- Email: [email protected]
The preparation of dry, monodisperse microspheres of [99mTc]albumin for lung ventilation imaging
0020-708x x2~121423-04$03.00.0 Copyright 0 1982 Pergamon Press Ltd
Illt. J. Appi. Rud~t. lsot. Vol. 33. pp. 1423 to 1426, 1982 Printed in Great dritain. All rights reserved
The Preparation of Dry, Monodisperse Microspheres of [ 99”Tc]Albumin for Lung Ventilation Imaging ALISTAIR WILLIAM
MACKAY JAMES and
Departments
MILLAR,’
HANNAN,’ ROBERT
LESLEY
PETER JOHN
McMILLAN,’
CHARLES
EMMETT3
AITKENj
Royal Infirmary. and ‘Medical Physics & Medical Engineering, Edinburgh EH3 9YW, Scotland and “The Institute of Occupational Medicine, 8 Roxburgh Place. Edinburgh EH8 9SU. Scotland
of ‘Pharmacy
(Receit&
29 Murc,h 1982)
to published methods. 2 pm microspheres of albumin were prepared and labelled with ‘9mTc according However, when the labelling technique was repeated without the inclusion of microspheres, apparent labelling, which was attributed to the formation of a 99mTc tin colloid, was still observed. It was not To establish conditions possible to identify 99mTc tin colloid in the presence of yymTc microspheres. labelling, with and without microspheres. was under which there was no 99mTc tin colloid formation, studied at low pH for a range of SnCl,.2H,O concentrations. A “kit” containing 0.4 mg of microspheres and I pg of SnCI,.ZH,O in 1 ml of 25 mM HCI gave optimum labelling of 88.3:)” (SD 3.9, II = 38). Unlabelled and labelled kits were stable for I and 3 h respectively. Before administration, the microspheres were resuspended in ethanol. m which they were stable for 6 h. None of the preparations contained 99mTctin colloid.
Introduction RADIONUCLIDE ventilation imaging of the lungs is an important diagnostic test in the investigation of suspected pulmonary embolism. The imaging has traditionally been performed using the radioactive gases xenon-133 (133Xe), xenon-127 (12’Xe), or krypton-81m (*lmKr). There are, however, disadvantages with the use of these gases. The 80 keV ;‘-ray emitted by ‘““Xe is too low for good image quality when using a y-camera. ‘2TXe is expensive and not readily available in most countries. The *lmKr generator has a working life of only one day due to the 4.7 h half-life of the “Rb parent and is therefore both expensive and not ideal if the site of use is remote from the site of production. A 99mT~ radiopharmaceutical for ventilation imaging of the lungs is therefore attractive since 99mTc is cheap and readily available in most centres performing radionuclide imaging. The various attempts which have been made at preparing a 99mTc radiopharmaceutical for ventilation imaging fall into two categories: those which involve the nebulisation of a solution of a 99mTc radiopharmaceutical,“*2’ and those which involve the preparation of 99mT~ labelled particles.(3-5’ The major problem with the technique of nebulising a 99mTc solution is the lack of uniformity of particle size obtained since the larger particles may be impacted in
the major airways. Attempts have been made by some workers to remove the large particles by the introduction of a settling bag between the nebuliser and the patient. (lo) The delivery efficiency of such systems is typically low (SSlOo/,) since most of the nebulised ssmTc is associated with large particles which are lost in the settling bag. It is therefore necessary to load the administration equipment with an activity of 99mTc which is up to 20 times that which will be delivered to the patient. Ventilation imaging of the lungs using 99mTc labelled particles has been criticised because of deposition of activity in the major airways’4,“’ and it has been shown that even when using particles of a narrow size distribution, adequate penetrance of the lung periphery is not achieved.‘4’ We believe that this may have been due to differences in size distribution between the particles dispersed in solution. and those inhaled by the patient. When particles are delivered from a saline solution, care must be taken to ensure adequate drying and that any saline residue left after drying does not alter the size of the airborne particles with which it is associated. We therefore decided to study the labelling of 2pm microspheres of human serum albumin with a view to delivering them in the dry state from a suspension in ethanol. This size has been shown to give a reasonably high alveolar deposition with only small losses in the extrathoracic and tracheobronchial regions.‘@
1423
Materials
and Methods
