘-Gd source would enable accurate, industry competitive 20~ scans, but would requin a specfh activit&3-10 times tbe current state of tbe art.’ The dual monochromatic energy emission at “43” keV and “100” keV of ’ Gd can be utbed to sfmultaneously recomstruct electron density and atomic number CT images. These techniques may be useful for the development of a low-cost dedicatedCT scanner for therapy planning. Isotope sources,
Computer
assisted tomography
(CAT) scanners,
INTRODUCTION
OF RADIOACTIVE X-RAY
The
stable
planning.
trum, and simplicity that can lead to cost savings and reduction in scanner complexity. A stable emission rate and stable energy spectrum are extremely important in CT scanning because the image reconstruction process demands a high degree of consistency in the data. These requirements are usually met using highly stabilized, constant-potential X-ray generators as well as monitor detectors that record variations in source intensity during the scan. This equipment could in principle be replaced by a radioactive source. In addition, gantry design could be simplified and improved because the need for high voltage cables and cooling hose wind-up mechanisms or alternatively, high voltage slip rings would be eliminated. Finally, the more compact size and reduced weight of a radioactive source would suggest further savings in gantry design. The use of a monoenergetic photon source as available from many isopopes would have a number of important advantages in CT. As in conventional diagnostic radiology more monochromatic sources permit dose reductions in images of comparable contrast resolution. This is because the contrast mechanism usually has a strong energy dependence, weighted toward lower energies while the transmission function increases with energy. These two functions balance at a particular optimum energy depending upon the specific material involved, i.e. iodine, calcium or soft tissue. Wickizer et al.”
Computerized-tomographic (CT) scanning of the head and body has been a recent and revolutionary development in the field of diagnostic X-ray imaging. CT scanning offers a sensitive, non-invasive, soft tissue imaging modality that is useful for the diagnosis and treatment planning of an increasing variety of diseases.’ All current commercial CT scanners use conventional diagnostic X-ray sources consisting of a 10-100 mA electron beam directed onto a tungsten target which produces bremsstrahlung X-radiation as well as peaks of characteristic radiation from the tungsten K-shell. These sources are typically operated at the maximum intensity allowed by target heating and cooling considerations. Recently, the introduction of fan beam scanning technology in combination with position-sensitive detector arrays has permitted larger solid angles and more efficient source utilization.’ This raises the possibility of using a radioactive isotopic source of X-rays in CT in place of conventional sources. This paper will discuss the advantages of isotopic sources, evaluate feasibility in terms of existing technology, and suggest alternative source types. ADVANTAGES
Treatment
ISOTOPE
SOURCES
advantages of isotopic sources lies in their emission rate. monochromatic photon spec53
54
Radiation Oncology 0 Biology 0 Physics
showed a substantial dose reduction by adding copper filtration to a CT scanner. A second important advantage of monoenergetic sources is the resulting improvement in the accuracy of the CT image and the elimination of artifacts. Errors in the values of the CT-numbers representing linear attenuation coefficients are introduced by the non-linearity of the polychromatic X-ray attenuation process. This is due to the progressive hardening of the beam as it passes through the body. These errors appear as subtle shifts in the true values of various tissues of up to 3% or as artifactual bands of elevated density near bone or lower values near air. Manufacturers provide software corrections to this problem which have varying degrees of success. Accuracy is important for a wide variety of diagnostic purposes ranging from the distinction between cystic and solid masses, to the problem of bone mineral quantitation. Another application for highly accurate CT images would be in the development of treatment plans for heavy ion therapy. A further use of monoenergetic sources would be to use scans at two energies to determine separately atomic number and electron density images. This method has been used for bone mineral densitometry using projection radiography and carries over to CT with the advantage of CT’s 3-dimensional image.’ Alvarez and Macovski’ have published one energydependent reconstruction scheme using scans with two polychromatic beams with differing mean energies. This method requires an accurate knowledge of the spectral shapes as well as the exact energy dependence of the Compton and photoelectric attenuation processes. Dual monochromatic scans would simplify the analysis substantially and improve the accuracy and precision. TECHNICAL FEASIBILITY A number of prototype CT scanners have been constructed using isotopic sources. A low intensity, low resolution scanner using 13’1is in use in Zurich’ for scanning the wrist and forearm for bone mineral determinations. A lT3Gd source was used at UCLA’ and a 241Amsource at Stanford’ for research into the principles of pencil beam scanning and fan beam scanning, respectively. None of these scanners have sufficient intensity for general purpose clinical use. In order to estimate the required source intensity it is useful to extrapolate from the design parameters of existing successful scanners. Referring to Table 1, one can see that the use of larger detector arrays has reduced scan time and reduced source intensity. In the last column of Table 1 the equivalent source strength in Curies is given. This is based on a calculation of equivalent dose rate output at 70 beV of the source and may be an overestimate of the true
