Patient self-attenuation and technologist dose in positron emission tomography

Transcription

1 Patient self-attenuation and technologist dose in positron emission tomography Benjamin W. Zeff a and Michael V. Yester Department of Radiology, University of Alabama at Birmingham, Birmingham, Alabama Received 6 August 2004; revised 4 January 2005; accepted for publication 13 January 2005; published 16 March 2005 Positron emission tomography PET, with 511-keV radiation and long patient-uptake times, presents unique radiation safety concerns. This two-part study considers aspects of PET radiation safety as they relate to PET suite design, dose to the public, and technologist occupational dose. In the first part of the study, the self-attenuation of radiation by patients bodies was quantified. The radiation exposure was measured at three positions from 64 patients injected with fluorine-18 fluorodeoxyglucose FDG during the uptake period. Compared with an in vitro control used as a point source, a significant decrease in exposure 40% at 1 m was observed due to nonuniform distribution of FDG and attenuation within the patients. The attenuation data are consistent with results from simulations M. E. Phelps, Comments and Perspectives, J. Nucl. Med. 45, that treat the body as a uniform, water-filled cylinder. As distance is often the principal source of protection for 511-keV radiation, the considerable self-attenuation may allow for more compact PET suites. However, despite high patient self-attenuation, shielding, and standard precautionary measures, PET technologist occupational doses can remain quite high 12 msv/year. The second part of this study tracked the daily dose received by PET technologists. Close technologist-patient interaction both during and following FDG administration, as much as 20 min/study, contribute to the high doses and point to the need for a more innovative approach to radiation protection for PET technologists American Association of Physicists in Medicine. DOI: / I. INTRODUCTION Positron emission tomography PET has become an indispensable diagnostic tool for the detection and monitoring of many cancers and other diseases. Within the past five years, the number of PET procedures performed in the U.S. has skyrocketed to around one million scans annually. 1 The number of PET scanners in use is increasing accordingly. Whether these scanners are placed in new custom suites or facilities are carved out of existing space, careful attention must be given to radiation safety in the design process. 2 5 At our PET facility, fluorine-18 fluorodeoxyglucose FDG T 1/2 =110 min for 18 F, a positron-emitting glucose analogue, is the primary radiopharmaceutical used. Patients are injected with an average of Mbq mci of FDG, after which they rest in one of two patient prep rooms to allow uptake and equilibration of the FDG. At the end of the uptake period, the patients are taken to the restroom to empty their bladder and then to the scanner. Once the FDG is administered, a patient becomes a radiation source, emitting 511-keV gamma rays. During the uptake period and the scan, technologists and others working in or near the PET suite must be sufficiently protected from radiation from the patient. The use of 511-keV photons for imaging necessitates a different type of radiation protection than is needed for most other nuclear medicine and x-ray procedures keV photons have a half-value layer in lead 3 of 4.1 mm for a narrow-beam geometry; for a broadly distributed radiation source such as a patient, the half-value layer is closer to 5 mm. Up to an inch of lead may be required to meet regulatory limits for radiation exposure in uncontrolled areas, so the weight and expense of lead shielding can be prohibitive. Another means of limiting exposure is the increase the distance from the radioactive source. While protecting the public and technologists is a key factor in PET suite design, estimating the exposure rate from patients is not straightforward. The simplest model is to treat the patient as a decaying, radiating point source. Certainly, such a model offers an upper limit to the exposure rates. In a real system, the exposure rate is reduced, perhaps significantly, from that upper limit. One such factor is the distribution of FDG in the patient s body in the heart, bladder, and elsewhere, which causes the exposure rate to diverge from that of a point source, especially near a patient. A second factor is the absorption and scatter of 511-keV radiation by the patient s body. Body habitus differs greatly between patients and strongly affects absorption and scatter. As it is difficult to decouple these factors, we refer to the decrease in the radiation exposure due to the combination of these factors simply as self-attenuation. In measurements from a small sample of patients, Kearfott et al. 6 found that point approximations overestimate instantaneous exposure rates by a factor of at a 1-meter distance from the patients lying on the examination table. Benetar et al. measured instantaneous dose rates from 115 patients immediately after injection, 7 though they did not attempt to quantify patient self-attenuation. Their data suggest only a very slight effect. Cronin et al. measured dose rates from 75 patients two hours after the injection of FDG. 8 In another study White et al. 9 looked at dose rates at four distances from PET patients after 861 Med. Phys. 32 4, April /2005/32 4 /861/5/$ Am. Assoc. Phys. Med. 861

