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

Concerning the relationship between benefit and radiation risk, and cancers detected and induced, in a breast screening programme

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 Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
In a breast screening programme based upon X-ray mammography it is necessary to demonstrate that benefit, from reduced mortality arising from earlier diagnosis, exceeds any potential risk from future induction of breast cancers by ionizing radiation. A rigorous treatment of this problem would be both complex and subject to large statistical uncertainty, even if all necessary data were available. A more simplified approach is to show that the number of cancers detected exceeds the number potentially induced by a sufficient margin. These numbers are relatively well established, but this approach is less satisfactory owing to the question of what would constitute a sufficient margin. This paper attempts to explore a possible relationship between the detection/induction ratio and the benefit/risk ratio, using treatment outcome data from three independent sources and mortality reduction data. Agreement between these four sources is considered to be fair, given the nature of the data. The future screening of older women (over 65 years) is also found to have a significant effect on the final outcome. When current trends in such screening are allowed for, the benefit/risk ratio is found to be only marginally less than the detection/induction ratio.

	Introduction

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
In any screening programme involving the use of ionizing radiation to detect cancer, it is necessary to show that the benefits of screening outweigh the radiation risks, quite apart from the risk of any other adverse outcome such as an incorrect diagnosis of cancer. In a breast screening programme, benefit should arise from earlier detection and diagnosis leading to improved prognosis, while the radiation risk is that of the potential induction of breast cancer at some time in the future. Benefit should take the form of reduced mortality, either specifically from breast cancer or from all causes, over a specified time span, or of increased lifespan. Either of these benefits may be compared with the corresponding increase in mortality, or decrease in lifespan, arising from any breast cancers induced by the use of ionizing radiation.

All estimates of changes in mortality or lifespan are subject to considerable statistical uncertainties. These arise from the practical difficulty (or impossibility) of complete ascertainment of treatment outcome data for any cohort of screened women, uncertainty regarding the extent to which that outcome was influenced by the screening procedure, and the statistical nature of treatment outcome data itself. For example, percentage survival over 5 years, 10 years or 20 years varies with the stage at which the disease is first detected, and the main objective of screening is to increase the proportion of patients first presenting at earlier stages and who therefore have a better prognosis. Such an approach provides one possible means of estimating benefit, and of the corresponding radiation detriment, but it is always subject to a combination of uncertainties.

A knowledge of the numbers of cancers detected, and also of those predicted to be induced, by a breast screening programme forms one foundation for the assessment of the balance between benefit and risk or detriment. These numbers are relatively easy to determine, though still subject to some degree of statistical uncertainty. Detection rates in past years for the National Health Service Breast Screening Programme (NHSBSP) are well established. These rates can be used to estimate future detection rates to within ±10–20% even when those rates are rising [1], due to the combined effects of improved techniques and greater experience.

Cancer induction rates are predicted from theproduct of breast radiation doses and cancer induction risk factors. Doses for screened populations can be estimated with some confidence [2] while doses for individual women can be estimated retrospectively with comparable confidence. Numerical values of cancer induction risk factors for exposure to ionizing radiation are still uncertain to within a factor of approximately two, but the estimate for breast cancer is probably better established than for any other organ. Therefore, the ratio of cancers detected/induced can usually be estimated to within little more than a factor of two.

In view of the additional and substantial uncertainties involved in any true estimates of benefit and detriment, some authors have restricted their discussion of this problem to numbers of cancers detected and induced, recognizing the need to demonstrate that detections must exceed inductions by a significant margin if we are to be confident that benefit exceeds detriment. The magnitude required for that margin is then left as a matter of subjective but informed judgement. A ratio of 100 might be thought to be ample, while a ratio of 10 may also be sufficient [3].

Survival and mortality data present particular problems in breast cancer. Whereas with a number of other forms of cancer 5-year survival may be a good index of "cure" in that subsequent life expectancy for patients equals that for a healthy population of the same age, life expectancy of breast cancer patients may take 20 years or more to approach this condition [4]. This tendency for breast cancer to recur over long periods affects both mortality and lifespan studies. It is therefore difficult to predict the benefits of screening based on actual survival studies until 20 years after introduction.

Numbers of cancers detected and likely to be induced in the current UK breast screening programme, NHSBSP, have been compared in another paper [5], and the amount by which the former should exceed the latter is briefly discussed there. This paper discusses several different types of data for survival and mortality and attempts to relate such data to the detection/induction ratio, the aim being to estimate the relationship between detection/induction and benefit/detriment in terms of breast cancer mortality reduction arising from screening and fatal cancers potentially induced by radiation.

	Sources of survival data

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
Four sources of data will be used, each representing a different type of data.

Nottingham Prognostic Indicator (NPI)
This has been in use for a number of years. It is a numerical value, equal to the sum of (0.2 x size in centimetres) + stage + grade. The lower the NPI, the better the prognosis. It was first derived empirically from observation and case records, and later verified in a study of 1662 further cases [6]. 15-year survival data for three different ranges of NPI values are available. The distribution of NPI values in a screened population is generally lower than in an unscreened population [6].

