1975, British Journal of Radiology, -AS", 327-351
MAY
1975
Review article-Bone scanning By M. V. Merrick, M.A., B.M., B.Ch. (Oxon.) M.Sc. (Lond.), D.M.R.D., F.R.C.R.* Radiodiagnostic Department, Royal Postgraduate Medical School, Du Cane Road, London, W.12. (Received May, 1974)
ABSTRACT
The discovery of a number of phosphate complexes labelled with " T c m that localize in bone has aroused widespread interest in bone scanning. The physiological properties of these and other clinically useful bone-seeking radiopharmaceuticals are compared, and their physical properties assessed in relation to the characteristics and limitations of available detector systems. A hypothesis is put forward to explain the behaviour of the technetiumlabelled agents. It is concluded that although there are differences in biochemical behaviour between these agents, strontium and fluorine, all three may, under suitable conditions, give similar clinical information. The radiation dose received by the patient is least with the usual dose of "Tc m , and the blood clearance of the diphosphonate and pyrophosphate preparations is faster than that of strontium, although slower than fluorine. The y-ray energy of technetium permits a much greater efficiency of detection than that of fluorine. These factors, together with the general availability of ltri Tcm and its relatively low cost make the technetium diphosphonate or pyrophosphate preparations the agents of choice for most skeletal radioisotope imaging. However, there are as yet insufficient follow-up studies to be able to assess the incidence of either false-negative or false-positive findings with these agents.
The earliest indication that the administration of radioisotopes to adults could lead to the accumulation (and retention) of radioactivity in the skeleton came as a result of observations in luminous dial painters (Bickel, 1912; Blum, 1924; Hoffman, 1924). It was at first thought that the radioisotopes were taken up by reticulo-endothelial cells, until it was shown by autoradiography of human post-mortem material that the bone itself was radioactive (Martland, 1926). However, until the incorporation of 32 P-phosphate into the skeleton of adult rats was demonstrated (Chiewitz and Hevesy, 1936) it was widely believed that bone was metabolically inactive. Volker et al. (1940) found in vitro that fluoride labelled with 18 F was taken up by hydroxyapatite, bone, enamel and dentin. The incorporation and concentration of a strontium isotope (90Sr) into bone tumours in animals and man was recorded by Treadwell et al. (1942). Radioactivity was not distributed uniformly throughout the tumours. With the greater range of radioisotopes that •Present address: Medical Radioisotope Department, Western General Hospital, Edinburgh EH4 2XU.
became available after 1945, and the interest in finding medical applications, a considerable effort was put into the development of a bone-seeking radionuclide to be used for treatment of bone tumours. In the course of this work high uptakes of gallium radioisotopes, in particular 72Ga, were observed (Dudley and Maddox, 1949; Dudley, Maddox and La Rue, 1949), although the difference between normal and abnormal areas was insufficient for therapeutic purposes (Dudley, Markowitz and Mitchell, 1956). The rectilinear scanner, derived from a prototype intended for thyroid imaging (Cassen et al., 1951) became available synchronously with these developments in radiopharmaceuticals. 87 Sr m (Meyers, 1960), 85Sr (Fleming, Mcllraith and King, 1961), 18 F (Blau, Nagler and Bender, 1962) and 68Ga (Hayes, Carlton and Byrd, 1965) were all employed clinically, and numerous potential alternatives were suggested (McAfee and Subramanian, 1966). For a number of reasons none was universally accepted, although all are capable of giving useful clinical information. The discovery of a group of bone-seeking compounds which could be labelled with 99Tcm (Subramanian and McAfee, 1971) made available, for the first time, bone-scanning agents which combined low radiation dose, high photon yield, good anatomical information and general availability. These factors, associated with the improving resolution of imaging devices, have been largely responsible for the recent resurgence of interest in bone scanning.
BASIC PHYSIOLOGY
Bone consists of an organic matrix in which is dispersed a crystalline mineral phase. The matrix is made up of polymers of glucuronic acid and hexosamines, in which are suspended collagen fibres, osteocytes and the vascular endothelium. Collagen is quantitatively the major organic constituent, but the mineral accounts for about two-thirds of the dry weight. Marrow is an unrelated tissue despite its situation within the medullary core of the skeleton (Copp and Shim, 1963).
327
VOL.