2 862 B. W. Zeff and M. V. Yester: Self-attenuation and public and technologist dose in PET 862 the uptake period; their data reflect a decrease in exposure rate of up to 38% relative to a point source at 0.5 m. In simulations that treated the body as a uniform water cylinder, Courtney et al. 2 estimated the effects of both radiopharmaceutical distribution and attenuation, with a reduction of the dose rate by a factor of 1.6 at 1 m. In order to better quantify the exposure that the public and technologists may get from patients, we have measured the exposure not the exposure rate for 64 patients over the full uptake period at three distances from the patient. The self-attenuation data were compared with the simulation results from Courtney et al.. The high self-attenuation factors measured have significance in the determining the necessary distance and shielding in a PET suite. Due to the high self-attenuation, public areas adjacent to the patient prep rooms would require less shielding. Distance, shielding, and patient self-attenuation notwithstanding, PET technologist doses can remain high. To understand the reasons, we followed up with a second series of measurements. We tracked the exposure received by the technologists during individual procedures as well as daily averages. In doing so, we were able to pinpoint highexposure tasks during each procedure. This data, together with the self-attenuation data, allow for improved radiation safety design and procedures for PET. II. METHODS A. Patient self-attenuation The PET- computed tomography CT suite at our institution has two patient prep rooms, a control and reading room, a scan room, and a hot lab for preparation of the radiopharmaceuticals for injection. During the uptake period, patients rest in a recliner with the lights dimmed to minimize motion and disturbances. The prep rooms are located across a small corridor from the control and reading room to make best use of distance as a radiation safety tool. This layout effectively reduces the exposure in the control and reading room to well within regulatory limits. Hence, the walls of the patient prep rooms that are within the suite are not lead lined. However, the walls adjacent to work areas outside the PET suite are lined with 5/8 in. lead shielding, a costly and difficult construction job that highlights the need for a better understanding of exposure from PET patients. Measurements of radiation exposure were made using two Rados Rad-60R MGP Instruments, Inc., Smyrna, GA silicon diode personal dosimeters affixed at three locations in and around one room two within the patient prep room and one outside the prep room Fig. 1. Initially, measurements were made at two locations: location 1 was 244 cm from patients sternum towards the feet, and location 2 was 112 cm away, laterally. Additional measurements were made at a third location 84 cm, immediately outside the prep room, chosen to give a more truly lateral measurement. Care was taken to make sure that this latter dosimeter was not located behind a stud. Two sheets of gypsum board have negligible attenuation, so no correction was applied to the measurements made through the wall. FIG. 1. Layout of the patient preparation room and dosimeter locations. The dosimeters were attached at three distances from the reclining patient measured from the sternum, marked with an x, d 1 =244 cm, d 2 =112 cm, and d 3 =84 cm. The PET/CT control room is across a corridor on the side of the room with dosimeter location 2. The bottom wall separates the prep room from a noncontrolled work area, and the right wall separates the two prep rooms. At the time of this study, an average of about six patient scans were performed each day. The two prep rooms are used alternately to increase patient throughput, so data were collected on only a few patients per day. The dosimeters were turned on in the morning, before any radioactivity was administered to patients. Background measurements were made to determine other sources of exposure readings on the dosimeters. The dosimeters were used in integrate mode and were not reset throughout the day. Instead, technologists recorded the exposure reading when any of the following occurred: a patient in either room was injected with FDG, a patient from either room was taken out to the restroom or for the scan after the uptake period, or a patient returned to the room after a scan. Additional readings were often recorded during the uptake period. For each reading, the time, the integrated exposure, and the occasion e.g., FDG administered to patient in room 1 in. were recorded. Those data were combined with data on patients weight, height, and radioactivity of FDG administered. To compare the exposures from actual patients with that from an unattenuated point source, we made a control measurement at locations 1 and 2. A syringe of known radioactivity of FDG was affixed to the recliner at chest level and the exposure and measurement duration were recorded. Exposure data were recorded at dosimeter locations 1 and 2 for 50 patients. Measurements of exposure for another 14 patients were taken at location 3. The total exposure, not the exposure rate, was recorded, and an average exposure was calculated for the time that the patient was in the recliner. B. PET technologist dose In terms of PET technologist radiation safety, the second part of this study is perhaps of greater consequence than the first. The shielding in the PET suite in which these measurements were taken was based on point-source calculations. A series of measurements using radiation survey meters and personal dosimeters have confirmed that the shielding calculations were, indeed, conservative. That is, within the control room, the dose received by the technologists is well below regulatory limits. The shielding and layout of the suite was