Human serum albumin microspheres, 2pm in dia., were prepared from a 0.2.5’:,, solution of human serum albumin (Kabi) using an air driven spinning top.“’ After production, the microspheres were denatured by heating in olive oil’*’ and the unwanted “satellite” particles were removed by centrifugation. The oil was removed by washing with ether and then the microspheres were prepared as a 0.8 mg/ml suspension in I”,, ethanol in water as required. The microspheres were then checked for size by light microscopy using a Timbre11 double image micrometer (Fleming Instruments Ltd). The full particle size distributions were obtained using a Coulter Counter (Coulter Electronics Ltd). Before use. the microspheres were resuspended by agitation in an ultrasonic bath for 10 min. The microspheres were then labelled by the method Of YbATES rt LI/.(3’ I 52 ml of the 0.8 mg/ml microsphere suspension, 80~11 of a solution of stannous chloride (SnCI,‘2H,O) (5 mg/ml in 1 M HCI) and 0.4 ml of a solution of sodium acetate (100 mg’ml) were mixed in a IO ml vial. To investigate the possibility of the formation of other labelled species, a second vial was prepared containing stannous chloride and sodium acetate as above but the suspension of microspheres replaced by water. Eight millilitrcs of sodium pertechnctate (yymT~) solution (100 MBq) from a ““‘Tc generator was added to each vial and the vials were placed in an ultrasonic bath for 5 min. The labelling efficiency of the contents of each vial was then measured by centrifuging the vial at 2000~ for 10 min. removing 1.0 ml of the supernatant and measuring the activity of ““‘Tc in the 1.O ml sample and the 9.0 ml remaining in the vial. Measurements of activity were made with a radioisotope dose calibrator (Pitman Model 238). The activity in the IOml of supernatant and hence the labelling efficiency were then calculated. When using the above technique. we observed a high labelling efficiency in the absence of microspheres, probably due to the formation of a ““‘Tc tin colloid. This Icd us to perform the following study of labelling at a lower pH. Twenty milligrammes of stannous chloride was dissolved in I ml of 5 M HCI then diluted to 100 ml with water. Dilutions of this solution were made with 50 mM HCl to give stannous chloride solutions ranging in concentration from 100 to 0.01 b(g 0.5 ml. A 0.5 ml aliquot of each stannous chloride solution \vas mixed with a 0.5 ml aliquot of the microsphere suspension (0.X mg:ml) in a IO ml vial. 0.5 ml of sodium pertcchnetate (““‘Tc) solution (50 MBq) was added to each vial and the vials incubated at room temperature for 20 min. A set of vials in which the 0.5 ml of microsphere suspension had been replaced by 0.5 ml of nater were processed in an identical manner. To measure the Inbelling etficicncy. X.5 ml of 50 mM HCI was added to each vial then the vials were ccntrifugcd and sampled as before. The above experiment demonstrated that optimum labclling was achieved v,ith 1pg of stannous chloride
and 0.4mg of microspheres in 1 ml of 25 mM HCI. For all subsequent work therefore, “kits” containing this formula were prepared. To study their shelf-life. a batch of eight such kits was prepared along with ;I batch of eight kits containing only 1 pg of stannous chloride in 1 ml of 25 mM HCI. Immediately after preparation. 0.5 ml of sodium pertechnetate (yymTc) solution (50MBq) was added to one of each type of kit. The vials were then incubated, diluted, centrifuged and sampled as before. and the efficiency of labelling was calculated. This procedure was repeated at hourly intervals up to 6 h after preparation of the kits. TO study the stability of the ““‘Tc labelled microspheres. batches of seven kits. with and without miccrospheres. were prepared as above. Immediately after preparation. 0.5 ml of sodium pertechnetate (“““‘Tc) solution (50 MBq) was added to each of the 14 kits. After 20 min incubation, one of each type of kit was diluted, centrifuged and sampled as before. and the efficiency of labelling was calculated. This procedure was repeated at hourly intervals up to 6 h after the addition of the yymTc. Since we intended to deliver the labelled microspheres to the patient from a suspension in ethanol. the stability of the labelled microspheres in ethanol was studied. A kit was labelled with y9”‘Tc as described in the above stability studies except that after centrifuging, all the supernatant was removed and 10 ml of ethanol was injected into the vial. The contents of the vial were mixed by vigorous shaking. centrifuged. sampled as before. and the labelling cfIiciency was calculated. The I.0 ml sample of supcrnatant was then returned to the vial and the contents of the vial were mixed by shaking. The process was repeated at hourly intervals up to 6 h after the addition of the ethanol. All experiments were performed on two separate occassions and all solutions and vials used wcrc purged with oxygen-free nitrogen.
