Correspondence time, in mammography physics. The data were presented by Dr K. C. Young of the centre at Symposium Mammographicum in April 1992 and will be published shortly by the NHS BSP. All sets measured had a mean glandular dose less than 3 mGy, with 98% less than 2 mGy. I will leave it to others to deal with the detailed technical questions Moores and Henshaw pose, but I can assure you that the NHS BSP takes radiation protection and the issues of benefits and disadvantages in screening very seriously indeed. Yours etc., J. PATNICK
Trent Regional Health Authority, Fulwood House, Old Fulwood Road, Sheffield S10 3TH (Received 28 July 1992 and accepted 1 September 1992) References MOORES, B. M. & HENSHAW, E. T., 1992. Radiation protection
associated with well woman breast cancer screening. British Journal of Radiology, 65, 552-553. NHS BSP, 1989. Quality Assurance Guidelines for Mammography, The Pritchard Report. IPSM, 1989. The Commissioning and Routine Testing ofMammographic X-ray Systems, Report No 59 (IPSM, York).
radiation detriment). This has already been carried out for the UK breast cancer screening programme where, for women aged between 50 and 65 years of age, the number of cancers detected and susceptible to cure was predicted to outweigh greatly the number of cancers induced (Department of Health and Social Security, 1986). The screening programme is therefore not a research project, as Moores and Henshaw suggest, since it has been justified and is now an accepted diagnostic practice. Although not recommending dose limits for medical exposures, ICRP has recently advocated the use of dose constraints in the process of optimization: these constraints should be set for common diagnostic procedures by professional and advisory bodies and applied with flexibility to allow higher doses when indicated by sound clinical judgement. The National Radiological Protection Board (1991) proposes to endorse this approach and will be giving further guidance in due course. For the purposes of optimization, dose constraints will relate to typical practices within a radiology department for various examinations and not to individual patients. The action to be taken if dose constraints are exceeded would be a local investigation of the circumstances and either modifications to techniques and equipment or justification of the need for the high doses. In summary, the 1990 recommendations of ICRP provide a framework for standards of radiation protection in mammography that are common to all medical and other types of exposure to ionizing radiation. Yours etc., P. C. SHRIMPTON B. F. WALL
Standards of radiation protection in mammographic screening THE EDITOR—SIR,
We wish to clarify some of the radiological protection issues raised by Moores and Henshaw (1992) in connection with the UK breast cancer screening programme using X-ray mammography. The International Commission on Radiological Protection (ICRP) has defined the category of "medical exposure" as principally the exposure of persons as part of their diagnosis or treatment with such exposure usually intended to provide a direct benefit to the exposed individual (ICRP, 1991). If the practice is justified and the protection optimized, the dose to the patient will be as low as is compatible with the medical purpose. ICRP suggests that the application of limits might be to the patient's detriment and therefore does not recommend dose limits for medical exposures. The purpose of mammographic screening is clearly the diagnosis of breast cancer, although the probability that any particular woman will have breast cancer is quite low. The irradiation of well women as part of such a programme should therefore be regarded as medical exposure according to the ICRP definition and so dose limits for members of the public are not relevant. British law (IRR, 1988) incorporates the ICRP principles of justification and optimization by requiring that all medical exposures be conducted in accordance with accepted diagnostic or therapeutic practice (justification), and with procedures selected to ensure that the dose to the patient is as low as reasonably practicable in order to achieve the required diagnostic or therapeutic purpose (optimization). In the case of mass screening by X-ray examination, justification requires the demonstration of a net positive benefit when weighing, in commensurate terms, all the benefits (both to individual persons and to the population) against all the costs (including
950
National Radiological Protection Board, Chilton, Didcot, Oxon OX 11 ORD (Received 30 July 1992 and accepted 1 September 1992) References DHSS, 1986. Report of a DHSS Working Group (Forrest, P., Chairman) to the Health Ministers of England, Wales, Scotland and Northern Ireland (Breast Cancer Screening, HMSO, London). ICRP, 1991. 1990 Recommendations of the International Commission on Radiological Protection. Publication 60. Annals of the ICRP, 21, Nos 1-3. IRR, 1988. The Ionising Radiation (Protection of Persons Undergoing Medical Examination or Treatment) Regulations. (HMSO, London) SI (1988), p. 778. MOORES, B. M. & HENSHAW, E. T., 1992. Radiation protection
associated with well woman breast cancer screening. British Journal of Radiology, 65, 552-553. NRPB, 1991. Board Advice Following Publication of the 1990 Recommendations ofthe ICRP. Consultative Document (NRPB, Chilton), M321.
