0 Special Feature
INTERIM REPORT ON THE NATIONAL BUREAU OF STANDARDS/BUREAU OF RADIOLOGICAL HEALTH @to TELETHERAPY SURVEY? DONALD Bureau of Radiological
L.
Ph.D.
Health. Food and Drug Administration. HAROLD
International
THOMPSON,
Commission
0.
on Radiation
WYCKOFF,
Rockville,
MD. U.S.A.
Ph.D.
Units and Measurements,
Washington.
DC. U.S.A.
and CHRISTOPHER National
Bureau of Standards.
G.
Department
SOARES,
Ph.D.
of Commerce.
Washington,
DC. U.S.A.
During the past three years tbe National Bureau of Standards and the Bureau of Radiological Health have been conducting a survey of “Co teietberapy facilities to determine their accuracy in exposing a phantom to a prescribed dose. As of May 1977, some 700 units were surveyed of the approximately 1000 which are licensed to administer therapy in tbe United States. Preliminary data indicate that about two-thirds of the respondents were able to calcuiate the spec&ed 300 rad dose within 5 rad, while 8% reported values which varied from the spe&ied dose by more than 15 rad. The mean exposure determined for a set of dosimeters differed from the 300 rad value by less than 5% for 83% of units; 4% of the units produced an exposure which differed by more than 10% of the target value. * “co tektherapy,
Survey, Dose accuracy,
Dose cakulations.
INTRODUCTION
confidential and that only statistical summaries would be released for publication. Responding facilities were provided with a questionnaire and a set of thermoluminescent dosimeters with instructions for their exposure. The survey variables identified were limited in scope to demographic characteristics of the facility, factors associated with calculation of the exposure time, and measurements of a simulated dose. The protocol did not address such factors as the qualifications of personnel, the make and model of therapy units nor the instruments used in calibration. The details of the
During the past 3 years, the National Bureau of Standards (NBS) and the Bureau of Radiological Health (BRH) have conducted a nationwide survey of (‘(‘Co teletherapy facilities to determine their accuracy in delivering a prescribed dose to a “phantom”. The purpose of the survey was to identify potential problems in dose delivery and develop appropriate remedial actions. In an effort to include in the survey all facilities with teletherapy service. a comprehensive list of licensees was developed by consulting the Nuclear Commission, individual Regulatory the state government radiation control directors. professional publications and manufacturers’ distribution lists. The nearly 1000 facilities identified were contacted in subsampies of approximately 50 each and invited to in the survey. Each respondent was participate guaranteed that individual results would be kep?
meetings of the American Association of Physicists in Medicine.‘.’ It is important to note that the inherent measurement error associated with a set of dosimeters was approximately 2.8%. A description of the dosimetric system will be included in the technical
+Presented at the annual meeting of the American Society of Therapeutic Radiologists. Denver. Colorado.
3 November 1978. Accepted for publication
dosimetric design and the questionnaire content were reported previously in papers presented at annual
17 May 1978.
1066
Radiation Oncology 0 Biology 0 Physics
November-December
section of the comprehensive report planned for publication early next year. This report also will include a statistical analysis as well as recommendations on quality assurance to improve the delivery of therapeutic radiation.
600 E
500
2
400
Divisions of the U.S. Total New England Middle Atlantic E. North Central W. North Central South Atlantic E. South Central W. South Central Mountain Pacific Territories
Total Facility Units 684 39 143 142 66 86 41 63 26 72 6
737 41 155 149 71 94 43 70 26 80 8
facilities and teletherapy Hospitals Facility Units 586 36 119 129 58 75 39 48 19 58 5
629 38 131 133 63 82 41 50 19 65 7
’
z
u. 300 0 : z
200
7 z
loo-
n_
n Hospitals
verify the initial data and to measure the nonrespondent impact on the findings. In the initial phase, in the United approximately 70% of the facilities
Table 1. Number of participating
-
L-2
PRELIMINARY FINDINGS The initial phase of the survey has been completed. A resurvey of selected subsamples has been done to
States accepted the invitation to participate by completing the questionnaire and exposing the survey dosimeters. As Fig. 1 indicates, 86% of the responding facilities were hospitals, 9% were clinics or group practices and 4% were private offices. Table 1 shows the distribution of participants by the major divisions of the U.S. The Middle Atlantic and East North Central divisions represented over 40% of the responding facilities with 142 and 143 facilities respectively. Most of the private offices in the survey were located in Middle Atlantic States while clinics were distributed more evenly throughout the country. The influence of nonrespondents on this demographic distribution data is unknown. As Table 2 indicates, the participating facilities were nearly evenly divided between those States under control of the Nuclear Regulatory Commission (non-agreement states) and those which are selfregulating under an agreement with the Nuclear Regulatory Commission. Both Table 2 and Fig. 2 indicate the average differences of the dosimeter readings from the target dose of 300rad for the non-agreement and agreement states. The average difference is slightly greater within the agreement states, [12.7 rad (4%)] than in the non-agreement states [lO.Orad (3%)]. For both categories of states the average difference from 300 rad was markedly higher in private offices. However, it should be noted that only 28 private offices are included, whereas
1978, Vol. 4, No. 11 and No. 12
Clinics-
Private
Groups
Off ices
TYPE
Fig. 1. Distribution
9 P
If Ss =< z.c z
Type Unknown
OF FACILITY
77
of participating
facilities.
