This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Law, J Articles by Faulkner, K Search for Related Content PubMed PubMed Citation Articles by Law, J Articles by Faulkner, K

Two-view screening and extending the age range: the balance of benefit and risk

¹ Edinburgh University Department of Medical Physics, Western General Hospital, Edinburgh EH4 2XU and ² Quality Assurance Reference Centre, Newcastle General Hospital, Newcastle-upon-Tyne NE4 6BE, UK

	Abstract

Top Abstract Introduction Method of calculation and... Discussion Conclusion References

 
The UK Breast Screening Programme intends to introduce two-view screening on all screening rounds, and also to extend the normal screening age range at its upper end from the present 64 years to 70 years. The implications of these changes for cancer detection:induction ratios and for benefit/risk ratios are discussed. It is shown that both ratios remain much greater than 1.0 at all screening ages. This is also true for younger women, subject to provisos regarding starting ages for annual screening as described in previous papers. The requirements for optimization in a breast screening programme are also discussed. It is suggested that benefit (rather than the benefit/risk ratio) should be maximized provided that benefit exceeds risk.

	Introduction

Top Abstract Introduction Method of calculation and... Discussion Conclusion References

 
The UK Breast Screening Programme (NHS BSP) will introduce two-view screening on all screening rounds by the end of March 2003, and not just on the first round as hitherto. In addition, the age range will be extended up to age 70 years by the end of March 2004. With a screening frequency unchanged at 3-yearly, this gives two additional screening rounds for all participants. As with all screening programmes, it is necessary to consider the "justification", in radiological protection terms, of these initiatives.

A trial of two-view vs single-view screening has shown a 24% increase in the cancer detection rate [1]. Whatever the magnitude of the benefit of breast screening may be, it is reasonable to suppose that it will vary in proportion to the cancer detection rate [1]. An increase of this magnitude may well be more than sufficient to justify the cost of the extra resources required. However, if the radiation dose is doubled in achieving this, it is not immediately obvious that such a dose increase is justified in the radiological protection sense of the term, as well as in the economic sense. The aim of this paper is to show that, in absolute terms, and for the NHS BSP, the increase in benefit is likely to exceed by a large margin the associated increase in radiation risk. A further aim is to consider the question of two-view screening in women below the age of 50 years, with and without family history of breast cancer.

For a procedure to be justified in radiological protection terms, the resulting benefit must exceed any associated risk or the adverse consequences of any side-effects. In the case of breast screening, optimization requires that the benefit be maximized, but with proper regard to the consequent risk. Sometimes this may correspond to maximizing the benefit/risk ratio, but that is not necessarily the optimum approach in all cases. In a breast screening programme the principal aim is a long-term reduction in mortality due to breast cancer, and it is this that should be maximized, subject to the constraint that it must always exceed any mortality increase due to the procedure. In practice this means that effort should be directed primarily to improving the detection rates for breast cancer.

	Method of calculation and results

Top Abstract Introduction Method of calculation and... Discussion Conclusion References

 
In the NHS BSP, two-view screening has been used in the first round since 1995. To assume a strict doubling of dose from single-view to two-view screening might appear reasonable, though as will be shown later that would be incorrect and an over-simplified view. On such a crude basis, for the current screening age range of 50–64 years and a total of five screening rounds per person, the relative increase in total dose to the screened population would be given by 4/6, i.e. 67%. This is because the first screening round already carries a "double dose". On the same basis, for a programme extended to age 70 years, and hence with two further screening rounds per person, the relative dose increase for two-view against single-view screening would be given by 6/8, i.e. 75%. However, because both benefit and radiation risk vary with age of the woman screened, the balance of benefit and risk must be considered at different ages or at each successive screening round. Thus, this modifying factor of 67% or 75% of dose is inappropriate and cannot be applied to women in later screening rounds.

In practice, and quite independently of that point, radiation doses do not double when moving from single to two-view screening, because the compressed breast thickness, and hence the radiation dose, varies with projection. Single-view screening employs the lateral–oblique (LO) view only, while two-view screening adds to this the craniocaudal (CC) view. The latest information available to us for doses in the NHS BSP comes from the survey by Young and Burch [2] and a further paper by Young [3]. This second paper by Young is mainly concerned with doses measured in the UK trial of screening of women aged 40–49 years, but it also describes in outline some recalculations of doses given in the earlier paper [2] to take account of factors to correct for breast composition and X-ray spectra. The mean dose per film for the oblique view (previously 2.03 mGy) is now given as 2.35 mGy, while for the CC view the figures are 1.65 mGy and 1.86 mGy, respectively. The corresponding mean doses per woman in the NHS BSP are not given in the second paper [3], but in the earlier survey [2] a value for the oblique view of 2.14 mGy per woman corresponds to 2.03 mGy per film, an increase of 0.11 mGy (5.4%). Applying this increase to the recalculated doses and rounding off slightly gives 2.5 mGy and 2.0 mGy as the best current estimates for the LO and CC views, respectively, in the NHS BSP, i.e. for women over 50 years of age. Thus the dose increase for two-view screening on each screening round is 80% rather than 100%.