3 863 B. W. Zeff and M. V. Yester: Self-attenuation and public and technologist dose in PET 863 FIG. 2. Probability distributions of self-attenuation for 50 patients at dosimeter locations 1 and 2. The selfattenuation is the percent decrease of the measured exposure from a patient source from that expected for a point-source approximation. A control measurement made with an unshielded syringe yielded SA=0 at both locations. The mean values of self-attenuation are SA 1 =28% and SA 2 =38%. The considerable spread around the mean self-attenuation values is due to variations in patient size, positioning, and radiopharmaceutical distribution. Error in individual patient measurements is due to the dosimeter accuracy and the accuracy of recorded measurement times and does not affect the mean self-attenuation values or standard deviation. designed to limit technologist exposure to 10 Sv/week from patient sources. According to radiation badge readings, however, the technologists could receive this much dose in only one day considering all tasks. While a number of papers have considered dose rates in PET imaging, few have directly related these measurements to actual technologist doses. To get a better picture of the source of the high technologist doses, two sets of measurements were made. First, each technologist was assigned one of the personal dosimeters and asked to record the exposure reading at the beginning and ending of each day over a 17-day period, corresponding to 75 patient studies. Additionally, the number of studies performed each day was recorded. A soft-tissue conversion was used to convert the dosimeter readings from exposure R to equivalent dose Sv for comparison to the badge readings 1 R=9.7 msv. More important than knowing the dose/study received by the technologists, however, is knowing the source of the dose. Hence, we recorded the dose received by each technologist while performing the following tasks during the 17 studies: radiopharmaceutical preparation, radiopharmaceutical administration, post-injection patient preparation, escorting the patient to the restroom, positioning the patient on the table, respositioning the patient between CT and PET scans, and helping the patient off the table after a scan. The duration of each task was recorded, along with the dose reading before and after each task. Analysis of the data revealed which task most increased technologist doses. III. RESULTS AND DISCUSSION A. Patient self-attenuation A significant decrease in exposure relative to that from a point source was observed at all three dosimeter locations. The exposure expected from a point source is given by the equation X point = A t r 2, where =5.73 R cm 2 mci 1 h 1 for FDG, A is the timeaveraged radioactivity level during the time period t found by integrating over the exponential decay, and r is the distance from the source. For any point-source calculation, a proper source location must be chosen. For both our calculations and the control measurement, this point was chosen at the patient s sternum, about halfway down the trunk from the head. The patient self-attenuation is defined simply as a percent decrease from the point-source value, X point at that location SA X point X measured X point =1 X measured X point. If a patient s body attenuates all radiation, then SA=1; if no radiation was attenuated, SA=0. A test sample of FDG with an radioactivity of 286 MBq 7.74 mci in a syringe placed in the recliner at chest level for a reclining patient was used as a control. At dosimeter locations 1 and 2, the exposures from the syringe were measured over a period of 104 minutes. The measured exposures 1.2 mr at location 1 and 5.1 mr at location 2 were the same as those predicted for a point source. That is, SA=0, as expected. The distribution of FDG changes over time, as it is collected in the bladder and other organs. The effect of this distribution is difficult to quantify, but it will cause a greater divergence from a point-source-inverse square law model close to the patient. Buildup effects, due to scatter, are also greatest near the patient, where they increase the exposure. At greater distances, the exposure is dominated by the inverse square dropoff. Hence, the self-attenuation will vary with location and distance from the patient. At distances from the patient of 2.5 m or more, however, the effects of distribution become insignificant; the patient can reasonably be modeled as a point source with a constant self-attenuation due to patient absorption and scatter. For the measurements at locations 1 and 2, the average injected radioactivity was A=540 MBq 14.6 mci, the average uptake period was t=61 min and the average measured exposures were X 1 =0.86 mr 244 cm and X 2 =3.51 mr 112 cm. At location 3 84 cm, the averages are A=537 MBq 14.5 mci, t=56 min, X 3 =3.86 mr. The mean self-attenuation factors for the three dosimeter locations were SA 1,mean =28% 244 cm, SA 2,mean =38% 112 cm, and SA 3,mean =56% 84 cm. Probability distributions of selfattenuation at locations 1 and 2 are shown in Fig. 2. Due to a smaller patient sample at location 3, no distribution is shown. Variation in patient size, distribution, and position-

