British Journal of Radiology (2006) 79, 479-482
© 2006 British Institute of Radiology
doi: 10.1259/bjr/33577478
Radiation benefit and risk at the assessment stage of the UK Breast Screening Programme
J Law, PhD1 and
K Faulkner, PhD2
1 Edinburgh University Department of Medical Physics, Chancellor's Building, Little France Crescent, Edinburgh EH16 4SB, 2 Quality Assurance Reference Centre, 9 Kingfisher Way, Silverlink Business Park, Wallsend NE28 9ND, UK
 |
Abstract
|
|---|
The balance between benefit and radiation risk in a breast cancer screening programme has received much attention at the initial screening stage. This paper extends that attention to first stage assessment, i.e. the first stage at which women are recalled for further investigation because of suspected lesions or other suspect film features, and prior to any biopsy. Numbers of films, including magnification films, taken at this stage, have been established in two UK regions by different methods. Average total mean glandular dose has been calculated using published data of dose per film from initial screening and a multiplying factor to allow for magnification film dose. It is concluded that the benefit/risk ratio is considerably higher at first stage assessment than at initial screening by a factor of between 4 and 9, because of the very much higher cancer detection rate in this well-defined sub-group of women. Qualitatively, this conclusion is unaffected by the quite wide variation between screening centres in the numbers of films taken at first stage assessment.
|
Introduction
|
|---|
Considerable attention has been given to benefit and risk in the initial screening stage of a breast cancer screening programme. This attention has sometimes centred on estimations of the numbers of cancers detected and induced as an indicator of the benefit/risk balance. Numbers of cancers detected and induced are not in themselves a direct measure of benefit and risk. Not all cancers detected by screening will be cured (though earlier detection usually gives a better prognosis), and not all of those induced will be fatal. A fuller discussion of these and related points has been given by us elsewhere [1]. Subsequent assessment stages, after the initial screening visit, have received very much less attention of this kind. This may be because it is easy to see that qualitatively the benefit/risk-balance will be better at assessment than at initial screening, but this point appears not to have been examined in quantitative detail.
In the UK Breast Screening Programme, screening rounds occur at 3-yearly intervals. Within each round, after the initial screening visit and examination, a proportion of the screened women are recalled for assessment because of suspicious or doubtful features on their films. This proportion is typically between 5% and 10% of those screened [2]. Women recalled in this way may have further mammograms taken in "contact" or conventional geometry and some will also have magnification films. Films of either of these kinds may be full field, i.e. cover the whole breast, or be collimated to the region of interest. This is sometimes referred to as first stage assessment, and that term will be used in this paper. Most of the women requiring such assessment are eventually given reassurance that all is well, while a minority (say 10–20%) of these are recalled for further assessment (second stage) which involves biopsy or needle samples [3].
This paper will concentrate principally on the benefit/risk balance at the first of these two stages, but some aspects of that balance at the second stage will also be discussed.
The ratio of cancers detected to estimates of cancers induced at the radiation dose levels involved is not the same as the benefit/risk ratio, but the two ratios have been shown to be related, with the benefit/risk ratio being probably in the broad region of half to two thirds of the detection/induction ratio [4, 5]. It is the detection/induction ratio which will be mainly considered in this paper, because the numbers required can be estimated with relative ease and reliability.
|
Numbers of films taken at first stage assessment
|
|---|
In order to estimate the numbers of cancers detected and induced at first stage assessment it is necessary to know the average numbers of films of each kind that are taken (i.e. conventional, magnification, coned or full field), the mean dose per film, and the cancer detection rate among women undergoing this assessment stage. Data for the second and third of these three can be obtained from published sources and will be discussed later. This section will describe attempts to find out numbers of films taken.
Initially, information was obtained from nine screening centres in the North East, Yorkshire and the Humber regions of England by means of a questionnaire sent to superintendent radiographers. The superintendents were asked to provide estimates of the numbers of films of each kind taken per woman attending in their own centre. In doing this, they were asked to use their own judgement regarding current normal practice rather than keeping detailed records for a prospective period, and to provide estimates of both the minimum numbers of films normally taken on all or most women and a maximum number. It was emphasised that this maximum should represent the upper end of the normal range and not the maximum ever known or the worst possible case, since the latter would tend to distort the general picture.