equivalent source strength because of the greater information per unit dose of a monoenergetic source. Thus the values in Table 1 may be conservative estimates of sources intensity by a factor of 2-3 or more. In a scanner optimized for an isotopic source, solid angles and scan time could be extended over the values typical of current commerical scanners. Hypothetical scanner 1 would relax the scan time of a GE or Varian design to 20 set and would require a 200 Ci (or less) isotopic source. Hypothetical scanner 2 would extend the solid angle by using an 80” fan with 600 detectors and achieve another factor of two reduction in source strength. In a scanner specialized for treatment planning for radiation therapy it could be argued that a 60 set scan time would be acceptable, providing that the design avoids streaks due to motion. One way to avoid motion streaking is to provide rapid gantry rotation so that the artifact produced by a movement is smeared over a reasonably large range of angles3 This could be accomplished by using several complete rotations during one 60 set scan. Thus hypothetical scanner 3 would provide blurring simulating conditions during a 60 set therapy session, but avoid streaks, and require 33 Ci (or less) source strength. Table 2 indicates two available high intensity sources with photon energies in the range typically employed in CT scanning. The short half-lives of ‘641b and lS3Gd help insure that the specific activity (Cilg) is adequate to get a high yield in a small area of 0.5 cm’ or less as required for target size in CT scanning. In addition self absorption in the target limits the maximum useable thickness to less than 1 mm so that effective quantities of isotope are typically substantially less than one gram. ‘@Yb has a number of contaminating high energy X-rays that become even more significant after the differential attenuation of the 52 keV line following passage through a body. One solution might be to use a detector such as Xenon that has a low detection efficiency above 50 keV. lT3Gd is a very promising candidate and offers two lines, one at a mean of 41.5 keV and the other at a mean of 100 keV. Theoretically lT3Gd could be produced with a specific activity of 3500 Cilg but in practice is limited to 6OCilg or less because of production difficulties leading to a large admixture of “*Gd along with the ‘“Gd. The maximum effective activity per cm* of area using a 60 Ci/g if “‘Gd would be 12 Ci. Higher activities would require an isotope separation step to enrich the fraction of lS3Gd relative to 15*Gd. Perhaps one of the new Uranium isotope separation technologies based on laser separation or plasma techniques could provide an economical solution to this problem.
Computerized-tomography: isotope sources 0 D. P. BOYD
Table 1. Representative
Type Delta body EMI body GElVarian Hypothetical Hypothetical Hypothetical
Table 2. Photon
Photon energy WV)
40 50 75 100 200 500
1 2 3
transmission
Scan time (set)
Number detectors
Ave. tube current (mA)
150 20 516 20 20 60
3lslice 30/slice 3OO/slice 3OO/slice 600 (80” fan) 600 (80” fan)
30 30 lo-20 3 1.5 0.5
Equiv. source strength 0) 2000 60?%0 200 100 33
Table 3. Diagnostic
vs energy
Transmission through 27 cm of tissue
requirement
(cm*/&
(%)
0)
0.249 0.214 0.180 0.166 0.135 0.096
0.12 0.31 0.78 1.13 2.61 7.49
470 212 100 74 40 20
Tissue attenuation coefficient
CT scanners
energy source candidates
Relative
source
Another solution to the specific activity problem would be to use a higher energy source. This would have the advantage of less self-absorption in the target, and more transmission through the body. Table 3 gives attenuation coefficients and per cent transmission through 27 cm of tissue as a function of energy. The last column gives the relative source strengths required to give equivalent contrast resolution in the CT image at each energy assuming the Compton process dominates the contrast mechanism. CT scans at high energies will have relatively less sensitivity to high 2 materials such as iodine and calcium than scans at low energy because of the rapid variation with energy of the photoelectric contrast mechanism. However, high energy CT images would be useful in particle therapy treatment planning where knowledge of the electron density distribution is required. Table 4 lists some possibilities for a higher energy source. ‘% is commercially available in 100 Ci strengths and emits photons at 308, 468 and 608 keV. At these energies there are substantial problems with detector arrays. Xenon ionization chambers would not be suitable because of low efficiency. NaI or BGO scintillators also have relatively low efficiencies in practical sizes. Crosstalk between detector elements in a closely packed array is likely to be a problem due to Compton scattering in the detector material. For these reasons a source in the 100-200 keV range such as “Co or “‘Eu may be better choices.