Radiation protection associated with well women breast cancer screening THE EDITOR—SIR,
The letter from Moores and Henshaw (1992) raises several points which must be answered. However, we believe the answers are relatively simple provided certain points are kept clearly in mind. We wish to make the following comments:
The British Journal of Radiology, October 1992
Correspondence 1. To answer the first question in their second paragraph, "Irradiation of well women for screening purposes" does constitute medical exposure, for which there is no dose limit. Paragraph 2.16 of the Guidance Notes to the Ionising Radiations Regulations 1985 (NRPB, 1988) applies, which states: "Screening programmes... should be undertaken only if the expected medical benefits to the individuals examined and to the population as a whole exceed the economic and social costs, including the risks associated with the radiation dose involved. Since benefits are not always the same for all members of the population, screening should be limited normally to particular groups". This aspect of the NHS Breast Screening Programme (BSP) was fully considered by the Forrest Committee when recommending that the BSP should be set up, in the light of information then available and a knowledge of breast doses at that time, as outlined in Annex D of the Forrest Report (DHSS, 1986). This has been confirmed to us recently by members of the Forrest Committee. Paragraphs 2.19-2.29 of the Guidance Notes refer specifically to medical research and thus plainly do not apply to the NHS BSP. It then follows that paragraphs 3-8 and paragraph 10 of their letter are irrelevant. 2. Returning to paragraph 2 of their letter, the "required diagnostic purpose" of breast cancer screening is the early detection of breast cancer. The "upper level of dose considered acceptable to achieve this purpose" was given in the Pritchard Report (Department of Health, 1989) as 5 mGy per film to a standard breast using a grid but may be revised to 3 mGy shortly. However, this value is intended only as a guideline and not as a limit. 3. Regarding the last sentence of paragraph 2 of their letter, the original target of the NHS BSP was a detection rate of 5 cancers per 1000 women on the first screen, and about 3.7 per 1000 on later rounds (Forrest report, paragraph 8.5). The target for the first screening round is understood to have been exceeded in all parts of the UK, but for present and future discussion the detection rate on subsequent screening rounds will be more important. Early indications are that this may be around 4 per 1000 and it should be noted that in the Edinburgh trial (Roberts et al, 1990) just over 3 cancers per 1000 were detected on the 2nd, 3rd and 4th rounds at 2 yearly intervals (in contrast to the 3 yearly interval at present employed in the NHS BSP). The important point here is not just the numerical value, but its constancy for repeated screening rounds. At each successive screening round the benefit/risk ratio increases, because the greater age of the women results in reduced risk of radiation induced cancers appearing while the detection rate remains about the same. 4. Regarding paragraph 9 of their letter, the UK Mammography Physics Group study referred to was an early and provisional survey carried out informally in 1989 by a group of physicists in the field, and intended for their information only, but was subsequently used by the Commission of the European Communities (CEC) with permission as the best up-to-date figures then available. The CEC guidelines referred to the upper quartile so that inevitably 25% of centres exceeded this value! 5. To answer those questions at the end of their letter which relate to the physics of mammography: Question 1. This has been answered in paragraph 1 above, and is also covered in a letter from Mrs J. Patnick, National Coordinator of the NHS BSP. Question 2. "Actual levels of radiation employed" have been measured in some regions for compressed breasts 4.5 cm thick and found not to differ significantly from those predicted employing phantoms of 4 cm Perspex, although compressed breast thicknesses
Vol. 65, No. 778