48, No. 569 M. V. Merrick
Most radioisotopic labels of bone involve an exchange with one or more of the inorganic components, but some tracers, notably gallium, zirconium, cerium and yttrium label the organic matrix (Neuman and Neuman, 1953). There is some evidence that plutonium may be attached to the surface of the osteocyte (Taylor and Chipperfield,
1970), and other transuranic elements may be expected to behave similarly. The principal mineral is a calcium salt with the stoiometric composition Caio(P04)e(OH)2. It occurs in the skeleton as hydroxyapatite. This is a crystallographic rather than a chemical description. When crystals of this complexity are formed the perfect
TABLE I PROPERTIES OF RADIONUCLIDES MENTIONED IN THIS ARTICLE
Useful
Gamma*
External photon* • i
(key)
0 / 0
J
Element
isotopes
Half life*
Fluorine
18p
1 -85 hours
5110 + )
32p
14-3 days
45
165 days 4-7 days
Phosphorus (as phosphate) Calcium
Gallium
47
67
Technetium (as
Bone
194
2-10
50-500
200-3,000
fi — only
—
0-2
2000
(3 — onlv 1310 830 480
76-3 5-7 5-7
0-005 0-01
75 30-70
4,000100,000 250-650 300
Ga
1-13 hours 65 days
511 0 +
87g r m
2-8 hours
99 r p c m
6 hours
Tc-SnEHDP)
131
Ba
11-5 days
Samarium
135 Ba m 153gm
1 -2 davs 1 -96 days
Dysprosium
157Dy
Barium
Whole body
3-25 days
85Sr
167
513
42 24 22 7 172 100
2 0-05-0-2
60 250-4,000
388
78
1-5
7-100
92 182 300 390
' 140
90-1
650
10-20
110-220
Blau etal., 1962 Spencer et al., 1967
f IRCP1971
J
2,0005,000
Saunders, Tavlor and Trott, 1973
800 1,50010,000 1003,000
Weber etal., 1969 Vennart and Alinski, 1962 IRCP, 1971
450-900
Castronovo and Callaghan, 1972 O'Mara and Subramanian, 1972; Spencer, Lange andTreves, 1970, 1971 Sved and Hosain, 1971 Karikaefa/,,1973
48 13 5 19 28 16 28
0-3
550
4,300
3 1
900 ?
3,500 3,300
8-1 hours
326
94
10
?
1,270
9-6 days
2 43 63 28 25
0-5
700
3,500
4
?
760
O'Mara etal., 1969
6 2-8
1
?
1,300
O'Mara et al., 1969
Erbium
7-5 hours
532 208 308 296 112
Lutetium
177Lu
6-7 days
208
Tm
2-5
Reference
496 373 250 216 124 268 103 70
171 E r
Thulium
a n YY~\ 1 t"l 1 Q —
aUlilllllo
tered (mCi)
Ga
68
Strontium
Ca Ca
Absorbed dose m rad from usual injected activity
Usual activity
Subramanian et al., 1971a Steinberg et al., 1973
n
p—
113
*Lederer, Hollander and Perlman (1968). Only gamma emissions with an external photon yield greater than 5 per cent are listed. f Where a range of dose is given, the lower figure assumes the lowest dose estimate and the smaller quoted mCi dose, and the upper the highest estimated absorbed dose and the higher quoted mCi dose.
328
Review article. Bone scanning
crystal structure does not develop immediately. As first precipitated they are small, imperfect and surrounded by a large hydration shell. Solution and recrystallization occurs spontaneously at a rapid rate (Neuman and Neuman, 1953; Pautard, 1964). The hydroxyapatite crystals present in bone measure 30nmX 30nmx 5 nm. There is, therefore, a large surface-to-volume ratio, high surface charge and large hydration shell. Ions in the surrounding solution can enter this hydration shell, and may then exchange with the layer of hydrated calcium, phosphate and hydroxide ions on the crystal surface. As a result of imperfections within the crystal lattice a further, albeit slower, exchange occurs between the surface of the crystal and its interior (Neuman and Xeuman, 1957). Maturation of bone is associated with perfection of its crystalline structure, leading to a mineral phase which is less hydrated and less readily accessible to ion movements (Heaney and Whedon, 1958). 99 per cent of the mineral in mature compact bone is in this state, and therefore relatively unavailable for exchange with administered radioisotopes (McLean, 1958). The distribution of 45Ca in bone slices incubated in vitro with 45CaCl2 (Dallemagne and Richelle, 1962) is similar to that observed when radioisotope is administered in vivo (Comar, Lotz and Boyd, 1952; Wood et al., 1970) indicating that, even in the absence of blood flow, uptake of radioisotope is affected by local variations in the state of the bone. Under physiological conditions blood supply and mineral turnover are interrelated (Brookes and Helal, 1970), so that it may not be possible to distinguish between their respective contributions to radioisotope uptake.
calcium and strontium are available to pass either into or out of plasma, the extraction ratio may be expected to fall. Measurements over the five minutes after injection have not been able to confirm this (Copp and Shim, 1964).