Source
Energies WV) 50 63 110 131 177 198 41.5 97 103
‘@Yb fl/2 = 30 days
“‘Gd 1I/2= 241 days
Table 4. Potential
Source
Energies (keV)
Yield @)
specific activity
185 45 18 11 22 35 60 30 22
(Cilg) - 25
-60
high energy sources Yield (%I
Activity
57co t llZ = 270 days
122 138
89 9
?
“‘Eu t 112 = 1.8 years
140 158 187 247
43 32 10 15
?
‘% t II2= 14 days
308 468 608
140 50 13
1OOCi in 3x4mm2
ALTERNATIVE SOURCES Several alternative techniques are available that offer many of the advantages of radioactive isotopic sources. The simplest alternative would be to use a highly filtered conventional bremmstrahlung source. However, filtering is achieved at a great penalty in intensity in most schemes. A more efficient method may be to use a secondary emitter near the tungsten target and produce characteristic radiation by means of X-ray fluorescence. Such a source may have a fluorescence efficiency in the range of 50%’ but suffers from poor solid angle coverage of the primary target if a small secondary source area is needed. A prototype secondary emission source has been described by Drexler et ~1.~This source uses a
56
Radiation
Oncology
0
Biology
0
Physics
well geometry to gain high geometrical efficiency. According to the authors, their source will produce 24 mR/min/mA at a distance of 50 cm and with a 4 mm spot size. This output would be in range of the requirement for hypothetical scanners 1 and 2 if the current 4 mA prototype can be scaled up to 100 mA. A third alternative would be to employ the energy dependent reconstruction scheme proposed by Alvarez and Macovski’ by acquiring projection data in two broad overlapping energy intervals. This can be accomplished by changing tube kVp between two values during the scan, or by switching between two filters, or by using an energy sensitive detection scheme such as by using alternatively filtered detectors in the array.
CONCLUSIONS The use of more monochromatic sources in CT scanning promises more accurate reconstructions and more specific information about selected materials. These benefits are expected to be particularly important in radiation therapy treatment planning. In addition, an isotopic source could reduce the cost and complexity of current commercial scanners and potentially offer an alternative approach to specific problems. Unfortunately, current source technology still falls short of the specific activity requirements for CT by a factor of 3-10. It is problematic whether the economic incentives of the CT industry will stimulate the necessary development in the isotopic source industry sufficiently to solve this problem.
REFERENCES
I. Alvarez, R.E., Macovski, A.: Energy selective reconPhys. struction in X-ray computerized-tomography. Med. Biol. 21: 733-734, 1976. 2. Boyd, D.: Instrumentation considerations for fan beam computerized tomography of the body. In Image Processing Project&s,
For 2-D and 3-D Reconstruction Stanford, 4-7 Aug. 1975 (Opt. Sot.
6.
7.
From
Am.), pp. TuA6-1 to 4. 3. Boyd, D., Korobkin, M., Moss, A.: Engineering status of computerized-tomographic scanning. Opt. Engng 16: 374, 1977 and Proc. SPIE 96: 1976. 4. Brooks, R.A., Di Chiro, G.: Principles of computer assisted tomography (CAT) in radiographic and radioisotopic imaging. Phys. Med. Biol. 21: 689-732, 1976. 5. Cho, Z.H., Ahn, IS., Isai, C.M.: IEEE Trans. Nucl. Sci. NS-21(l): 218. 1974.
8. 9.
10.
Drexler, G., Gossran, M., Panzer, W.: A new Auorescence X-ray source. Biomedical Dosimetry: Proc. Symp. in Vienna, IO-14 Mar. 1975, Vienna, IAEA, 1975, p. 499. Genant, H.K., Boyd, D., Korobkin, M., Norman, D.: Quantitative bone mineral determination using dual energy scanning (Abstract). Invest. Rad. 11 (6): 1976. Hoffman, E.J., Phelps, M.E.: Production of monoenergetic X-rays from 8-87 keV. Phys. Med. Biol. 19: 19-35, 1974. Ruegsegger, P., Elsasser, U., Anliker, M.: Quantification of bone mineralization using computed-tomography. Wickizer, C.R., Zacher. R., Krippner, K.: Analysis of X-ray beam hardening in CT. In Applications of Optical Instrumentation in Medicine V, Proceedings SPIE, Vol. 1976, %.