The results of this enquiry are given in Table 1
. All the centres concerned, except one, now take two views (normally craniocaudal (CC) and lateral oblique (LO)) at the initial visit on each screening round, and this will tend to reduce the numbers of films required at assessment. In the absence of sufficient detailed information, it is assumed for calculation purposes that all films taken in conventional contact geometry are full-field, though this is almost certainly not the case. All the films will have been taken on one breast only, unless a suspicious feature was seen in each breast; in the latter case, the consequent greater dose is balanced by the corresponding increased probability of a cancer being identified.
View this table:
[in this window]
[in a new window]
|
Table 1. Average numbers of films taken per woman attending at first stage assessment– England (North East, Yorkshire and The Humber) (9 screening centres)
|
|
In a separate exercise, data on film numbers were obtained from four Scottish Screening Centres, who extracted numbers from their records retrospectively. These were therefore actual numbers for a defined period chosen essentially at random, e.g. the previous month, and did not depend on memory or judgement. The results obtained in this way are shown in Table 2. This region has not yet adopted two-view screening except for the first screening round. Consequently at first stage assessment films are taken of the CC view on each breast unless these have already been taken at the basic screening visit. Occasionally the mediolateral (ML) view is substituted for the CC view. When two-view screening on all rounds is adopted, all these films will be taken at the basic screening visit, and therefore their numbers are excluded from the following account. All other films are categorised as plain films with full field exposed, plain films coned down, and magnification films, all of which are coned down in this region. Average numbers of films of each kind in each of four centres are given in Table 2.
View this table:
[in this window]
[in a new window]
|
Table 2. Average numbers of films taken per woman attending at first stage assessment– Scotland (four Screening Centres)
|
|
It is clear that practice varies between these four centres in some respects. For example, Centre C does three times as many magnification films as Centres A and B; this is a general policy in that centre, not related to the preference of any one radiologist. In some centres, films taken with a small compression paddle are routinely coned down to about the area of the paddle; in other centres the beam may not be coned in this way, to allow the area of interest to be viewed in relation to the architecture of the breast as a whole. Centre B follows this latter policy so that all their plain films are assumed to be effectively full field for dose estimation purposes.
In this region, film numbers are broadly similar to those in Region A, but with many fewer magnification films. Average numbers of films from both these regions are combined in Table 3.
View this table:
[in this window]
[in a new window]
|
Table 3. Average numbers of films taken per woman attending at first stage assessment: results combined from two UK regions given equal weight, based on two-view initial screening
|
|
|
Results
|
|---|
Mean glandular breast dose at first stage assessment
Mean breast dose in the UK Breast Screening Programme has been estimated at a little under 2.5 mGy [6] per woman for the LO view and 2.0 mGy for the CC view, giving a mean dose of 2.25 mGy per view to the assessed breast. A more recent estimate [7] suggests very slightly lower values, but the earlier higher values are retained in this paper. Future estimates of mean breast dose may fluctuate around these values. The ratio of magnification film dose to contact geometry film dose has very recently been estimated to be 2.2±0.15, giving magnification film doses of 5.5 mGy (LO) and 4.4 mGy (CC) [8]. Since magnification films are equally likely to be taken on either view, the mean magnification film dose is 5.0 mGy.
Coned films are assumed, for this calculation, to cover approximately one third of the full breast area, on average [9]. In practice, this proportion varies with sizes of individual breasts and with exact sizes of defining cones on different X-ray units, but one third is considered to be a reasonable average value.
In general, for a suspicious feature or features in one breast only, only one breast is imaged at assessment. All estimates of risk factor for breast cancer induction are based on assumptions of equal radiation exposure to both breasts. Therefore, if the usual breast cancer induction risk factors [1, 8] are taken, the effective breast dose (and hence the effective risk) is half the actual dose to the one breast (Table 4).
View this table:
[in this window]
[in a new window]
|
Table 4. Mean glandular dose at first stage assessment(i.e. number of films x dose per film). For further calculation these numbers are divided by 2 because only one breast is irradiated (i.e. effective mean glandular dose to a single breast 5.1 mGy)
|
|
The mean glandular breast dose is calculated in this way in Table 4
, using film numbers from Table 3. Because ranges of film numbers were given in Table 1, the upper ends of those ranges have been used in this calculation.