RADIOPHARMACEUTICALS
1. Calcium and its analogues
Although 47Ca has been used for bone scanning (Corey et al., 1961, 1962; Jasinski, 1962) it is not an ideal radioisotope for imaging because of the high energy of its y emissions, and unavoidable contamination with 45Ca (Table I). Approximately 7 per cent of the calcium in bone mineral can be replaced by strontium in vitro. The rate of exchange is governed in the early stages by the calcium to strontium ratio, but at equilibrium there is discrimination against strontium (Harrison et al., 1959; Schoenberg, 1963). This discrimination also operates in vivo. After the intravenous injection of a bolus of strontium, there is a high extraction— of the order of 70 per cent—on the first passage through bone. On subsequent passages, when both
FIG. 1. Posterior scan started two hours after 4 mCi of 87Srm. There is a lesion in the lower thoracic spine, and probably also one at the lumbosacral junction. The irregular uptake in the upper thoracic spine is associated with the presence of degenerative changes and early osteophyte formation.
329
VOL.
48, No. 569 M. V. Merrick
Strontium is thus taken up rapidly by bone, levels of radioactivity reaching half their maximum value within ten minutes, and 90 per cent within one hour (Charkes and Sklaroff, 1965; Kshirsagar, Lloyd and Vaughan, 1966). The amount in the skeleton represents 20-50 per cent of the administered dose. The highest values are found in children, and in patients with skeletal metastases (Guerin and Guerin, 1960; Cederquist, 1964). Despite the rapid skeletal uptake, plasma clearance (of the fraction not taken up by bone) is relatively slow, 10-20 per cent of the administered dose of either strontium or calcium still being in the circulation at one hour, and 1 per cent at 24 hours (Harrison, Carr and Sutton, 1957; Weber et al., 1969). Excreted strontium appears both in urine and faeces. Most of the urinary excretion occurs within the first 24 hours (Spencer, Laszlo and Brothers, 1957). Faecal excretion is sufficient to cause difficulties in scan interpretation. There are two radioisotopes of strontium with physical characteristics suitable for imaging (Table I). The half-life of 87 Sr m is 2-8 hours. The scan must thus be performed on the day of injection, at an interval which is a compromise between the plasma clearance and radioactive decay. A delay of two to three hours has been suggested (Jasinski et al., 1968). At this time extra-osseous radioactivity is readily detectable (Alexander and Gillespie, 1971). Moreover the high count-rate from non-target tissues reduces any percentage difference between lesion and normal bone, and can conceal abnormalities (Charkes, 1969). The long physical (65-day) and biological (800day) half-life of 85 Sr limits the amount which can be administered (Charkes, Sklaroff and Bierly, 1964). The extended examination times associated with a low7 count-rate usually preclude examination of the entire skeleton. To overcome this difficulty a rapid whole-body scan displayed at reduced size has been suggested as a screening procedure, followed by a slower scan of any areas where the interpretation is in doubt (Braunstein, Hernberg and Chandra, 1971). The alternative is to use a profile or point-counting technique (Bauer and Wendeberg, 1959; Rosenthal, 1965; Papworth and Andrews, 1967; Robillard et al., 1970; Taskinen and Vahatalo, 1971). There is evidence that the latter is the more sensitive (McCready, French and MacDonald, 1970; DeNardo et al., 1967). When patients with symptoms suggestive of metastatic disease are scanned with either of these radioisotopes, deposits are found on the scan in up
to 20 per cent more cases than are visible radiologically (Charkes and Sklaroff, 1964; Erjavec, 1965; Simpson and Orange, 1965; De Nardo and Volpe, 1966; Briggs, 1967; Bessler, 1968; Benoit, Torres and Peterson, 1968; Scott and Adams, 1974). However, about 5 per cent of deposits identified on radiographs are not seen on the scan. This is particularly the case when there is fully mineralized reactive bone (Charkes, Sklaroff and Young, 1966) readily apparent radiologically, but indicating healing following cessation of tumour growth. This is most commonly associated with a favourable response to endocrine therapy, in patients with carcinoma of the prostate or breast. The highest uptakes occur in lesions showing both bone destruction and new bone formation. There is rather less uptake in predominantly lytic deposits (Charkes, Young and Sklaroff, 1968; Charkes, 1972). The poor statistics inevitable with 85Sr are to some extent offset by the high target to background ratios, which are usually maximal one to two weeks after injection. Where there is rapid turnover the peak may occur much earlier (Core}' et al., 1962a). The better statistics obtainable with 87Srm give clearer images. As a result there are some lesions which may