20
15
IO
2 Y
5
d 9
a
0
Hospitols
Private
Clinics-
Offices
Groups
TYPE
TYPO Unknown
OF FACILITY
Fig. 2. Average difference
NO”
of dosimeter
77
readings.
results were obtained from 545 hospitals and 64 clinics or group practices. This table does not include data from the 40 federal facilities which are licensed by the Nuclear Regulatory Commission even when states. located in agreement In this survey, 98% of the participating hospitals were general medical-surgical institutions. The preliminary analysis of results for such hospitals has been tabulated according to the type of ownership (Table 3, Fig. 3). It appears that there may be a significant difference in the performance of some of these groups, particularly those identified as investorunits by type of facility and divisions Private offices Facility Units 28 0 17 1 0 4 0 2 1 3 0
28 0. 17 1 0 4 0 2 1 3 0
Clinics or Groups Facility Units
of the United States Type unknown Facility Units
64 2 7 12 6 7 2 II 6 10
74 2 7 15 6 8 2 16 6 11
6 1 0 0 2 0 0 2 0 1
6 1 0 0 2 0 0 2 0 1
1
I
0
0
Interim report on the National Bureau of Standards 0 D. L.
THOMPSON
1067
et al.
Table 2. Number of participating facilitiest and average differences of dosimeter readingst by agreement state status and type of facility Total
Mean differ .
Facility
Type of facility Total Hospitals Private offices Clinics or groups Type unknown
644 546 28 64 6
Agreement
states
Facility
Mean differ.
11.3 10.7 23.4 11.4 6.5
tExcludes federal hospitals. $Difference in rad between dosimeter
313 252 23 36 2
Nonagreement states
Facility
12.7 12.1 23.9 10.1 4.3
Mean differ.
331 294 5 28 4
10.0 9.5 21.2 13.1 7.7
readings and target dose of 300 rad.
l”YCI,or
Owned
-
TYPE OF OWNERSHIP
Fig.
3. Average
NO” 77
difference of dosimeter general hospitals.
readings
for DOSE (RAD) NOV 77
Fig. 4. Distribution
of participating facilities by calculated dose and by average dosimeter reading.
The average difference of dosimeter readings from 3OOrad was 13.9rad for federal facilities, about the same as facilities in the agreement states. Figure 4 depicts the distribution of the facilities for all sets of dosimeter readings. As expected, small errors occur frequently, large errors are observed less often. About 83% of the participants delivered an average dose to the dosimeters within 25% of 300rad. The data are skewed toward larger doses owned
operations.
Table 3. Number
reflecting some tendency to err in the positive direction. Based on the exposure rates or dose rates reported in the questionnaires, the determination of exposure times required for 300 rad were recalculated. In most instances the exposure times listed by the participant would not have delivered exactly 300rad. The distribution of facilities by calculated values is also
of participating general hospitalst and average differences dosimeter readings* by type of ownership
Type of ownership Total State City and/or County Hospital district Church operated Nongovernment-not for profit Investor owned Federal
General hospitals
Mean difference
572 25 50 28 133 284
10.7 10.0 10.5
12 40
19.0 13.9
;:: 11.0
?Excludes pediatric, orthopedic and other specialty hospitals. SDifference in rad between dosimeter readings and target dose of 300 rad.