Results and Discussion The labelling etticiency of the ““‘Tc microsphcrcs prepared by the method of YFATIS c’t trl.‘.” was greater than 99”,. However. when the procedure was repeated in the absence of microspheres. the labelling eficiency was apparently still greater than 99”,,. The pH of these mixtures was measured as 4.7. It has been shonn previously that the “““‘Tc in stannous chloride sodium pertechnetate (“““‘Tc) mixtures, exists as a souble species below pH 1.5 and above pH IO. but labelled to ;I colloid at pH values of between 2.5 and IO.‘“’ and that this colloid makes the solution cloudy. It is therefore likely that a tin colloid was present in the kit of YI Art:s c’t trl.‘“’ but its prcscncc \+as masked by the cloudiness imparted to the solution b! the microspheres. Therefore, using this method for the prcparation of ““mTc microspheres. it is possible that the 9“mTc was not only labelled to the microspheres hut also to the tin colloid and that subsequent pulmonar!
I425
Labelllng
Labellng
efflclency
eff1cwlcy
(%I
(%I
I 0 01
L
01
1 SnCl2.2H20
10
0
100
1
2
3
Tjme
lpgl
4
after
5
preparation
6
(hrsi
FIG. 1. The effect of stannous chloride concentration on: the ““‘Tc labelling of albumin microspheres (0) and. the 9ymT~ labelling in the absence of microspheres apparent (0). All measurements were made at pH < 2. Each point is the mean of two measurements.
FIG. 2. The effect of time on the stability of: the kit containing I pg stannous chloride and 0.4 mg of albumin microspheres in I ml of 25 mM HCI (0) and. the kit containing only 1 /lg stannous chloride in I ml of 25 mM HCI (0).
investigations”” were performed after the inhalation of this mixture of unknown size distribution. Since the alveolar deposition of inhaled particles is a critical function of particle size,“’ the presence of a 99mT~ tin colloid could result in large impaction losses in the major airways. It is technically difficult to demonstrate whether the 99mTc in ““‘Tc labelled microspheres is present as 94”‘Tc tin colloid or a ““‘Tc labelled microspheres. mixture of both. Therefore, to eliminate the possibility of having a tin colloid present in the microsphere solution which was to be labelled with 99mTc. we decided to perform all our subsequent labelling experiments at low pH by using a solution of stannous chloride in 50 mM HCL To demonstrate the absence of 9ymTc tin colloids. all labelling experiments were performed with and without microspheres. Figure I shows the effect of different concentrations of stannous chloride on the efficiency of labelling of the microspheres. It can be seen that a reasonably high labelling efficiency was obtained when between 0.5 p(g and 5 pg of stannous chloride was used. but that labelling fell off rapidly outside these limits. A labelling efficiency of 90.2:,, was obtained with I pg of stannous chloride. This was therefore chosen as the amount of stannous chloride required for optimum labelling. At no concentration of stannous chloride was any appreciable amount of 99mTc tin colloid present. For all subsequent experiments, kits containing 0.5 ml of the microsphere suspension (0.8 mg/ml) and 0.5 ml of stannous chloride (1 pg/O.S ml of 50 mM HCI) were used. The pH of these kits was less than 2.0. To date. 38 kits with this formula have been labelled, giving an average labelling efficiency of 88.3”/, (SD 3.9). The results of the tests on the stability of this kit are shown in Fig. 2. At 0 and 1 h after preparation,
the labelling efficiencies were 87.3”,, and 82.6”,, respectively but beyond I h Isbelling uas much reduced. At no time was any appreciable amount of “‘“‘Tc tin colloid present. One hour after preparation uas therefore chosen as the shelf-life of the unlabellcd kit. Figure 3 shows the results of the measurements on the stability of the labelled kit. Three hours after labelling, the labelling efficiency was 90.4”,, but fell thereafter. Three hours was therefore chosen as the shelflife of the labelled microspheres in aqueous solution. Figure 3 also shows that the labelling efficiency of the 99mTc microspheres, once resuspended in ethanol. remained greater than 94.0”,, over a period of 6 h. Ethanol was chosen as the delivery medium due to its high volatility and low toxicity. A high volatility is
Each point is the mean of two measurements.