be detected using either radioisotope but are not seen with the other (Charkes, 1969; Sauer, 1971). The availability of bone-seeking technetium complexes (q.v. infra) has greatly reduced the use of strontium. Where strontium is still used the most sensitive technique is to perform profile scans two to seven days after the administration of 85Sr, followed by a rectilinear scan of any areas suspicious on the profile. Barium is cleared from the blood much more rapidly than strontium, only 1 per cent of an injected dose remaining in the circulation after three hours (Bauer, Carlsson and Lindquist, 1957; Weber et al., 1969). However, the extraction of barium by bone is slightly lower than that of strontium (Wooton, unpublished), and more is excreted, mainly via the faeces (Harrison et al., 1967). Clinical experience with barium radioisotopes has been restricted by the limited availability and high cost of preparations of sufficiently high specific activity. The clinical results which have been published (Spencer, Lange and Treves, 1971; O'Mara and Subramanian, 1972; Syed, Hosain and Wagner, 1972) have not demonstrated any advantage over more readily available agents. 2. Technetium-labelled compounds
Neither pertechnetate (TcO4~~)northe technetium (Tc 44 ) ion localize to any useful extent in bone,
330
MAY
1975
Review article. Bone scanning
but a number of complexes of technetium with phosphorus-containing compounds are concentrated in the skeleton. The first such substances to be described (Subramanian and McAfee, 1971) were based upon condensed phosphates, compounds formed when orthophosphates are heated: OH OH OH OH O-P-OH
o
+ HO-p-cr. .-o-p-o-p-cr o
o
+H 2 O
o
(pyrophosphate) On further heating this process is repeated to produce chains, which may be of any length, and either straight or branching. The final composition is dependent upon the conditions under which the reaction is carried out. Separation of a particular length chain can be accomplished with difficulty, but on standing slow hydrolysis ultimately results in a mixture of shorter-chain compounds. In vitro condensed phosphates inhibit the deposition from solution of a number of calcium salts, including the orthophosphate, and prevent the conversion of amorphous calcium phosphate to hydroxyapatite. Hydroxyapatite crystals which have been exposed to solutions of condensed phosphates, then washed, dissolve more slowly than untreated crystals. These effects are believed to be the result of replacement of orthophosphate groups on the surface of the crystal or nidus by condensed phosphate, thus altering the charge distribution and reactiveness. Pyrophosphate has been identified as one of the naturally occurring inhibitors of calcium deposition in vivo (Russell and Smith, 1973). Intravenously administered ortho or condensed phosphates, labelled with 32 P, or chelated to labelled metals such as 51Cr (Anghileri and Miller, 1971, 1972; Anghileri, 1972) accumulate to a greater extent in the skeleton than in any other tissue. More radioactivity accumulates in bone when straight chains are given than when branching chain fractions are used. However, in vivo condensed phosphates are rapidly hydrolysed to orthophosphate by phosphatases. The original bone-scanning agents based upon these compounds utilized a short chain of three phosphate residues (tripolyphosphate) (Subramanian and McAfee, 1971). It was subsequently found that more rapid blood clearance and higher bone-tosoft tissue ratios were obtained with a straight chain of 46 residues (PP 46) (Subramanian et al., 1971b). The initial blood clearance in rabbits was more rapid than that of strontium, while levels in most tissues,
including bone, were similar. Technetium levels in liver were appreciably higher. Some workers have reached similar conclusions about optimum chain length (Dewanjee, Fletcher and Davis, 1972), but other groups obtained their best results with pyrophosphate or tripolyphosphate (Boc et al., 1973; Hegesippe et al., 1973; Hopkins, Creighton and Van Deripe, 1973; Hosain et al., 1973; King et al., 1973; McKamey, Artis and Hansen, 1973; Murray et al., 1972; Huberty, Hattner and Powell, 1974). This discrepancy may be due in part to the labile nature of the longer chains (Bowen and Garnett, 1974). However, every batch of every preparation has to be tested finally by a bio-assay as, despite the most meticulous control of their chemical composition, they occasionally fail to localize in bone (Subramanian et al., 1912%). This implies that no chemical test yet devised is specific for the biologically active compound. Thin layer chromatography on hydroxyapatite plates may prove a useful innovation in this respect (Radiochemical Centre, to be published). Before technetium can be labelled to any of the condensed phosphates it must first be reduced from pertechnetate (the form in which it is eluted from its generator) to Tc. Although there are a number of reducing agents which can effect this transformation, bone localization of the technetium occurs only when the stannous ion (Sn 2+ ) or Ti