For younger women aged 40–49 years, Young's later paper [3] reveals that, when their doses are compared with those for women over 50-years-old on the same imaging and processing equipment in the same centres, the increase is less than 2% for the LO view and less than 6% for the CC view, and is not statistically significant in either case. Accordingly, the same rounded doses of 2.5 mGy and 2.0 mGy for the two views will be used for all age groups in the following discussion.

Numerical values for risk factors for potential cancer inductions due to X-rays have been provided by the National Radiological Protection Board (NRPB). These are considered to be those most appropriate for a UK population and refer to 5-year age bands [4]. They are the same as those used in an earlier paper [5] and are given in Table 1. For doses described earlier in this section, estimates of potential cancer inductions derived from these factors are given in Table 2a for ages 50–70 years. The latest available cancer detection rates are also taken from that earlier paper [5] and these, increased by 24% for all screening rounds except the first, are also given in Table 2a. These detection rates increase with age but only apply up to ages 60–64 years. The Age Extension Trial of screening, in the age range 65–69 years, has recently reported cancer detection rates in this age range [6], and their value of 8.1 per 1000, for women screened in the previous 5 years, is used here. For women not screened in the previous 5 years, the detection rate was even higher at 8.7 per 1000. Data are available from the NHS BSP for self-referrals and GP referrals in women aged 45 years and over, and these show cancer detection rates around 50% greater than for invited women aged 60–64 years. However, this sub-group will not be typical of the general female population of that age range.

	Discussion

Top Abstract Introduction Method of calculation and... Discussion Conclusion References

 
A number of studies have been carried out on the effect on cancer detection rates of two-view compared with single-view screening, and increases reported have varied considerably [1, 9, 10]. Some of these studies have been based on comparison of past radiographs known to show cancers with a small number of controls, e.g. [9], and show an increase of only 9%. It has been pointed out by Wald et al [1] and also by Blanks et al [10] that this approach tends to lead to underestimation of the true increase. Some other studies have been combined with comparisons of double and single film reading, and increases greater than 24% have also been found. At present the study by Wald et al [1] appears to be the only randomized, controlled clinical trial of two-view screening, so that their value of 24% is taken to be the best estimate for the increase in cancer detections. It is assumed here that the 24% extra cancers detected from the second view will have the same prognosis, and hence potential benefit, as those already being detected from single-view screening. This 24% increase has so far been observed only in the first or prevalent screening round, but it was also observed that the increase in detection rate for cancers less than 15 mm maximum diameter, i.e. those likely to have a better prognosis, was 45% [10]. Thus the increase in benefit is likely to exceed 24%, though at the present time this probable greater increase cannot be so easily quantified. A further assumption is also made that two-view screening in later or incident rounds will show a similar 24% increase. This assumption has been suggested as reasonable by Wald et al [1].

It is understood that the decision by the NHS BSP to adopt two-view screening on all rounds was based on the 24% increase in cancers detected reported by Wald et al [1], reinforced by the 45% increase reported for smaller cancers [10], which are likely to have a better prognosis.

For the present NHS BSP age range of 50–64 years, the ratios of detections/inductions are all over 100 both for single- and two-view screening. At each age band the ratio for two-view screening is less than that for single-view, as must be expected because the detection rate has increased by a smaller proportion (24%) than has the potential cancer induction rate (80%). When the ratio of the increase in detections to the increase in inductions is considered, this is smaller again than the ratio for two-view screening, but still quite large and it also increases with increased age of the women screened.

This ratio of cancer detections to inductions is not a true measure of the benefit/risk ratio, but is nevertheless one indication of its likely magnitude. The relationship between these two ratios has been explored in another recent paper [11] where it is suggested that the benefit/risk ratio is probably not much less than the detection/induction ratio, the difference possibly being as little as 15–20%, and certainly smaller than the uncertainties in either ratio. The relationship depends on the reduction in mortality due to breast cancer that can be attributed to screening, and the future mortality of any radiation induced breast cancers due to screening. Some data on both these quantities are available, but are less firmly established than present screening detection rates or past mortality due to breast cancer. It is also pointed out in that paper [11] that the degree of risk from any radiation-induced cancers will depend on whether women continue to obtain screening in later life, beyond the normal screening age range, whether this be up to 64 years or 70 years. This is because there is a time delay of approximately 10 years between radiation exposure and the first appearance of any consequent breast cancers, so that a proportion of any cancers induced by screening will appear in women who are over the screening age range. If all women were screened from age 50 years for the rest of their lives, benefit/risk would be an even greater proportion of the detection/induction ratio.