Cancer detection rate at first stage assessment
Typical cancer detection rates are around 6 per 1000 women who attend for basic screening, with an age-dependant range mostly within 5–7 per 1000 [2]. The proportion of women recalled for assessment is usually between 5% and 10% [2], and all cancers detected are detected within this sub-group. Thus, for a 10% assessment recall rate, the detection rate at assessment is 10 times what it is at initial screening, and for a 5% recall rate it will be 20 times that figure.
|
Discussion and conclusions
|
|---|
Benefit and risk are difficult to define, but one simple possibility is to regard lives saved from death by breast cancer as benefit and lives lost to radiation induced breast cancer as risk. Cancers detected and induced are not the same as these, but there are reasonable grounds for thinking that the benefit/risk ratio is around half to two-thirds of the detection/induction ratio [4, 5].
It has already been shown that compared with initial screening, the cancer detection rate has increased 10 times in the sub-group recalled for assessment, if the recall rate is 10%, or 20 times if the recall rate is only 5%. The effective mean glandular dose per woman at first stage assessment has been shown in Table 4 to be up to about 5.1 mGy as against 4.5 mGy at initial screening, an increase of 15%. If we consider only the 5.1 mGy received at assessment, the cancer detection/induction ratio increases by a factor of 10/1.15 = 8.7 compared with its value at initial screening; if the doses at each stage are combined to give 9.6 mGy, this factor becomes 10/2.15 = 4.7. In both cases these factors are doubled if the assessment recall rate is 5% rather than 10%.
Table 5 shows ratios of cancers detected/induced at various ages for the UK Breast Screening Programme. Column two shows values previously published [1], based on the most recent survey of mean glandular breast doses, and breast cancer induction risk factors from NRPB [1, 10]. These values refer to two-view screening. Column three shows values for first stage assessment, assuming a 10% recall rate; if the recall rate were 5%, these values would be twice as large. Values in column three are greater than those in column two by the factor of 8.7 as outlined in the previous paragraph. Values in column four result from combining the mean glandular dose received at the initial screening visit with that received at first stage assessment, and are obtained by multiplying the values in column two by the factor 4.7 from the previous paragraph. Again, the resulting detection/induction ratios would be doubled if the assessment recall rate were 5% rather than 10%.
View this table:
[in this window]
[in a new window]
|
Table 5. Cancers detected/induced at first stage assessment: two-view screening. Assumes 10% of women are assessed (these numbers would be doubled if the assessment rate were only 5%)
|
|
Uncertainties in the numbers quoted in column two of Table 5 have been discussed in the paper from which these figures have been taken. They will be at least a factor of two in each direction due to the uncertainty in the underlying values for cancer induction per unit dose. Uncertainties in the numbers in columns three and four will be further increased because of the variability of protocol and practice between screening centres at the assessment stage. These considerations do not, however, affect the main conclusion that the benefit/risk balance is much more favourable at first-stage assessment than at initial screening.
The numerical risk factors for numbers of cancers induced per unit dose, which underlie all these ratios, may change at some time in the future. Any such change will not affect the relationship between ratios given in column two and those of columns three and four in Table 5, and thus will not affect our main qualitative conclusion.
Thus, whatever the relation between these figures and benefit/risk ratios, assuming that this relation is similar at all screening stages, the benefit/risk ratio is greater at first stage assessment than at initial screening, despite the greater dose level, because of the greatly increased cancer detection rate. At later stages of assessment, when biopsy or needle-aspiration and stereoscopic X-ray pairs may occur, the dose will be greater still, but again the cancer detection rate will rise, typically to the region of 30–50% as against 6 per 1000 at initial screening, or 6–12% at first stage assessment, so that again the benefit/risk ratio will increase compared with that at earlier screening stages.