MAY
1975
Review article-Bone scanning By M. V. Merrick, M.A., B.M., B.Ch. (Oxon.) M.Sc. (Lond.), D.M.R.D., F.R.C.R.* Radiodiagnostic Department, Royal Postgraduate Medical School, Du Cane Road, London, W.12. (Received May, 1974)
ABSTRACT
The discovery of a number of phosphate complexes labelled with " T c m that localize in bone has aroused widespread interest in bone scanning. The physiological properties of these and other clinically useful bone-seeking radiopharmaceuticals are compared, and their physical properties assessed in relation to the characteristics and limitations of available detector systems. A hypothesis is put forward to explain the behaviour of the technetiumlabelled agents. It is concluded that although there are differences in biochemical behaviour between these agents, strontium and fluorine, all three may, under suitable conditions, give similar clinical information. The radiation dose received by the patient is least with the usual dose of "Tc m , and the blood clearance of the diphosphonate and pyrophosphate preparations is faster than that of strontium, although slower than fluorine. The y-ray energy of technetium permits a much greater efficiency of detection than that of fluorine. These factors, together with the general availability of ltri Tcm and its relatively low cost make the technetium diphosphonate or pyrophosphate preparations the agents of choice for most skeletal radioisotope imaging. However, there are as yet insufficient follow-up studies to be able to assess the incidence of either false-negative or false-positive findings with these agents.
The earliest indication that the administration of radioisotopes to adults could lead to the accumulation (and retention) of radioactivity in the skeleton came as a result of observations in luminous dial painters (Bickel, 1912; Blum, 1924; Hoffman, 1924). It was at first thought that the radioisotopes were taken up by reticulo-endothelial cells, until it was shown by autoradiography of human post-mortem material that the bone itself was radioactive (Martland, 1926). However, until the incorporation of 32 P-phosphate into the skeleton of adult rats was demonstrated (Chiewitz and Hevesy, 1936) it was widely believed that bone was metabolically inactive. Volker et al. (1940) found in vitro that fluoride labelled with 18 F was taken up by hydroxyapatite, bone, enamel and dentin. The incorporation and concentration of a strontium isotope (90Sr) into bone tumours in animals and man was recorded by Treadwell et al. (1942). Radioactivity was not distributed uniformly throughout the tumours. With the greater range of radioisotopes that •Present address: Medical Radioisotope Department, Western General Hospital, Edinburgh EH4 2XU.
became available after 1945, and the interest in finding medical applications, a considerable effort was put into the development of a bone-seeking radionuclide to be used for treatment of bone tumours. In the course of this work high uptakes of gallium radioisotopes, in particular 72Ga, were observed (Dudley and Maddox, 1949; Dudley, Maddox and La Rue, 1949), although the difference between normal and abnormal areas was insufficient for therapeutic purposes (Dudley, Markowitz and Mitchell, 1956). The rectilinear scanner, derived from a prototype intended for thyroid imaging (Cassen et al., 1951) became available synchronously with these developments in radiopharmaceuticals. 87 Sr m (Meyers, 1960), 85Sr (Fleming, Mcllraith and King, 1961), 18 F (Blau, Nagler and Bender, 1962) and 68Ga (Hayes, Carlton and Byrd, 1965) were all employed clinically, and numerous potential alternatives were suggested (McAfee and Subramanian, 1966). For a number of reasons none was universally accepted, although all are capable of giving useful clinical information. The discovery of a group of bone-seeking compounds which could be labelled with 99Tcm (Subramanian and McAfee, 1971) made available, for the first time, bone-scanning agents which combined low radiation dose, high photon yield, good anatomical information and general availability. These factors, associated with the improving resolution of imaging devices, have been largely responsible for the recent resurgence of interest in bone scanning.
BASIC PHYSIOLOGY
Bone consists of an organic matrix in which is dispersed a crystalline mineral phase. The matrix is made up of polymers of glucuronic acid and hexosamines, in which are suspended collagen fibres, osteocytes and the vascular endothelium. Collagen is quantitatively the major organic constituent, but the mineral accounts for about two-thirds of the dry weight. Marrow is an unrelated tissue despite its situation within the medullary core of the skeleton (Copp and Shim, 1963).
327
VOL.