From that standpoint alone, the recent decision to extend the age range of the NSH BSP to 65–70 years is to be welcomed as it will increase the number of cancers detected. The predicted ratios of cancer detection to inductions given in Table 2b are all greater for this age band than for any other, so that if they are satisfactory at ages 50–64 years they are even more so at this older age band.

Fears are sometimes expressed regarding the long-term effect of the increasing cumulative X-ray dose with each screening round. This cumulative dose must increase when all women are invited for two further screening rounds. Such fears, however, are misplaced since further cancers are detected at each screening round, regardless of and entirely additional to those detected in any previous round. Moreover, the detection rate is found to increase with each successive screening round, in line with the rising incidence of breast cancer with age, while the induction rate decreases with age.

In discussing the ratios of benefit/risk and of cancer detection/induction, it should be remembered that the latter ratio can be estimated with reasonable confidence. The greatest source of uncertainty for this ratio is the factor of at least two on the induction risk factors per unit X-ray dose [4]. In contrast, any estimate of benefit/risk will carry an uncertainty compounded of that factor of at least two and various other sources of uncertainty regarding treatment outcome, few of which are likely to be small. Clearly the numbers of cancers detected should exceed the numbers induced, preferably by a fair margin, though it may be difficult to make a precise estimate of how large that margin should be. In an earlier paper [5] it was suggested that this ratio should exceed a factor of approximately 5 if we are to be reasonably confident that benefit exceeds risk in the NHS BSP, taking all sources of uncertainty into account. In Table 2b all values of the ratio exceed 5 by a very large margin. Nevertheless, the reader must appreciate that the uncertainties on the risk factors are rather difficult to quantify. Much depends on the assumptions and risk models employed. The figures used here are those provided by the NRPB [4] and are considered by them to be the best values currently available for the UK population.

In discussing optimization of breast screening, and in relation to the balance of benefit and risk, it might be supposed that the optimum balance would be represented by maximizing the benefit/risk ratio. This is not the position of the NHS BSP, since all values of cancers detected and/or induced in Table 2b are lower for two-view than for single-view screening. However, the aim of the NHS BSP is to maximize mortality reduction from breast cancer, and this is likely to be achieved by maximizing cancer detections, especially small cancers at early stages of development. Thus optimization of breast screening does not mean maximizing the benefit/risk ratio but achieving the maximum cancer detection rate, subject only to the constraint that this should sufficiently exceed the potential induction rate, e.g. by the factor of 5 already mentioned. When making a change to a screening programme, such as the introduction of two-view screening, it is desirable that the resulting increase in benefit should exceed the increase in risk. For the interest of readers, the ratios of these increases are also given in Tables 2b to 4b, and in every case they exceed 1.0. However, this is not a strict requirement; it is only necessary that benefit is increased by the change and that the final total benefit exceeds the final total risk.

For younger women, with or without a family history of breast cancer, Table 3(b) and Table 4(b) show that detections exceed inductions by more than a factor of 5 at all ages considered, for a 3-yearly screening interval. If annual screening is employed, this is only true above age 35 years, or above age 40 years if this margin is required for the ratio of increases of levels of detections and inductions.

For women with a family history of breast cancer given annual two-view screening, detections are predicted to exceed inductions with sufficient margin for uncertainty only above age 30 years. The factor of two uncertainty on the cancer induction risk data may rise to a factor of four for women below 30 years of age, so that even greater caution may be advisable in that age group. The proviso made in earlier papers regarding the possibility of women with family history having enhanced susceptibility to X-ray induction of breast cancer still stands. If radiation risk of cancer induction is broadly proportional to the underlying natural incidence of cancer of the same type, then those with increased risk of developing breast cancer may have enhanced radiation susceptibility in exactly the same proportion. If that were so, then the results given in Table 3 would apply to the "family history" group as well as to those without a family history of breast cancer. However, it is not yet established that radiation risk is proportional to natural incidence of breast cancer to the degree that it may be for some other cancers.

There seems to be some tendency for women below the NHS BSP age range to obtain screening for themselves if they can do so, sometimes at annual intervals. Where this occurs it is often outside the NHS, at centres that are not covered in surveys of dose and image quality carried out for the NHS BSP. Even where their standards match those of the NHS BSP, based upon this data it would seem best to avoid annual screening below age 30 years for those with family history, and below age 35 years for those without.