4 864 B. W. Zeff and M. V. Yester: Self-attenuation and public and technologist dose in PET 864 FIG. 3. Technologist dose/study for a technologist 1 and b technologist 2 over a period of 18 days. The dose from a single study is highly dependent upon the specific needs of the patient. See Table I ing are reflected in the considerable spread of data around the mean. The presence of a radiation-emitting patient in the second patient prep room had no measurable effect on the data. The general trend shows the self-attenuation factor dropping with distance from the patient, approaching a constant, nonzero value. Previous measurements showed greater self-attenuation laterally, which our data neither confirm nor contradict. The numerical simulations, however, were only performed for a lateral position. At the three dosimeter distances, the numerical, water-phantom simulations gave SA 1 =28%, SA 2 =36%, and SA 3 =40%, in good agreement with our data at distances greater than 1 m. In planning radiation protection for a PET suite, it is good to be conservative. It is best, perhaps, to begin any shielding calculations with a point-source approximation. High selfattenuation at all distances assures that such approximations are quite conservative. Patient self-attenuation data may then be used in a number of ways. Radiation protection for 511- kev radiation is tricky, often requiring large distances or massive lead shielding. In many cases, the thickness of lead called for is simply impractical to mount in a wall or is prohibitively expensive, or the available space for the suite is tight. In such cases, accounting for self-attenuation may save a meter or two of distance or reduce the required of lead shielding by 1/8 in. For distances greater than 2.5 m from the patient, an assumption of SA=20% would still yield conservative estimates of exposure. At closer distances, even greater savings can be expected, and again, the selfattenuation adds in a conservative factor. B. PET technologist dose Badge readings from the two technologists showed averages of 9.5 and 12.1 Sv/study. Doses recorded on the personal dosimeters during the same period were lower, averaging 7.2 and 9.7 Sv/study Fig. 3. Data from Benatar et al. are consistent with these numbers; they measure a technologist dose of 5.5 Sv/study for 352 MBq 9.5 mci injected radioactivity, corresponding to an 8.9 Sv/study for 518 MBq 14 mci as is used at our institution. Chiesa et al. measured 11.9 Sv/study 10 after scaling their data for 14 mci of radioactivity. In all cases, 10 Sv/study is a reasonable estimate. The discrepancy between the badge readings, which use optically stimulated luminescence OSL, and dosimeter readings reflects differences in the method of calculating equivalent dose as well as differences in accuracy. The dosimeters have an inherent calibration accuracy of ±5%. The Nuclear Regulatory Commission NRC limits occupational equivalent dose to 50 msv/year. At the time of this study, approximately 1000 studies per year were performed in our PET center, and our technologists received approximately 10 msv/year. While this dose is within the upper limit, the as low as reasonably achievable ALARA principle calls for lower doses. In accordance with common practice, this limit has been lowered by a factor of ten at our institution, to 5 msv/ year. By this standard, doses received by our PET technologists are too high. In our PET suite, technologists in the control room are separated from resting patients in prep room 1 by 4.8 m and in prep room 2 by 6.9 m. This distance provides the primary radiation protection during the uptake period. Additionally, the wall and window between the control room and the PET/CT scanner are shielded with 1.6 mm of lead, which provides protection during the CT scan. The time-averaged patient radioactivity and uptake period are 454 MBq 12 mci and 1 h, respectively. Assuming a point-source approximation for the patient as a radiation source with 28% selfattenuation, the technologist dose would be 2.1 Sv/study for a patient in room 1 and 1.0 Sv/study for a patient in room 2. Assuming equal use of the two rooms, technologists in the control room should receive a dose of 1.6 Sv/study from patients in the prep rooms. Actual dose measurements in the control room were even lower, 1.1 Sv/study. Hence, nearly 90% of technologist dose is received during close interactions with the patient and radiopharmaceutical, when shielding is especially difficult. In tracking the dose during individual studies, we found that a technologist is in close proximity to FDG or a radiation-emitting patient for 9 20 min, an unexpectedly large period of time. A breakdown of technologist tasks during a study and the dose received is shown in Table I. The tasks shown in Table I would not typically all be performed by a single technologist, so the cumulative dose of Sv/study would likewise be shared. FDG administration, however, is responsible for the highest technologist dose. At times, both technologists were present during this process. Ideally, only one technologist is present during the