60 Labelllng efficiency WI
4O !
L 0
1
2 Trne
1
L
3
4
after
preparation
I 5
6
(hrsl
FIG. 3. The effect of time on: the stability of the ““Tc microspheres in HCl (0). the formation of ‘““‘Tc tin colloid in HCI in the absence of microspheres (0) and. the stability of the 99”Tc microspheres once resuspended in ethanol (A). Each point is the mean of two measurements
essential if the previously discussed problems associated with the use of saline as the medium for final dispersal are to be avoided. Before administration to patients. the 94mT~ microspheres were washed and then resuspended in ethanol as follows: the labelled kit was centrifuged as above and the supernatant removed. 10 ml of ethanol was added to the vial which was vigorously shaken then centrifuged. The supernatant was removed and the labelled microspheres were finally resuspended in 0.3 ml of ethanol. This suspension was then dispersed into the administration apparatus by an air jet nebulizer and the ethanol droplets quickly evaporated leaving the dry airborne microspheres. It has been shown that this amount of ethanol does not affect respiratory function when measured by standard lung function tests.“” The washing in ethanol was most important since it ensured that once the ethanol had evaporated in the administration apparatus, there would be no water droplets associated with the airborne microspheres. Removal of aqueous supernatant and washing in cthanol means that the relatively poor labelling efficiency of the microspheres in the aqueous kit is less important than it might be with other 99mTc radiopharmaceuticals since the unlabelled yymTc is removed before the microspheres are administered to the patient. have Clinical results with the 99”Tc-microspheres been extremely encouraging, with minimal deposition of activity in the major airways and good penetrance to the lung periphery. The apparatus used to adminis-
ter the microspheres and the clinical results of a comparison with S’mKr are described elsewhere.“” Acknow/edgrmcnt.\-The authors wish to acknowledge partial funding of this project by the European Coal Steel Community and the U.K. National Coal Board.
the and
References I. H~vris M. for TAPLIN G. V. Sunri~. Nuci. Med. 10, 243 (1980). ~KA~(.IS R. A.. AGNEW J. E.. SLTTOU P. P.. PAVIA D. and CLARKE S. W. 5u(~l. ,L!&. Conu?~ur~.2. X3 ( IYXI ). YEATES D. B.. WARBI(.K A. and ASPIC N. Iut. J. Appi.
Radiut. lsor. 25, 518 (1974). A. P.. MINIATI M. and Phj,siopcrrh. Resp. 16, 287 (1980).
GKEENING
6. 7.
FA~IO F. &r/i. tlrr,.
FAZIO F.. SANTOLI(.AXDKO A.. SOI.FA~TLLI S., PALL,\ A.. FORXAI E. and GI~NTIXI C. Br. J. Radio/. (Spwitrl ,Vo.) Report 15 (Ed. LAVENDER J. P.) 130 (1978). STAHLH~F~N W., G~BHART J. and Hturx:~ J. Am. (ut/. Hyq. Assoc. J. 33. 237 (1972). WALTON W. H. and PR~WFTT W. C. Pm. Phrs. SOCK.
(London) B62, 341 (1949). ALIIRICH J. E. and JOHNSOK J. R. Iur. .I. 4ppI. Rodrtrl. Isot. 25. IS (1974). 9. CALLAHAN R. J. and CASTRONO~O F. P. Anlcr. J. Ho.tp, Plzurr~~.30, 614 (1973). IO. Y~ATFS D. B.. ASPIN N., Lrv~so~ H.. JONES M. T. and BRYAN A. C. J. Appl. Ph_wio/. 39. 487 (1975). Il. EMMETT P. C. Ph.D. Thesis, Universit] of Edinburgh (1980). 12. HANKAX W. J., EMMETT P. C.. AITKLN R. J.. Love R. G.. MILLAR A. M. and MUIR A. L. J. !~‘wi. !\led. accepted for publication (1982). 8.