of
1068
Radiation Oncology @ Biology 0 Physics
depicted in Fig. 4. Calculated doses are rather symmetrical around the target value. The plots of readings and calculated values all fall within 510%. Values outside the scale have not been shown; these include some 32 sets of readings and 5 calculations. In many cases it is possible to identify the source of the computational errors, and a listing is being prepared to identify the problems that occur more frequently. In addition, further analysis of the calculated values and their correlations with dosimeter readings can yield information on the reasons for variation in delivered doses. In a random sample of 92 calculations checked for this presentation 43% listed the wrong value for conversion of roentgen to rad. Many respondents cited the tissue-to-air factor in spite of instructions to expose as if in water. It is uncertain how many failed to follow instructions and how many were not aware of the application of conversion factors. In the same sample, 24% of the participants entered an incorrect value for backscatter and at least 37% erred in their value for the decay correction. One participant gave four different sets of calibration data with a spread of 9%. Occasionally values for factors were not listed on the returned form and it was not always possible to determine whether the factors had been omitted in the calculation. This initial review also has raised
November-December
1978, Vol. 4, No. 11dnd No. 12
questions about the use of displacement or attenuation factors and the appropriate intervals for decay correction. QUALITY ASSURANCE PROGRAM Based in part on the response to the survey, a long-term voluntary program of quality assurance is being considered by the Bureau of Radiological Health. The program initially may focus on calibration procedures and dose calculations but probably will be expanded to include training projects. particularly to reinforce earlier career education. In addition to the American Society of Therapeutic Radiologists, an attempt will be made to coordinate the program with the National Bureau of Standards, the Nuclear Regulatory Commission, the Conference of State Radiation Control Program Directors, the American College of Radiology, the American Association of Physicists in Medicine and other professional groups. The comprehensive report planned for early 1978 will contain recommendations on such a program utilizing the detailed findings of the survey, the review of responses to the Food and Drug Administration Federal Register Notice of Intent to propose rules and develop guidelines for radiation therapy equipment (42FRlS428; 22 March 1977), and the interest of other groups in participating in the program.
REFERENCES I. Ehrlich, M., Welter, G.: Nationwide survey of “OCo teletherapy dosimetry. J. Res. NBS, 80A: No. 4, 662668. July-August, 1976. 2. Soares, C., Ehrlich, M.: Nationwide survey of “uCo
teletherapy
dosimetry. Am. Assoc. Phys. Med., Ohio. Aug. 1977. (In preparation as NBS Technical Note). Cincinnati.
INTERIM REPORT ON THE NATIONAL BUREAU OF STANDARDS/BUREAU OF RADIOLOGICAL HEALTH @to TELETHERAPY SURVEY? DONALD Bureau of Radiological
L.
Ph.D.
Health. Food and Drug Administration. HAROLD
International
THOMPSON,
Commission
0.
on Radiation
WYCKOFF,
Rockville,
MD. U.S.A.
Ph.D.
Units and Measurements,
Washington.
DC. U.S.A.
and CHRISTOPHER National
Bureau of Standards.
G.
Department
SOARES,
Ph.D.
of Commerce.
Washington,
DC. U.S.A.
During the past three years tbe National Bureau of Standards and the Bureau of Radiological Health have been conducting a survey of “Co teietberapy facilities to determine their accuracy in exposing a phantom to a prescribed dose. As of May 1977, some 700 units were surveyed of the approximately 1000 which are licensed to administer therapy in tbe United States. Preliminary data indicate that about two-thirds of the respondents were able to calcuiate the spec&ed 300 rad dose within 5 rad, while 8% reported values which varied from the spe&ied dose by more than 15 rad. The mean exposure determined for a set of dosimeters differed from the 300 rad value by less than 5% for 83% of units; 4% of the units produced an exposure which differed by more than 10% of the target value. * “co tektherapy,
Survey, Dose accuracy,
Dose cakulations.
INTRODUCTION
confidential and that only statistical summaries would be released for publication. Responding facilities were provided with a questionnaire and a set of thermoluminescent dosimeters with instructions for their exposure. The survey variables identified were limited in scope to demographic characteristics of the facility, factors associated with calculation of the exposure time, and measurements of a simulated dose. The protocol did not address such factors as the qualifications of personnel, the make and model of therapy units nor the instruments used in calibration. The details of the
During the past 3 years, the National Bureau of Standards (NBS) and the Bureau of Radiological Health (BRH) have conducted a nationwide survey of (‘(‘Co teletherapy facilities to determine their accuracy in delivering a prescribed dose to a “phantom”. The purpose of the survey was to identify potential problems in dose delivery and develop appropriate remedial actions. In an effort to include in the survey all facilities with teletherapy service. a comprehensive list of licensees was developed by consulting the Nuclear Commission, individual Regulatory the state government radiation control directors. professional publications and manufacturers’ distribution lists. The nearly 1000 facilities identified were contacted in subsampies of approximately 50 each and invited to in the survey. Each respondent was participate guaranteed that individual results would be kep?
meetings of the American Association of Physicists in Medicine.‘.’ It is important to note that the inherent measurement error associated with a set of dosimeters was approximately 2.8%. A description of the dosimetric system will be included in the technical
+Presented at the annual meeting of the American Society of Therapeutic Radiologists. Denver. Colorado.