are more usually around 5.5 cm rather than the 4.5 cm initially assumed. Question 3. Perspex phantoms have been employed because they provide a quick, practical check of dose by a standard method which can be used and repeated in all centres. As stated in IPSM Report 59, p. 34, "Dose will also be affected by the size and composition of the breast with the former varying both within and between populations and the latter throughout the life of the woman. This diversity of both dose specification and breast dimensions and composition makes dose comparisons very difficult and it is important for such purposes that doses be specified for a standard breast and in a standard way" (IPSM, 1989). The use of phantoms is not unknown in other radiological examinations, as shown by the recent CEC workshop on phantoms held in Wurzburg, and their use should probably be encouraged. Questions 4 and 5. It is important to distinguish dose limits from reference dose levels for diagnostic X-ray examinations. There are no dose limits for medical exposure. We leave the remaining question for others to answer. We believe that the fundamental radiation protection aspects of the NHS BSP were properly considered at its inception, and are being kept under review. In view of the relatively high benefit to risk ratio, it is probably more important to ensure that image quality remains high (to maximize cancer detection rate), than to show excessive concern over minimizing the dose involved. Yours etc., J. LAW D. R. DANCE K. FAULKNER M. C. FITZGERALD M. L. RAMSDALE A. ROBINSON
(IPSM Mammography Working Party) Department of Medical Physics and Medical Engineering Western General Hospital Crewe Road Edinburgh EH4 2XU (Received 30 July 1992 and accepted 1 September 1992) References DEPARTMENT OF HEALTH, 1989. Guidelines on the Establishment of
a Quality Assurance System for the Radiological Aspects of Mammography Used for Breast Screening (Department of Health, London). DHSS, 1986. Breast Cancer Screening. A Report of a Working Group Chaired by Professor Sir Patrick Forrest (HMSO, London), pp. 81-82. NRPB, 1988. Guidance Notesfor the Protection of Persons Against Ionising Radiations Arising from Medical and Dental Use (HMSO, London). IPSM, 1989. Commissioning and Routine Testing of Mammographic X-ray Systems. Report No. 59 (IPSM, York). MOORES, B. M. & HENSHAW, E. T., 1992. Radiation protection
associated with well woman breast cancer screening. British Journal of Radiology, 65, 552-553. ROBERTS, M. M., ALEXANDER, F. E., ANDERSON, T. J., CHETTY, U., DONNAN, P. T., FORREST, P., HEPBURN, W., HUGGINS, A., KIRKPATRICK, A. E., LAMB, J., MUIR, B. B., PRESCOTT, R. J.,
1990. Edinburgh trial of screening for breast cancer: mortality at seven years. Lancet, 335, 241-246.
951
Trent Regional Health Authority, Fulwood House, Old Fulwood Road, Sheffield S10 3TH (Received 28 July 1992 and accepted 1 September 1992) References MOORES, B. M. & HENSHAW, E. T., 1992. Radiation protection
associated with well woman breast cancer screening. British Journal of Radiology, 65, 552-553. NHS BSP, 1989. Quality Assurance Guidelines for Mammography, The Pritchard Report. IPSM, 1989. The Commissioning and Routine Testing ofMammographic X-ray Systems, Report No 59 (IPSM, York).
radiation detriment). This has already been carried out for the UK breast cancer screening programme where, for women aged between 50 and 65 years of age, the number of cancers detected and susceptible to cure was predicted to outweigh greatly the number of cancers induced (Department of Health and Social Security, 1986). The screening programme is therefore not a research project, as Moores and Henshaw suggest, since it has been justified and is now an accepted diagnostic practice. Although not recommending dose limits for medical exposures, ICRP has recently advocated the use of dose constraints in the process of optimization: these constraints should be set for common diagnostic procedures by professional and advisory bodies and applied with flexibility to allow higher doses when indicated by sound clinical judgement. The National Radiological Protection Board (1991) proposes to endorse this approach and will be giving further guidance in due course. For the purposes of optimization, dose constraints will relate to typical practices within a radiology department for various examinations and not to individual patients. The action to be taken if dose constraints are exceeded would be a local investigation of the circumstances and either modifications to techniques and equipment or justification of the need for the high doses. In summary, the 1990 recommendations of ICRP provide a framework for standards of radiation protection in mammography that are common to all medical and other types of exposure to ionizing radiation. Yours etc., P. C. SHRIMPTON B. F. WALL
Standards of radiation protection in mammographic screening THE EDITOR—SIR,