Illt. J. Appi. Rud~t. lsot. Vol. 33. pp. 1423 to 1426, 1982 Printed in Great dritain. All rights reserved
The Preparation of Dry, Monodisperse Microspheres of [ 99”Tc]Albumin for Lung Ventilation Imaging ALISTAIR WILLIAM
MACKAY JAMES and
Departments
MILLAR,’
HANNAN,’ ROBERT
LESLEY
PETER JOHN
McMILLAN,’
CHARLES
EMMETT3
AITKENj
Royal Infirmary. and ‘Medical Physics & Medical Engineering, Edinburgh EH3 9YW, Scotland and “The Institute of Occupational Medicine, 8 Roxburgh Place. Edinburgh EH8 9SU. Scotland
of ‘Pharmacy
(Receit&
29 Murc,h 1982)
to published methods. 2 pm microspheres of albumin were prepared and labelled with ‘9mTc according However, when the labelling technique was repeated without the inclusion of microspheres, apparent labelling, which was attributed to the formation of a 99mTc tin colloid, was still observed. It was not To establish conditions possible to identify 99mTc tin colloid in the presence of yymTc microspheres. labelling, with and without microspheres. was under which there was no 99mTc tin colloid formation, studied at low pH for a range of SnCl,.2H,O concentrations. A “kit” containing 0.4 mg of microspheres and I pg of SnCI,.ZH,O in 1 ml of 25 mM HCI gave optimum labelling of 88.3:)” (SD 3.9, II = 38). Unlabelled and labelled kits were stable for I and 3 h respectively. Before administration, the microspheres were resuspended in ethanol. m which they were stable for 6 h. None of the preparations contained 99mTctin colloid.
Introduction RADIONUCLIDE ventilation imaging of the lungs is an important diagnostic test in the investigation of suspected pulmonary embolism. The imaging has traditionally been performed using the radioactive gases xenon-133 (133Xe), xenon-127 (12’Xe), or krypton-81m (*lmKr). There are, however, disadvantages with the use of these gases. The 80 keV ;‘-ray emitted by ‘““Xe is too low for good image quality when using a y-camera. ‘2TXe is expensive and not readily available in most countries. The *lmKr generator has a working life of only one day due to the 4.7 h half-life of the “Rb parent and is therefore both expensive and not ideal if the site of use is remote from the site of production. A 99mT~ radiopharmaceutical for ventilation imaging of the lungs is therefore attractive since 99mTc is cheap and readily available in most centres performing radionuclide imaging. The various attempts which have been made at preparing a 99mTc radiopharmaceutical for ventilation imaging fall into two categories: those which involve the nebulisation of a solution of a 99mTc radiopharmaceutical,“*2’ and those which involve the preparation of 99mT~ labelled particles.(3-5’ The major problem with the technique of nebulising a 99mTc solution is the lack of uniformity of particle size obtained since the larger particles may be impacted in
the major airways. Attempts have been made by some workers to remove the large particles by the introduction of a settling bag between the nebuliser and the patient. (lo) The delivery efficiency of such systems is typically low (SSlOo/,) since most of the nebulised ssmTc is associated with large particles which are lost in the settling bag. It is therefore necessary to load the administration equipment with an activity of 99mTc which is up to 20 times that which will be delivered to the patient. Ventilation imaging of the lungs using 99mTc labelled particles has been criticised because of deposition of activity in the major airways’4,“’ and it has been shown that even when using particles of a narrow size distribution, adequate penetrance of the lung periphery is not achieved.‘4’ We believe that this may have been due to differences in size distribution between the particles dispersed in solution. and those inhaled by the patient. When particles are delivered from a saline solution, care must be taken to ensure adequate drying and that any saline residue left after drying does not alter the size of the airborne particles with which it is associated. We therefore decided to study the labelling of 2pm microspheres of human serum albumin with a view to delivering them in the dry state from a suspension in ethanol. This size has been shown to give a reasonably high alveolar deposition with only small losses in the extrathoracic and tracheobronchial regions.‘@
1423
Materials
and Methods