3 November 1978. Accepted for publication
dosimetric design and the questionnaire content were reported previously in papers presented at annual
17 May 1978.
1066
Radiation Oncology 0 Biology 0 Physics
November-December
section of the comprehensive report planned for publication early next year. This report also will include a statistical analysis as well as recommendations on quality assurance to improve the delivery of therapeutic radiation.
600 E
500
2
400
Divisions of the U.S. Total New England Middle Atlantic E. North Central W. North Central South Atlantic E. South Central W. South Central Mountain Pacific Territories
Total Facility Units 684 39 143 142 66 86 41 63 26 72 6
737 41 155 149 71 94 43 70 26 80 8
facilities and teletherapy Hospitals Facility Units 586 36 119 129 58 75 39 48 19 58 5
629 38 131 133 63 82 41 50 19 65 7
’
z
u. 300 0 : z
200
7 z
loo-
n_
n Hospitals
verify the initial data and to measure the nonrespondent impact on the findings. In the initial phase, in the United approximately 70% of the facilities
Table 1. Number of participating
-
L-2
PRELIMINARY FINDINGS The initial phase of the survey has been completed. A resurvey of selected subsamples has been done to
States accepted the invitation to participate by completing the questionnaire and exposing the survey dosimeters. As Fig. 1 indicates, 86% of the responding facilities were hospitals, 9% were clinics or group practices and 4% were private offices. Table 1 shows the distribution of participants by the major divisions of the U.S. The Middle Atlantic and East North Central divisions represented over 40% of the responding facilities with 142 and 143 facilities respectively. Most of the private offices in the survey were located in Middle Atlantic States while clinics were distributed more evenly throughout the country. The influence of nonrespondents on this demographic distribution data is unknown. As Table 2 indicates, the participating facilities were nearly evenly divided between those States under control of the Nuclear Regulatory Commission (non-agreement states) and those which are selfregulating under an agreement with the Nuclear Regulatory Commission. Both Table 2 and Fig. 2 indicate the average differences of the dosimeter readings from the target dose of 300rad for the non-agreement and agreement states. The average difference is slightly greater within the agreement states, [12.7 rad (4%)] than in the non-agreement states [lO.Orad (3%)]. For both categories of states the average difference from 300 rad was markedly higher in private offices. However, it should be noted that only 28 private offices are included, whereas
1978, Vol. 4, No. 11 and No. 12
Clinics-
Private
Groups
Off ices
TYPE
Fig. 1. Distribution
9 P
If Ss =< z.c z
Type Unknown
OF FACILITY
77
of participating
facilities.
20
15
IO
2 Y
5
d 9
a
0
Hospitols
Private
Clinics-
Offices
Groups
TYPE
TYPO Unknown
OF FACILITY
Fig. 2. Average difference
NO”
of dosimeter
77
readings.
results were obtained from 545 hospitals and 64 clinics or group practices. This table does not include data from the 40 federal facilities which are licensed by the Nuclear Regulatory Commission even when states. located in agreement In this survey, 98% of the participating hospitals were general medical-surgical institutions. The preliminary analysis of results for such hospitals has been tabulated according to the type of ownership (Table 3, Fig. 3). It appears that there may be a significant difference in the performance of some of these groups, particularly those identified as investorunits by type of facility and divisions Private offices Facility Units 28 0 17 1 0 4 0 2 1 3 0
28 0. 17 1 0 4 0 2 1 3 0
Clinics or Groups Facility Units
of the United States Type unknown Facility Units
64 2 7 12 6 7 2 II 6 10
74 2 7 15 6 8 2 16 6 11
6 1 0 0 2 0 0 2 0 1
6 1 0 0 2 0 0 2 0 1
1
I
0
0
Interim report on the National Bureau of Standards 0 D. L.
THOMPSON
1067
et al.
Table 2. Number of participating facilitiest and average differences of dosimeter readingst by agreement state status and type of facility Total
Mean differ .