5 865 B. W. Zeff and M. V. Yester: Self-attenuation and public and technologist dose in PET 865 TABLE I. A breakdown of time and dose for technologist tasks during a single PET study. These tasks may be split among multiple technologists to spread out the dose. For other tasks, both technologists may be present. In certain cases, a very sick patient might require much more attention, thereby increasing the interaction time and dose. Technologist Task Time minutes Dose msv dose preparation dose administration post-dose preparation patient to restroom position patient on table reposition patient for PET patient off table dose administration, with the technologists alternating this duty. Postinjection preparation may simply involve helping a patient into the reclining position, though some patients are nervous or ill and require more assistance from the technologists during the uptake period. Positioning a patient on the scanner table can usually be performed quite easily, but in certain cases can be quite complicated. One patient with serious breathing difficulty due to illness needed to be repeatedly moved out of the scanner and readjusted, requiring close interaction with a technologist for more than five minutes. The task of positioning the patient for the PET scan was due to the fact that a combined PET/CT scanner is employed. Originally, the table had to be indexed from the CT position to the PET position by the operator. This function is now performed remotely by the technologist from the control booth and was performed remotely during this study. By the end of the scan, the level of radioactivity in the patient is much lower and doses to the technologists are less significant. The tasks listed in Table I are necessary and can only be performed in close proximity to a patient. Still, a number of steps can be taken to minimize technologist exposure. The majority of technologist exposure occurs during radiopharmaceutical dose preparation and administration. When preparing and transporting the FDG, a full syringe shield, with no glass window, should be used. Technologists at our facility use tungsten syringe shields when preparing and administering injections. Before bringing the FDG out of the hot lab, a patient s IV should already be started, a practice we also follow. A very short leader tube into the IV should be used, since the radioactivity in the tube is entirely unshielded. The postinjection procedure should be explained to the patient and all questions dealt with prior to administration of the radiopharmaceutical. If the patient has questions during the uptake period, the technologist should talk with the patient from as great a distance as possible. In general, tasks should be planned to be as efficient as possible. If high doses are a problem, it may also be useful to audit technologist dose during a few individual studies. Using a monthly radiation badge cycle instead of quarterly for PET technologists allows quicker feedback on how well procedural changes are working and help technologists to self-monitor. In certain situations, rotating jobs amongst technologists may be necessary. IV. CONCLUSIONS The design of shielding and radiation safety procedures for PET presents challenges for the medical physicist. During long uptake times, controlled and noncontrolled personnel, and the public must be protected from patient sources of 511-keV radiation. For noncontrolled areas, a simple distance-and-shielding model for radiation protection is appropriate to estimate dose, as personnel or the public in these areas will have no other contact with the radioactive source. Factoring patient self-attenuation into a point-source approximation can provide a conservative estimate of radiation exposure with significant distance, shielding, and cost reduction. Protection of PET personnel is more complicated, as 90% of technologist occupational dose comes from close interaction with the patient. Certainly, appropriate shielding is necessary to protect technologists from patient sources. Any significant reduction of technologist dose, however, requires careful monitoring and limiting of technologist and patient interactions. ACKNOWLEDGMENTS The authors would like to thank T. Mahone and B. Jackson for their extensive role in data collection. We would also like to acknowledge Sharon White for her helpful comments on the manuscript. a Electronic mail: bzeff@uabmc.edu 1 M. E. Phelps, Comments and Perspectives, J. Nucl. Med. 45, J. C. Courtney, P. Mendez, O. Hidalgo-Salvatierra, and S. Bujenovic, Photon Shielding for a Positron Emission Tomography Suite, Operational. Radiation Safety J. 81, S24 S A. Bixler, G. Springer, and R. Lovas, Practical Aspects of Radiation Safety for Using Flourine-18, J. Nucl. Med. Technol. 27, T. F. Brown and N. J. Yasillo, Radiation Safety Considerations for PET Centers, J. Nucl. Med. Technol. 25, M. A. Dell, Radiation Safety Review for 511-keV Emitters in Nuclear Medicine, J. Nucl. Med. Technol. 25, K. J. Kearfott, J. E. Carey, M. N. Clemenshaw, and D. B. Faulkner, Radiation Protection Design for a Clinical Positron Emission Tomography Imaging Suite, Health Phys. 63, N. A. Benetar, B. F. Cronin, and M. J. O Doherty, Radiation Dose Rates from Patients Undergoing PET: Implications for Technologists and Waiting Areas, Eur. J. Nucl. Med. 27, B. F. Cronin, B. K. Marsden, and M. J. O Doherty, Are Restrictions to Behaviour of Patients Required Following Fluorine-18 Fluorodeoxyglucose Positron Emission Topographic Studies, Eur. J. Nucl. Med. 26, S. White, D. Binns, V. Johnston, M. Fawcett, B. Greer, F. Ciavarella, and R. Hicks, Occupational Exposure in Nuclear Medicine and PET, Clin. Positron Imaging 3, C. Chiesa, V. De Sanctis, F. Crippa, M. Schiavini, C. E. Fraigola, A. Bogni, C. Pascali, D. Decise, R. Marchesini, and E. Bombadieri, Radiation Dose to Technicians Per Nuclear Medicine Procedure: Comparison Between Technetium-99m, Gallium-67, and Iodine-131 Radiotracers and Fluorine-18 Fluorodeoxyglucose, Eur. J. Nucl. Med. 24,