48, No. 569 M. V. Merrick
Most radioisotopic labels of bone involve an exchange with one or more of the inorganic components, but some tracers, notably gallium, zirconium, cerium and yttrium label the organic matrix (Neuman and Neuman, 1953). There is some evidence that plutonium may be attached to the surface of the osteocyte (Taylor and Chipperfield,
1970), and other transuranic elements may be expected to behave similarly. The principal mineral is a calcium salt with the stoiometric composition Caio(P04)e(OH)2. It occurs in the skeleton as hydroxyapatite. This is a crystallographic rather than a chemical description. When crystals of this complexity are formed the perfect
TABLE I PROPERTIES OF RADIONUCLIDES MENTIONED IN THIS ARTICLE
Useful
Gamma*
External photon* • i
(key)
0 / 0
J
Element
isotopes
Half life*
Fluorine
18p
1 -85 hours
5110 + )
32p
14-3 days
45
165 days 4-7 days
Phosphorus (as phosphate) Calcium
Gallium
47
67
Technetium (as
Bone
194
2-10
50-500
200-3,000
fi — only
—
0-2
2000
(3 — onlv 1310 830 480
76-3 5-7 5-7
0-005 0-01
75 30-70
4,000100,000 250-650 300
Ga
1-13 hours 65 days
511 0 +
87g r m
2-8 hours
99 r p c m
6 hours
Tc-SnEHDP)
131
Ba
11-5 days
Samarium
135 Ba m 153gm
1 -2 davs 1 -96 days
Dysprosium
157Dy
Barium
Whole body
3-25 days
85Sr
167
513
42 24 22 7 172 100
2 0-05-0-2
60 250-4,000
388
78
1-5
7-100
92 182 300 390
' 140
90-1
650
10-20
110-220
Blau etal., 1962 Spencer et al., 1967
f IRCP1971
J
2,0005,000
Saunders, Tavlor and Trott, 1973
800 1,50010,000 1003,000
Weber etal., 1969 Vennart and Alinski, 1962 IRCP, 1971
450-900
Castronovo and Callaghan, 1972 O'Mara and Subramanian, 1972; Spencer, Lange andTreves, 1970, 1971 Sved and Hosain, 1971 Karikaefa/,,1973
48 13 5 19 28 16 28
0-3
550
4,300
3 1
900 ?
3,500 3,300
8-1 hours
326
94
10
?
1,270
9-6 days
2 43 63 28 25
0-5
700
3,500
4
?
760
O'Mara etal., 1969
6 2-8
1
?
1,300
O'Mara et al., 1969
Erbium
7-5 hours
532 208 308 296 112
Lutetium
177Lu
6-7 days
208
Tm
2-5
Reference
496 373 250 216 124 268 103 70
171 E r
Thulium
a n YY~\ 1 t"l 1 Q —
aUlilllllo
tered (mCi)
Ga
68
Strontium
Ca Ca
Absorbed dose m rad from usual injected activity
Usual activity
Subramanian et al., 1971a Steinberg et al., 1973
n
p—
113
*Lederer, Hollander and Perlman (1968). Only gamma emissions with an external photon yield greater than 5 per cent are listed. f Where a range of dose is given, the lower figure assumes the lowest dose estimate and the smaller quoted mCi dose, and the upper the highest estimated absorbed dose and the higher quoted mCi dose.
328
Review article. Bone scanning
crystal structure does not develop immediately. As first precipitated they are small, imperfect and surrounded by a large hydration shell. Solution and recrystallization occurs spontaneously at a rapid rate (Neuman and Neuman, 1953; Pautard, 1964). The hydroxyapatite crystals present in bone measure 30nmX 30nmx 5 nm. There is, therefore, a large surface-to-volume ratio, high surface charge and large hydration shell. Ions in the surrounding solution can enter this hydration shell, and may then exchange with the layer of hydrated calcium, phosphate and hydroxide ions on the crystal surface. As a result of imperfections within the crystal lattice a further, albeit slower, exchange occurs between the surface of the crystal and its interior (Neuman and Xeuman, 1957). Maturation of bone is associated with perfection of its crystalline structure, leading to a mineral phase which is less hydrated and less readily accessible to ion movements (Heaney and Whedon, 1958). 99 per cent of the mineral in mature compact bone is in this state, and therefore relatively unavailable for exchange with administered radioisotopes (McLean, 1958). The distribution of 45Ca in bone slices incubated in vitro with 45CaCl2 (Dallemagne and Richelle, 1962) is similar to that observed when radioisotope is administered in vivo (Comar, Lotz and Boyd, 1952; Wood et al., 1970) indicating that, even in the absence of blood flow, uptake of radioisotope is affected by local variations in the state of the bone. Under physiological conditions blood supply and mineral turnover are interrelated (Brookes and Helal, 1970), so that it may not be possible to distinguish between their respective contributions to radioisotope uptake.
calcium and strontium are available to pass either into or out of plasma, the extraction ratio may be expected to fall. Measurements over the five minutes after injection have not been able to confirm this (Copp and Shim, 1964).