Facility
Type of facility Total Hospitals Private offices Clinics or groups Type unknown
644 546 28 64 6
Agreement
states
Facility
Mean differ.
11.3 10.7 23.4 11.4 6.5
tExcludes federal hospitals. $Difference in rad between dosimeter
313 252 23 36 2
Nonagreement states
Facility
12.7 12.1 23.9 10.1 4.3
Mean differ.
331 294 5 28 4
10.0 9.5 21.2 13.1 7.7
readings and target dose of 300 rad.
l”YCI,or
Owned
-
TYPE OF OWNERSHIP
Fig.
3. Average
NO” 77
difference of dosimeter general hospitals.
readings
for DOSE (RAD) NOV 77
Fig. 4. Distribution
of participating facilities by calculated dose and by average dosimeter reading.
The average difference of dosimeter readings from 3OOrad was 13.9rad for federal facilities, about the same as facilities in the agreement states. Figure 4 depicts the distribution of the facilities for all sets of dosimeter readings. As expected, small errors occur frequently, large errors are observed less often. About 83% of the participants delivered an average dose to the dosimeters within 25% of 300rad. The data are skewed toward larger doses owned
operations.
Table 3. Number
reflecting some tendency to err in the positive direction. Based on the exposure rates or dose rates reported in the questionnaires, the determination of exposure times required for 300 rad were recalculated. In most instances the exposure times listed by the participant would not have delivered exactly 300rad. The distribution of facilities by calculated values is also
of participating general hospitalst and average differences dosimeter readings* by type of ownership
Type of ownership Total State City and/or County Hospital district Church operated Nongovernment-not for profit Investor owned Federal
General hospitals
Mean difference
572 25 50 28 133 284
10.7 10.0 10.5
12 40
19.0 13.9
;:: 11.0
?Excludes pediatric, orthopedic and other specialty hospitals. SDifference in rad between dosimeter readings and target dose of 300 rad.
of
1068
Radiation Oncology @ Biology 0 Physics
depicted in Fig. 4. Calculated doses are rather symmetrical around the target value. The plots of readings and calculated values all fall within 510%. Values outside the scale have not been shown; these include some 32 sets of readings and 5 calculations. In many cases it is possible to identify the source of the computational errors, and a listing is being prepared to identify the problems that occur more frequently. In addition, further analysis of the calculated values and their correlations with dosimeter readings can yield information on the reasons for variation in delivered doses. In a random sample of 92 calculations checked for this presentation 43% listed the wrong value for conversion of roentgen to rad. Many respondents cited the tissue-to-air factor in spite of instructions to expose as if in water. It is uncertain how many failed to follow instructions and how many were not aware of the application of conversion factors. In the same sample, 24% of the participants entered an incorrect value for backscatter and at least 37% erred in their value for the decay correction. One participant gave four different sets of calibration data with a spread of 9%. Occasionally values for factors were not listed on the returned form and it was not always possible to determine whether the factors had been omitted in the calculation. This initial review also has raised
November-December
1978, Vol. 4, No. 11dnd No. 12
questions about the use of displacement or attenuation factors and the appropriate intervals for decay correction. QUALITY ASSURANCE PROGRAM Based in part on the response to the survey, a long-term voluntary program of quality assurance is being considered by the Bureau of Radiological Health. The program initially may focus on calibration procedures and dose calculations but probably will be expanded to include training projects. particularly to reinforce earlier career education. In addition to the American Society of Therapeutic Radiologists, an attempt will be made to coordinate the program with the National Bureau of Standards, the Nuclear Regulatory Commission, the Conference of State Radiation Control Program Directors, the American College of Radiology, the American Association of Physicists in Medicine and other professional groups. The comprehensive report planned for early 1978 will contain recommendations on such a program utilizing the detailed findings of the survey, the review of responses to the Food and Drug Administration Federal Register Notice of Intent to propose rules and develop guidelines for radiation therapy equipment (42FRlS428; 22 March 1977), and the interest of other groups in participating in the program.
REFERENCES I. Ehrlich, M., Welter, G.: Nationwide survey of “OCo teletherapy dosimetry. J. Res. NBS, 80A: No. 4, 662668. July-August, 1976. 2. Soares, C., Ehrlich, M.: Nationwide survey of “uCo
teletherapy
dosimetry. Am. Assoc. Phys. Med., Ohio. Aug. 1977. (In preparation as NBS Technical Note). Cincinnati.