Any problems arising from lead time in relation to percentage survival after a stated period are taken into account by considering changes to the distribution of cases between recognized stages, i.e. the effect of screening is to increase the proportion of cases presenting at earlier stages having a better prognosis and, correspondingly, to decrease the proportion of cases presenting at later stages having a poorer prognosis.

Surgical data
5-year survival data are available for those treated by surgery in the Yorkshire region both just before and just after the introduction of the NHSBSP [7].

Radiotherapy data
5-year survival data stratified by stage are available and can be combined with changes in distribution between stages before and after the introduction of screening (G R Kerr, personal communication).

Long-term mortality studies
Two recent papers [8, 9] have reported long-term reduction in mortality due to breast cancer in women invited for screening during the trials of early detection of breast cancer by mammographic screening in Edinburgh and Guildford between 1979 and 1987. Both papers report mortality reductions of around 30% over periods of 14years and 16 years.

It might be hoped that the results to be obtained from using these four different sources of data would ideally agree with each other, but any such hope is unrealistic. Considerable variation is only to be expected, and this variation may provide some indication of the general uncertainty or reliability of any of the underlying data of this kind. The fact that the four sources are so varied may, in this sense, be an advantage.

	General treatment of survival data from all sources

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
Screening aims to detect cancers earlier than would happen otherwise, with a consequent increase in percentage survival at any given time after first treatment. The benefit of screening is not the proportion of those with screen detected cancers who survive a given period, but the difference between that proportion and the corresponding proportion who would have been predicted to survive that long if they had been in a comparable unscreened population. The risk from cancer induction may also be considered correspondingly as the proportion of inductions which prove fatal, also over a stated time period. However, to relate cancer detections and inductions to benefit and risk, it is necessary to consider outcomes both with and without future screening of older age groups, i.e. those older than the normal screening age range. There is a latent period of approximately 10 years between radiation exposure and the appearance of the first cancers arising from this exposure. It is therefore necessary to consider whether such future cancers will be detected by a screening programme or by symptomatic services. Given the present screening age range of 50–64 years and this 10 year latent period, a proportion of any cancers that may be induced will occur in older women who are not then being screened. This will remain true to some extent, even when the screening age range is extended to 70 years in the next few years, and irrespective of any rising trend of screening self-referrals in those aged 65 years and over.

In relating benefit and risk to numbers of cancers detected and induced, one possible way to proceed is as follows. Of the cancers detected by screening, only a proportion of those individuals will survive for any specified subsequent period. Let this proportion be denoted by A%. It is assumed that all these cancers would, if not detected by screening or in the absence of any screening service, develop to the point where they would be detected by other services following the onset of symptoms. Of the cancers detected symptomatically by non-screening services, as with cancers detected by screening, only a proportion of individuals will survive for any specified subsequent period. This proportion may be denoted byB%. B will be less than A because a greater percentage of screen-detected cancers present at earlier stages than do those cancers detected symptomatically, and correspondingly have a better prognosis. Hence, the benefit attributable to screening in terms of "lives saved", i.e. decrease in lives lost to breast cancer in the specified subsequent period, is given by cancers detected x (A–B)%.

Of the small number of cancers induced by the small radiation dose arising from the screening process, not all will be fatal. Percentage mortality in this group may be denoted by M%, where M% also equals 100-percentage survival in this group. Because of the 10-year latent period already referred to, the majority (though not all) of these cancers are likely to appear in women over the normal routine screening upper age range, at present 64 years. Thus two quite different values for M will each need to be considered separately, one for women who receive screening in all subsequent periods, e.g. by self-referral as encouraged in the NHSBSP, and the other for women who receive no subsequent screening and whose survival rate is that of symptomatically detected cases. Hence, the risk attributable to screening in terms of "lives lost" to breast cancers induced by screening itself, is given by cancers induced x M%, where M may take either of the values described above. Thus, one possible measurement of benefit/risk may be taken to be:

This expression may be re-written as:

M is defined here in terms of symptomatic detection because it must be assumed at present that the majority of induced cancers will appear in women over the age of 64 years who do not refer themselves for screening. If screening of all older age groups in future years could be assumed, themortality of screen detected cancers would replace that for cancers detected symptomatically. Because of the imminent extension of the screening age range, and the increasing number of self-referrals by women over the upper screening age limit, the proportion of women aged 65 years and over who receive screening will continue to increase. Therefore, any value of M based on symptomatic detection only will be an underestimate. This point will be discussed further later in this paper.