Human serum albumin microspheres, 2pm in dia., were prepared from a 0.2.5’:,, solution of human serum albumin (Kabi) using an air driven spinning top.“’ After production, the microspheres were denatured by heating in olive oil’*’ and the unwanted “satellite” particles were removed by centrifugation. The oil was removed by washing with ether and then the microspheres were prepared as a 0.8 mg/ml suspension in I”,, ethanol in water as required. The microspheres were then checked for size by light microscopy using a Timbre11 double image micrometer (Fleming Instruments Ltd). The full particle size distributions were obtained using a Coulter Counter (Coulter Electronics Ltd). Before use. the microspheres were resuspended by agitation in an ultrasonic bath for 10 min. The microspheres were then labelled by the method Of YbATES rt LI/.(3’ I 52 ml of the 0.8 mg/ml microsphere suspension, 80~11 of a solution of stannous chloride (SnCI,‘2H,O) (5 mg/ml in 1 M HCI) and 0.4 ml of a solution of sodium acetate (100 mg’ml) were mixed in a IO ml vial. To investigate the possibility of the formation of other labelled species, a second vial was prepared containing stannous chloride and sodium acetate as above but the suspension of microspheres replaced by water. Eight millilitrcs of sodium pertechnctate (yymT~) solution (100 MBq) from a ““‘Tc generator was added to each vial and the vials were placed in an ultrasonic bath for 5 min. The labelling efficiency of the contents of each vial was then measured by centrifuging the vial at 2000~ for 10 min. removing 1.0 ml of the supernatant and measuring the activity of ““‘Tc in the 1.O ml sample and the 9.0 ml remaining in the vial. Measurements of activity were made with a radioisotope dose calibrator (Pitman Model 238). The activity in the IOml of supernatant and hence the labelling efficiency were then calculated. When using the above technique. we observed a high labelling efficiency in the absence of microspheres, probably due to the formation of a ““‘Tc tin colloid. This Icd us to perform the following study of labelling at a lower pH. Twenty milligrammes of stannous chloride was dissolved in I ml of 5 M HCI then diluted to 100 ml with water. Dilutions of this solution were made with 50 mM HCl to give stannous chloride solutions ranging in concentration from 100 to 0.01 b(g 0.5 ml. A 0.5 ml aliquot of each stannous chloride solution \vas mixed with a 0.5 ml aliquot of the microsphere suspension (0.X mg:ml) in a IO ml vial. 0.5 ml of sodium pertcchnetate (““‘Tc) solution (50 MBq) was added to each vial and the vials incubated at room temperature for 20 min. A set of vials in which the 0.5 ml of microsphere suspension had been replaced by 0.5 ml of nater were processed in an identical manner. To measure the Inbelling etficicncy. X.5 ml of 50 mM HCI was added to each vial then the vials were ccntrifugcd and sampled as before. The above experiment demonstrated that optimum labclling was achieved v,ith 1pg of stannous chloride
and 0.4mg of microspheres in 1 ml of 25 mM HCI. For all subsequent work therefore, “kits” containing this formula were prepared. To study their shelf-life. a batch of eight such kits was prepared along with ;I batch of eight kits containing only 1 pg of stannous chloride in 1 ml of 25 mM HCI. Immediately after preparation. 0.5 ml of sodium pertechnetate (yymTc) solution (50MBq) was added to one of each type of kit. The vials were then incubated, diluted, centrifuged and sampled as before. and the efficiency of labelling was calculated. This procedure was repeated at hourly intervals up to 6 h after preparation of the kits. TO study the stability of the ““‘Tc labelled microspheres. batches of seven kits. with and without miccrospheres. were prepared as above. Immediately after preparation. 0.5 ml of sodium pertechnetate (“““‘Tc) solution (50 MBq) was added to each of the 14 kits. After 20 min incubation, one of each type of kit was diluted, centrifuged and sampled as before. and the efficiency of labelling was calculated. This procedure was repeated at hourly intervals up to 6 h after the addition of the yymTc. Since we intended to deliver the labelled microspheres to the patient from a suspension in ethanol. the stability of the labelled microspheres in ethanol was studied. A kit was labelled with y9”‘Tc as described in the above stability studies except that after centrifuging, all the supernatant was removed and 10 ml of ethanol was injected into the vial. The contents of the vial were mixed by vigorous shaking. centrifuged. sampled as before. and the labelling cfIiciency was calculated. The I.0 ml sample of supcrnatant was then returned to the vial and the contents of the vial were mixed by shaking. The process was repeated at hourly intervals up to 6 h after the addition of the ethanol. All experiments were performed on two separate occassions and all solutions and vials used wcrc purged with oxygen-free nitrogen.