RADIOPHARMACEUTICALS
1. Calcium and its analogues
Although 47Ca has been used for bone scanning (Corey et al., 1961, 1962; Jasinski, 1962) it is not an ideal radioisotope for imaging because of the high energy of its y emissions, and unavoidable contamination with 45Ca (Table I). Approximately 7 per cent of the calcium in bone mineral can be replaced by strontium in vitro. The rate of exchange is governed in the early stages by the calcium to strontium ratio, but at equilibrium there is discrimination against strontium (Harrison et al., 1959; Schoenberg, 1963). This discrimination also operates in vivo. After the intravenous injection of a bolus of strontium, there is a high extraction— of the order of 70 per cent—on the first passage through bone. On subsequent passages, when both
FIG. 1. Posterior scan started two hours after 4 mCi of 87Srm. There is a lesion in the lower thoracic spine, and probably also one at the lumbosacral junction. The irregular uptake in the upper thoracic spine is associated with the presence of degenerative changes and early osteophyte formation.
329
VOL.
48, No. 569 M. V. Merrick
Strontium is thus taken up rapidly by bone, levels of radioactivity reaching half their maximum value within ten minutes, and 90 per cent within one hour (Charkes and Sklaroff, 1965; Kshirsagar, Lloyd and Vaughan, 1966). The amount in the skeleton represents 20-50 per cent of the administered dose. The highest values are found in children, and in patients with skeletal metastases (Guerin and Guerin, 1960; Cederquist, 1964). Despite the rapid skeletal uptake, plasma clearance (of the fraction not taken up by bone) is relatively slow, 10-20 per cent of the administered dose of either strontium or calcium still being in the circulation at one hour, and 1 per cent at 24 hours (Harrison, Carr and Sutton, 1957; Weber et al., 1969). Excreted strontium appears both in urine and faeces. Most of the urinary excretion occurs within the first 24 hours (Spencer, Laszlo and Brothers, 1957). Faecal excretion is sufficient to cause difficulties in scan interpretation. There are two radioisotopes of strontium with physical characteristics suitable for imaging (Table I). The half-life of 87 Sr m is 2-8 hours. The scan must thus be performed on the day of injection, at an interval which is a compromise between the plasma clearance and radioactive decay. A delay of two to three hours has been suggested (Jasinski et al., 1968). At this time extra-osseous radioactivity is readily detectable (Alexander and Gillespie, 1971). Moreover the high count-rate from non-target tissues reduces any percentage difference between lesion and normal bone, and can conceal abnormalities (Charkes, 1969). The long physical (65-day) and biological (800day) half-life of 85 Sr limits the amount which can be administered (Charkes, Sklaroff and Bierly, 1964). The extended examination times associated with a low7 count-rate usually preclude examination of the entire skeleton. To overcome this difficulty a rapid whole-body scan displayed at reduced size has been suggested as a screening procedure, followed by a slower scan of any areas where the interpretation is in doubt (Braunstein, Hernberg and Chandra, 1971). The alternative is to use a profile or point-counting technique (Bauer and Wendeberg, 1959; Rosenthal, 1965; Papworth and Andrews, 1967; Robillard et al., 1970; Taskinen and Vahatalo, 1971). There is evidence that the latter is the more sensitive (McCready, French and MacDonald, 1970; DeNardo et al., 1967). When patients with symptoms suggestive of metastatic disease are scanned with either of these radioisotopes, deposits are found on the scan in up
to 20 per cent more cases than are visible radiologically (Charkes and Sklaroff, 1964; Erjavec, 1965; Simpson and Orange, 1965; De Nardo and Volpe, 1966; Briggs, 1967; Bessler, 1968; Benoit, Torres and Peterson, 1968; Scott and Adams, 1974). However, about 5 per cent of deposits identified on radiographs are not seen on the scan. This is particularly the case when there is fully mineralized reactive bone (Charkes, Sklaroff and Young, 1966) readily apparent radiologically, but indicating healing following cessation of tumour growth. This is most commonly associated with a favourable response to endocrine therapy, in patients with carcinoma of the prostate or breast. The highest uptakes occur in lesions showing both bone destruction and new bone formation. There is rather less uptake in predominantly lytic deposits (Charkes, Young and Sklaroff, 1968; Charkes, 1972). The poor statistics inevitable with 85Sr are to some extent offset by the high target to background ratios, which are usually maximal one to two weeks after injection. Where there is rapid turnover the peak may occur much earlier (Core}' et al., 1962a). The better statistics obtainable with 87Srm give clearer images. As a result there are some lesions which may be detected using either radioisotope but are not seen with the other (Charkes, 