At the initial screening stage, the female population is assumed to be symptom-free and therefore "healthy". For this reason they are not called patients. Thus, at first stage assessment, the status of an individual may change from being one of a selected group of the general population to a patient, since there are now specific grounds to suspect that disease may be present, which must be confirmed or excluded. For patients, justification of an examination disregards prior exposure, for whatever reason. It is therefore right that at this stage particular regard should be paid to ensuring that the balance of benefit and risk is favourable. At second stage assessment there is a relatively high probability that disease is present. Many other patients will routinely receive X-ray examinations to confirm or exclude the presence of a disease, and often where the probability of disease being present is less than it is at this stage of breast screening. Justification of X-ray exposure at second stage assessment should thus certainly be made as it is for other patients. At first stage assessment, this point is perhaps less clear-cut, but the recalls have been made on suspicion of disease being present at a probability in the region of 5–10%. Many hospital X-ray examinations will be routinely performed at a lower probability of disease or fracture than this, and the women concerned cannot be regarded as "healthy" in the way they are at initial screening.
The collective dose to the population of screened women is affected by the recall to assessment rate. Women who attend for first stage assessment receive an additional mean glandular dose of 10.2 mGy per woman (Table 4). If this were to be averaged across the screened population, then it would lead to an increase in the mean glandular dose to a typical woman of 1 mGy to 5.5 mGy assuming a recall rate of 10%. If the recall rate was 5%, then the mean glandular dose to a typical woman would be 5.0 mGy. However, this is not the approach we have chosen to adopt, as justification of an examination disregards prior exposure.
Regardless of these considerations, once it appears that breast cancer may be present it is essential that all necessary X-ray exposures are made to confirm or reject that possibility. The dose levels reported in this paper for first stage assessment are such that justification of the exposure on the balance of benefit and risk should be straightforward. Thus, there should be no problem in terms of benefit/risk balance in taking whatever films are necessary to this end. As in any X-ray investigation that does not imply that films should be taken without good cause. Similar arguments should also apply at second stage assessment. The worst possible outcome of a screening visit is a false-negative or a false-positive report.
|
Acknowledgments
|
|---|
We wish to thank the Superintendent Radiographers in Breast Screening Centres throughout the North East, Yorkshire and the Humber Region and Scotland for answering our questions and supplying information on screening procedures and numbers of films taken at assessment. This research was partially supported by the European Commission's Radiation Protection Research Programme (DIMOND III and SENTINEL Projects).
Received for publication April 7, 2005.
Revision received July 20, 2005.
Accepted for publication August 31, 2005.
|
References
|
|---|
- Law J, Faulkner K. Cancers detected and induced, and associated risk and benefit, in a breast screening programme. Br J Radiol 2001;74:1121–7.[Abstract/Free Full Text]
- National Health Service Breast Screening Programme Reviews, 1994-5 to 2003-4. Statistical Bulletin, London: Department of Health
- NHS BSP Annual Review 2004 Changing Lives. Sheffield NHS BSP 2004.
- Young KC, Faulkner K, Wall BF, Muirhead C. Review of Radiation risk in Breast Screening. NHS BSP Report 54; Sheffield, NHS BSP 2003.
- Law J, Faulkner K. Concerning the relationship between benefit and radiation risk, and cancers detected and induced, in a breast screening programme. Br J Radiol 2002;75:678–84.[Abstract/Free Full Text]
- Young KC. Radiation doses in the UK trial of breast screening in women aged 40-48 years. Br J Radiol 2002;75:362–70.[Abstract/Free Full Text]
- Young KC, Burch A, Oduko JM. Radiation doses received in the UK Breast Screening Programme in 2001 and 2002. Br J Radiol 2005;78:207–18.[Abstract/Free Full Text]
- Law J. Breast dose from magnification films in mammography. Br J Radiol 2005;78:816–20.[Abstract/Free Full Text]
- Faulkner K, Bennison K. An assessment of digital stereotaxis in the NHSBSP. Radiat Prot Dosim 2005. 117:327–9.
- Stokell PJ, Robb JD. SPIDER-1 Software for evaluating the detriment associated with radiation exposure. NRPB-SR261. Chilton: NRPB 1994.
This article has been cited by other articles:
|
|
|
|
 
BJR review of the year - 2006
Br. J. Radiol.,
March 1, 2007;
80(951):
147 - 151.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