Results and Discussion The labelling etticiency of the ““‘Tc microsphcrcs prepared by the method of YFATIS c’t trl.‘.” was greater than 99”,. However. when the procedure was repeated in the absence of microspheres. the labelling eficiency was apparently still greater than 99”,,. The pH of these mixtures was measured as 4.7. It has been shonn previously that the “““‘Tc in stannous chloride sodium pertechnetate (“““‘Tc) mixtures, exists as a souble species below pH 1.5 and above pH IO. but labelled to ;I colloid at pH values of between 2.5 and IO.‘“’ and that this colloid makes the solution cloudy. It is therefore likely that a tin colloid was present in the kit of YI Art:s c’t trl.‘“’ but its prcscncc \+as masked by the cloudiness imparted to the solution b! the microspheres. Therefore, using this method for the prcparation of ““mTc microspheres. it is possible that the 9“mTc was not only labelled to the microspheres hut also to the tin colloid and that subsequent pulmonar!
I425
Labelllng
Labellng
efflclency
eff1cwlcy
(%I
(%I
I 0 01
L
01
1 SnCl2.2H20
10
0
100
1
2
3
Tjme
lpgl
4
after
5
preparation
6
(hrsi
FIG. 1. The effect of stannous chloride concentration on: the ““‘Tc labelling of albumin microspheres (0) and. the 9ymT~ labelling in the absence of microspheres apparent (0). All measurements were made at pH < 2. Each point is the mean of two measurements.
FIG. 2. The effect of time on the stability of: the kit containing I pg stannous chloride and 0.4 mg of albumin microspheres in I ml of 25 mM HCI (0) and. the kit containing only 1 /lg stannous chloride in I ml of 25 mM HCI (0).
investigations”” were performed after the inhalation of this mixture of unknown size distribution. Since the alveolar deposition of inhaled particles is a critical function of particle size,“’ the presence of a 99mT~ tin colloid could result in large impaction losses in the major airways. It is technically difficult to demonstrate whether the 99mTc in ““‘Tc labelled microspheres is present as 94”‘Tc tin colloid or a ““‘Tc labelled microspheres. mixture of both. Therefore, to eliminate the possibility of having a tin colloid present in the microsphere solution which was to be labelled with 99mTc. we decided to perform all our subsequent labelling experiments at low pH by using a solution of stannous chloride in 50 mM HCL To demonstrate the absence of 9ymTc tin colloids. all labelling experiments were performed with and without microspheres. Figure I shows the effect of different concentrations of stannous chloride on the efficiency of labelling of the microspheres. It can be seen that a reasonably high labelling efficiency was obtained when between 0.5 p(g and 5 pg of stannous chloride was used. but that labelling fell off rapidly outside these limits. A labelling efficiency of 90.2:,, was obtained with I pg of stannous chloride. This was therefore chosen as the amount of stannous chloride required for optimum labelling. At no concentration of stannous chloride was any appreciable amount of 99mTc tin colloid present. For all subsequent experiments, kits containing 0.5 ml of the microsphere suspension (0.8 mg/ml) and 0.5 ml of stannous chloride (1 pg/O.S ml of 50 mM HCI) were used. The pH of these kits was less than 2.0. To date. 38 kits with this formula have been labelled, giving an average labelling efficiency of 88.3”/, (SD 3.9). The results of the tests on the stability of this kit are shown in Fig. 2. At 0 and 1 h after preparation,
the labelling efficiencies were 87.3”,, and 82.6”,, respectively but beyond I h Isbelling uas much reduced. At no time was any appreciable amount of “‘“‘Tc tin colloid present. One hour after preparation uas therefore chosen as the shelf-life of the unlabellcd kit. Figure 3 shows the results of the measurements on the stability of the labelled kit. Three hours after labelling, the labelling efficiency was 90.4”,, but fell thereafter. Three hours was therefore chosen as the shelflife of the labelled microspheres in aqueous solution. Figure 3 also shows that the labelling efficiency of the 99mTc microspheres, once resuspended in ethanol. remained greater than 94.0”,, over a period of 6 h. Ethanol was chosen as the delivery medium due to its high volatility and low toxicity. A high volatility is
Each point is the mean of two measurements.