We wish to clarify some of the radiological protection issues raised by Moores and Henshaw (1992) in connection with the UK breast cancer screening programme using X-ray mammography. The International Commission on Radiological Protection (ICRP) has defined the category of "medical exposure" as principally the exposure of persons as part of their diagnosis or treatment with such exposure usually intended to provide a direct benefit to the exposed individual (ICRP, 1991). If the practice is justified and the protection optimized, the dose to the patient will be as low as is compatible with the medical purpose. ICRP suggests that the application of limits might be to the patient's detriment and therefore does not recommend dose limits for medical exposures. The purpose of mammographic screening is clearly the diagnosis of breast cancer, although the probability that any particular woman will have breast cancer is quite low. The irradiation of well women as part of such a programme should therefore be regarded as medical exposure according to the ICRP definition and so dose limits for members of the public are not relevant. British law (IRR, 1988) incorporates the ICRP principles of justification and optimization by requiring that all medical exposures be conducted in accordance with accepted diagnostic or therapeutic practice (justification), and with procedures selected to ensure that the dose to the patient is as low as reasonably practicable in order to achieve the required diagnostic or therapeutic purpose (optimization). In the case of mass screening by X-ray examination, justification requires the demonstration of a net positive benefit when weighing, in commensurate terms, all the benefits (both to individual persons and to the population) against all the costs (including
950
National Radiological Protection Board, Chilton, Didcot, Oxon OX 11 ORD (Received 30 July 1992 and accepted 1 September 1992) References DHSS, 1986. Report of a DHSS Working Group (Forrest, P., Chairman) to the Health Ministers of England, Wales, Scotland and Northern Ireland (Breast Cancer Screening, HMSO, London). ICRP, 1991. 1990 Recommendations of the International Commission on Radiological Protection. Publication 60. Annals of the ICRP, 21, Nos 1-3. IRR, 1988. The Ionising Radiation (Protection of Persons Undergoing Medical Examination or Treatment) Regulations. (HMSO, London) SI (1988), p. 778. MOORES, B. M. & HENSHAW, E. T., 1992. Radiation protection
associated with well woman breast cancer screening. British Journal of Radiology, 65, 552-553. NRPB, 1991. Board Advice Following Publication of the 1990 Recommendations ofthe ICRP. Consultative Document (NRPB, Chilton), M321.
Radiation protection associated with well women breast cancer screening THE EDITOR—SIR,
The letter from Moores and Henshaw (1992) raises several points which must be answered. However, we believe the answers are relatively simple provided certain points are kept clearly in mind. We wish to make the following comments:
The British Journal of Radiology, October 1992
Correspondence 1. To answer the first question in their second paragraph, "Irradiation of well women for screening purposes" does constitute medical exposure, for which there is no dose limit. Paragraph 2.16 of the Guidance Notes to the Ionising Radiations Regulations 1985 (NRPB, 1988) applies, which states: "Screening programmes... should be undertaken only if the expected medical benefits to the individuals examined and to the population as a whole exceed the economic and social costs, including the risks associated with the radiation dose involved. Since benefits are not always the same for all members of the population, screening should be limited normally to particular groups". This aspect of the NHS Breast Screening Programme (BSP) was fully considered by the Forrest Committee when recommending that the BSP should be set up, in the light of information then available and a knowledge of breast doses at that time, as outlined in Annex D of the Forrest Report (DHSS, 1986). This has been confirmed to us recently by members of the Forrest Committee. Paragraphs 2.19-2.29 of the Guidance Notes refer specifically to medical research and thus plainly do not apply to the NHS BSP. It then follows that paragraphs 3-8 and paragraph 10 of their letter are irrelevant. 2. Returning to paragraph 2 of their letter, the "required diagnostic purpose" of breast cancer screening is the early detection of breast cancer. The "upper level of dose considered acceptable to achieve this purpose" was given in the Pritchard Report (Department of Health, 1989) as 5 mGy per film to a standard breast using a grid but may be revised to 3 mGy shortly. However, this value is intended only as a guideline and not as a limit. 