	Conclusion

Top Abstract Introduction Method of calculation and... Discussion Conclusion References

 
The introduction of two-view screening in the NHS BSP is expected to increase the number of cancers detected by around 24% while increasing the total X-ray dose to the screened population by approximately 80%. Total numbers of cancers detected will still far exceed the total numbers induced by radiation, and the increase in numbers detected will also exceed the increase in numbers induced. The corresponding benefit/risk ratios will also greatly exceed 1.0, even though they may be slightly lower than the detection/induction ratios. Optimization in breast screening is best represented by maximizing benefit, provided benefit also exceeds risk.

The age range extension to 70 years is expected to increase the cancer detection/induction ratio because detection rates increase with age while induction rates decrease.

Similar conclusions also apply to the screening of younger women, both with and without a family history of breast cancer, but subject to caution below ages 30–35 years as discussed in an earlier paper [5].

	Acknowledgments

 
We thank Mr B F Wall of the NRPB, Dr R G Blanks and Dr S M Moss of the Institute of Cancer Research and Dr K C Young of the National Physics Centre for Mammographic Screening, for providing information and for helpful discussions.

Received for publication February 12, 2002. Revision received May 29, 2002. Accepted for publication May 31, 2002.

	References

Top Abstract Introduction Method of calculation and... Discussion Conclusion References

 

1. Wald NJ, Murphy P, Major P, Parkes C, Townsend J, Frost C. UKCCCR multicentre randomised controlled trial of one and two view mammography in breast cancer screening. BMJ 1995;311:1189–93.[Abstract/Free Full Text]
2. Young KC, Burch A. Radiation doses received in the UK BSP in 1997 and 1998. Br J Radiol 2000;73:278–87.[Abstract]
3. Young KC. Radiation doses in the UK trial of breast screening in women aged 40–48. Br J Radiol 2002;75:302–70.
4. Stokell PJ, Robb JD. SPIDER-I software for evaluating the detriment associated with radiation exposure, NRPB-SR261. Chilton, UK: NRPB, 1994.
5. 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]
6. Moss SM, Brown J, Garvican L, Coleman DA, Johns LE, Blanks RG, et al. Routine breast screening for women aged 65–69: results from evaluation of the demonstration sites. Br J Cancer 2001;85:1289–94.[Medline]
7. Law J. Cancers detected and induced in mammographic screening: new screening schedules and younger women with family history. Br J Radiol 1997;70:62–9.[Abstract]
8. Houlston RS, McCarter E, Parbhoo S, Scurr JH, Slack J. Family history and risk of breast cancer. J Med Genet 1992;29:154–7.[Abstract/Free Full Text]
9. Blanks RG, Given-Wilson RM, Moss SM. Efficiency of cancer detection during routine repeat (incident) mammographic screening: two versus one view mammography. JMS 1998;5:141–5.
10. Blanks RG, Moss SM, Wallis MG. Use of two view mammography compared with one view in the detection of small invasive cancers: further results from the National Health Service breast screening programme. JMS 1997;4:98–101.
11. 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]

This article has been cited by other articles:

				K. Faulkner Ethical concerns arising from screening procedures such as mammography and self-referral Radiat Prot Dosimetry, July 1, 2009; 135(2): 90 - 94. [Abstract] [Full Text] [PDF]

				C. J. Martin RADIATION DOSIMETRY FOR DIAGNOSTIC MEDICAL EXPOSURES Radiat Prot Dosimetry, March 28, 2008; (2008) ncm495v1. [Abstract] [Full Text] [PDF]

				E. Samei, R. S. Saunders Jr, J. A. Baker, and D. M. Delong Digital Mammography: Effects of Reduced Radiation Dose on Diagnostic Performance Radiology, May 1, 2007; 243(2): 396 - 404. [Abstract] [Full Text] [PDF]

				J Law, K Faulkner, and K C Young Risk factors for induction of breast cancer by X-rays and their implications for breast screening Br. J. Radiol., April 1, 2007; 80(952): 261 - 266. [Abstract] [Full Text] [PDF]

				G J Heyes, A J Mill, and M W Charles Enhanced biological effectiveness of low energy X-rays and implications for the UK breast screening programme. Br. J. Radiol., March 1, 2006; 79(939): 195 - 200. [Abstract] [Full Text] [PDF]

				K. Faulkner and K. Bennison An assessment of digital stereotaxis in the National Health Service Breast Screening Programme Radiat Prot Dosimetry, December 1, 2005; 117(1-3): 327 - 329. [Abstract] [Full Text] [PDF]

				K. Faulkner and J. Law Mammographic breast cancer screening for women previously treated with high breast doses for diseases such as Hodgkin's Radiat Prot Dosimetry, December 1, 2005; 117(1-3): 330 - 333. [Abstract] [Full Text] [PDF]

				J Law Breast dose from magnification films in mammography Br. J. Radiol., September 1, 2005; 78(933): 816 - 820. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Law, J Articles by Faulkner, K Search for Related Content PubMed PubMed Citation Articles by Law, J Articles by Faulkner, K