60 Labelllng efficiency WI
4O !
L 0
1
2 Trne
1
L
3
4
after
preparation
I 5
6
(hrsl
FIG. 3. The effect of time on: the stability of the ““Tc microspheres in HCl (0). the formation of ‘““‘Tc tin colloid in HCI in the absence of microspheres (0) and. the stability of the 99”Tc microspheres once resuspended in ethanol (A). Each point is the mean of two measurements
essential if the previously discussed problems associated with the use of saline as the medium for final dispersal are to be avoided. Before administration to patients. the 94mT~ microspheres were washed and then resuspended in ethanol as follows: the labelled kit was centrifuged as above and the supernatant removed. 10 ml of ethanol was added to the vial which was vigorously shaken then centrifuged. The supernatant was removed and the labelled microspheres were finally resuspended in 0.3 ml of ethanol. This suspension was then dispersed into the administration apparatus by an air jet nebulizer and the ethanol droplets quickly evaporated leaving the dry airborne microspheres. It has been shown that this amount of ethanol does not affect respiratory function when measured by standard lung function tests.“” The washing in ethanol was most important since it ensured that once the ethanol had evaporated in the administration apparatus, there would be no water droplets associated with the airborne microspheres. Removal of aqueous supernatant and washing in cthanol means that the relatively poor labelling efficiency of the microspheres in the aqueous kit is less important than it might be with other 99mTc radiopharmaceuticals since the unlabelled yymTc is removed before the microspheres are administered to the patient. have Clinical results with the 99”Tc-microspheres been extremely encouraging, with minimal deposition of activity in the major airways and good penetrance to the lung periphery. The apparatus used to adminis-
ter the microspheres and the clinical results of a comparison with S’mKr are described elsewhere.“” Acknow/edgrmcnt.\-The authors wish to acknowledge partial funding of this project by the European Coal Steel Community and the U.K. National Coal Board.
the and
References I. H~vris M. for TAPLIN G. V. Sunri~. Nuci. Med. 10, 243 (1980). ~KA~(.IS R. A.. AGNEW J. E.. SLTTOU P. P.. PAVIA D. and CLARKE S. W. 5u(~l. ,L!&. Conu?~ur~.2. X3 ( IYXI ). YEATES D. B.. WARBI(.K A. and ASPIC N. Iut. J. Appi.
Radiut. lsor. 25, 518 (1974). A. P.. MINIATI M. and Phj,siopcrrh. Resp. 16, 287 (1980).
GKEENING
6. 7.
FA~IO F. &r/i. tlrr,.
FAZIO F.. SANTOLI(.AXDKO A.. SOI.FA~TLLI S., PALL,\ A.. FORXAI E. and GI~NTIXI C. Br. J. Radio/. (Spwitrl ,Vo.) Report 15 (Ed. LAVENDER J. P.) 130 (1978). STAHLH~F~N W., G~BHART J. and Hturx:~ J. Am. (ut/. Hyq. Assoc. J. 33. 237 (1972). WALTON W. H. and PR~WFTT W. C. Pm. Phrs. SOCK.
(London) B62, 341 (1949). ALIIRICH J. E. and JOHNSOK J. R. Iur. .I. 4ppI. Rodrtrl. Isot. 25. IS (1974). 9. CALLAHAN R. J. and CASTRONO~O F. P. Anlcr. J. Ho.tp, Plzurr~~.30, 614 (1973). IO. Y~ATFS D. B.. ASPIN N., Lrv~so~ H.. JONES M. T. and BRYAN A. C. J. Appl. Ph_wio/. 39. 487 (1975). Il. EMMETT P. C. Ph.D. Thesis, Universit] of Edinburgh (1980). 12. HANKAX W. J., EMMETT P. C.. AITKLN R. J.. Love R. G.. MILLAR A. M. and MUIR A. L. J. !~‘wi. !\led. accepted for publication (1982). 8.