1969; Sauer, 1971). The availability of bone-seeking technetium complexes (q.v. infra) has greatly reduced the use of strontium. Where strontium is still used the most sensitive technique is to perform profile scans two to seven days after the administration of 85Sr, followed by a rectilinear scan of any areas suspicious on the profile. Barium is cleared from the blood much more rapidly than strontium, only 1 per cent of an injected dose remaining in the circulation after three hours (Bauer, Carlsson and Lindquist, 1957; Weber et al., 1969). However, the extraction of barium by bone is slightly lower than that of strontium (Wooton, unpublished), and more is excreted, mainly via the faeces (Harrison et al., 1967). Clinical experience with barium radioisotopes has been restricted by the limited availability and high cost of preparations of sufficiently high specific activity. The clinical results which have been published (Spencer, Lange and Treves, 1971; O'Mara and Subramanian, 1972; Syed, Hosain and Wagner, 1972) have not demonstrated any advantage over more readily available agents. 2. Technetium-labelled compounds
Neither pertechnetate (TcO4~~)northe technetium (Tc 44 ) ion localize to any useful extent in bone,
330
MAY
1975
Review article. Bone scanning
but a number of complexes of technetium with phosphorus-containing compounds are concentrated in the skeleton. The first such substances to be described (Subramanian and McAfee, 1971) were based upon condensed phosphates, compounds formed when orthophosphates are heated: OH OH OH OH O-P-OH
o
+ HO-p-cr. .-o-p-o-p-cr o
o
+H 2 O
o
(pyrophosphate) On further heating this process is repeated to produce chains, which may be of any length, and either straight or branching. The final composition is dependent upon the conditions under which the reaction is carried out. Separation of a particular length chain can be accomplished with difficulty, but on standing slow hydrolysis ultimately results in a mixture of shorter-chain compounds. In vitro condensed phosphates inhibit the deposition from solution of a number of calcium salts, including the orthophosphate, and prevent the conversion of amorphous calcium phosphate to hydroxyapatite. Hydroxyapatite crystals which have been exposed to solutions of condensed phosphates, then washed, dissolve more slowly than untreated crystals. These effects are believed to be the result of replacement of orthophosphate groups on the surface of the crystal or nidus by condensed phosphate, thus altering the charge distribution and reactiveness. Pyrophosphate has been identified as one of the naturally occurring inhibitors of calcium deposition in vivo (Russell and Smith, 1973). Intravenously administered ortho or condensed phosphates, labelled with 32 P, or chelated to labelled metals such as 51Cr (Anghileri and Miller, 1971, 1972; Anghileri, 1972) accumulate to a greater extent in the skeleton than in any other tissue. More radioactivity accumulates in bone when straight chains are given than when branching chain fractions are used. However, in vivo condensed phosphates are rapidly hydrolysed to orthophosphate by phosphatases. The original bone-scanning agents based upon these compounds utilized a short chain of three phosphate residues (tripolyphosphate) (Subramanian and McAfee, 1971). It was subsequently found that more rapid blood clearance and higher bone-tosoft tissue ratios were obtained with a straight chain of 46 residues (PP 46) (Subramanian et al., 1971b). The initial blood clearance in rabbits was more rapid than that of strontium, while levels in most tissues,
including bone, were similar. Technetium levels in liver were appreciably higher. Some workers have reached similar conclusions about optimum chain length (Dewanjee, Fletcher and Davis, 1972), but other groups obtained their best results with pyrophosphate or tripolyphosphate (Boc et al., 1973; Hegesippe et al., 1973; Hopkins, Creighton and Van Deripe, 1973; Hosain et al., 1973; King et al., 1973; McKamey, Artis and Hansen, 1973; Murray et al., 1972; Huberty, Hattner and Powell, 1974). This discrepancy may be due in part to the labile nature of the longer chains (Bowen and Garnett, 1974). However, every batch of every preparation has to be tested finally by a bio-assay as, despite the most meticulous control of their chemical composition, they occasionally fail to localize in bone (Subramanian et al., 1912%). This implies that no chemical test yet devised is specific for the biologically active compound. Thin layer chromatography on hydroxyapatite plates may prove a useful innovation in this respect (Radiochemical Centre, to be published). Before technetium can be labelled to any of the condensed phosphates it must first be reduced from pertechnetate (the form in which it is eluted from its generator) to Tc. Although there are a number of reducing agents which can effect this transformation, bone localization of the technetium occurs only when the stannous ion (Sn 2+ ) or Ti