3. Regarding the last sentence of paragraph 2 of their letter, the original target of the NHS BSP was a detection rate of 5 cancers per 1000 women on the first screen, and about 3.7 per 1000 on later rounds (Forrest report, paragraph 8.5). The target for the first screening round is understood to have been exceeded in all parts of the UK, but for present and future discussion the detection rate on subsequent screening rounds will be more important. Early indications are that this may be around 4 per 1000 and it should be noted that in the Edinburgh trial (Roberts et al, 1990) just over 3 cancers per 1000 were detected on the 2nd, 3rd and 4th rounds at 2 yearly intervals (in contrast to the 3 yearly interval at present employed in the NHS BSP). The important point here is not just the numerical value, but its constancy for repeated screening rounds. At each successive screening round the benefit/risk ratio increases, because the greater age of the women results in reduced risk of radiation induced cancers appearing while the detection rate remains about the same. 4. Regarding paragraph 9 of their letter, the UK Mammography Physics Group study referred to was an early and provisional survey carried out informally in 1989 by a group of physicists in the field, and intended for their information only, but was subsequently used by the Commission of the European Communities (CEC) with permission as the best up-to-date figures then available. The CEC guidelines referred to the upper quartile so that inevitably 25% of centres exceeded this value! 5. To answer those questions at the end of their letter which relate to the physics of mammography: Question 1. This has been answered in paragraph 1 above, and is also covered in a letter from Mrs J. Patnick, National Coordinator of the NHS BSP. Question 2. "Actual levels of radiation employed" have been measured in some regions for compressed breasts 4.5 cm thick and found not to differ significantly from those predicted employing phantoms of 4 cm Perspex, although compressed breast thicknesses
Vol. 65, No. 778
are more usually around 5.5 cm rather than the 4.5 cm initially assumed. Question 3. Perspex phantoms have been employed because they provide a quick, practical check of dose by a standard method which can be used and repeated in all centres. As stated in IPSM Report 59, p. 34, "Dose will also be affected by the size and composition of the breast with the former varying both within and between populations and the latter throughout the life of the woman. This diversity of both dose specification and breast dimensions and composition makes dose comparisons very difficult and it is important for such purposes that doses be specified for a standard breast and in a standard way" (IPSM, 1989). The use of phantoms is not unknown in other radiological examinations, as shown by the recent CEC workshop on phantoms held in Wurzburg, and their use should probably be encouraged. Questions 4 and 5. It is important to distinguish dose limits from reference dose levels for diagnostic X-ray examinations. There are no dose limits for medical exposure. We leave the remaining question for others to answer. We believe that the fundamental radiation protection aspects of the NHS BSP were properly considered at its inception, and are being kept under review. In view of the relatively high benefit to risk ratio, it is probably more important to ensure that image quality remains high (to maximize cancer detection rate), than to show excessive concern over minimizing the dose involved. Yours etc., J. LAW D. R. DANCE K. FAULKNER M. C. FITZGERALD M. L. RAMSDALE A. ROBINSON
(IPSM Mammography Working Party) Department of Medical Physics and Medical Engineering Western General Hospital Crewe Road Edinburgh EH4 2XU (Received 30 July 1992 and accepted 1 September 1992) References DEPARTMENT OF HEALTH, 1989. Guidelines on the Establishment of
a Quality Assurance System for the Radiological Aspects of Mammography Used for Breast Screening (Department of Health, London). DHSS, 1986. Breast Cancer Screening. A Report of a Working Group Chaired by Professor Sir Patrick Forrest (HMSO, London), pp. 81-82. NRPB, 1988. Guidance Notesfor the Protection of Persons Against Ionising Radiations Arising from Medical and Dental Use (HMSO, London). IPSM, 1989. Commissioning and Routine Testing of Mammographic X-ray Systems. Report No. 59 (IPSM, York). MOORES, B. M. & HENSHAW, E. T., 1992. Radiation protection
associated with well woman breast cancer screening. British Journal of Radiology, 65, 552-553. ROBERTS, M. M., ALEXANDER, F. E., ANDERSON, T. J., CHETTY, U., DONNAN, P. T., FORREST, P., HEPBURN, W., HUGGINS, A., KIRKPATRICK, A. E., LAMB, J., MUIR, B. B., PRESCOTT, R. J.,
1990. Edinburgh trial of screening for breast cancer: mortality at seven years. Lancet, 335, 241-246.
951