The foregoing approach may be applied to treatment outcome data, since these are commonly presented in terms of percentage survival in cancer cases. With mortality reduction data, a slightly different approach is required since such data are normally derived from breast cancer mortality, e.g. breast cancer deaths per 100 000 popu<~?show=[fo]>lation, without reference to numbers of breast cancer cases. Nevertheless, there must be a strong link between these numbers, since mortality reduction will only result from improvements in cancer case survival. In as far as we are comparing situations with and without screening as well as the proportionate changes resulting from screening, the proportionate change in percentage survival must be equivalent to the proportionate change in mortality, i.e. mortality reduction, provided the timescale is sufficiently long and no other material changes occur, such as a change in breast cancer incidence. In the data discussed below the timescale is 14–16 years, which, if not ideal, is better than 5-year survival data.

Thus, from the equations above, (A-B)=percentage mortality reduction attributed to screening, and M=percentage mortality in the absence of screening, or alternatively in the presence of screening if all older women continued to be screened.

In equation (2), (A-B)/M represents the conversion factor (C) from detection/induction to benefit/risk ratios. Numerical values for this factor will now be estimated from treatment outcome data, and from mortality reduction data, obtained from the various sources outlined above.

	Results

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
Nottingham Prognostic Indicator
Of the various sources of data to be considered, that based on the NPI appears to be the most firmly based, and it is certainly the easiest to interpret and apply. It will therefore be given correspondingly greater weight in the ensuing discussion. Data for 15-year survival for three different value ranges of NPI are given in Table 1. Data are also available for the proportions of women falling within these three NPI value ranges both with and without screening. Values of C are deduced from these data and given in Table 2, again with and without future screening of older women. Details of the calculation involved are given in the Appendix.

Radiotherapy data
Data for 5-year survival by stage have been obtained from Edinburgh for two 1-year periods; 1986–1987 (before national screening) and 1993–1994 (with screening fully established), and are given in Table 3 (G R Kerr, personal communication). These figures include those patients receiving either chemotherapy or radiotherapy, orboth. Women with the best prognosis would beoffered surgery alone, whereas radiotherapy would be offered in other cases. Accordingly, prognosis for that group of women treated by radiotherapy is poorer than for those receiving surgery. The figures for the later period include all ages, not only 50–64 years. Allowing for a screening uptake of around 75% in 1993–1994, only approximately one third of the women in those years' figures will actually have been screened. There is a further complication in that Edinburgh was one of the centres in the trial that preceded the NHSBSP, with half the women aged 45–64 years invited for screening. As 60% of these accepted, the women included in 1986–1987 will include a small percentage who had been screened.

View this table: [in this window] [in a new window]   Table 3. (a) Radiotherapy survival data used for values given in Table 2

Long-term mortality studies
Two recent papers [8, 9] have reported long-term reduction of mortality due to breast cancer of approximately 30% over periods of 14 years and 16 years, and they attribute these reductions to screening. This implies future mortality with screening of 70% of its previous value without screening. Hence, conversion factors to benefit/risk of (100-70)/100=0.30 without future screening of older women, or 30/70=0.43 with future screening, can be derived. However, the 30% mortality reduction was observed in women invited for screening, of whom between 60% and 72% actually received first round screening, with somewhat lower rates for later rounds [8, 9]. This suggests that the mortality reduction in women screened may be approximately 45%, which would lead to values of C of 0.45 without future screening of older women, or 45/55=0.82 with future screening.

Statistical estimates of errors on data in Table 1 and Table 3 are not available in the original sources from which they have been obtained and, consequently, are not provided here. Therefore estimates of uncertainty on single values of C given in Table 2 can only be speculative and, in the authors' view, not worth making. Estimates of uncertainty on mean values of C in the final column of Table 2 would require numerical weighting values to be applied to each of four values on which the means are based, and these would be rather subjective as outlined in the next section. Also, even if equal weighting were applied, any standard error would assume a normal distribution of the underlying four values, which is itself open to question.

	Discussion

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
All calculations in this paper assume that changes in survival, the relative distribution of presenting cases between recognized stages, or in treatment outcome, which can be demonstrated within a few years before and after the introduction of the NHSBSP, are largely attributable to that programme. Improvements in treatment arising for other reasons, or effects of changes in demographic factors, are normally only detectable over periods measured in a few decades rather than a few years.

In attempting to convert from detection/induction to benefit/risk, data from a variety of sources have been used. Of these data sources, that from the NPI may be the best for this purpose. The NPI is derived from parameters that are relatively easy to determine, and its relationship to survival has been well established. Survival at 15 years is a much better basis than the 5-year survival widely used for other cancers. However, the age corrected 15-year survival of 96% for the group having the best prognosis does seem remarkably high. Mortality reductions over 14–16 years also have that advantage, plus the further advantage of avoiding lead-time bias, i.e. the possibility that screening merely detects cancer earlier with no effect on outcome or its timing, but it is more difficult to interpret and calculate implications of mortality reduction for radiological justification.

The radiotherapy data presented here may be a poorer indication of long-term survival because of the shorter time of follow-up, but are otherwise likely to underestimate benefit because they will not include those with better prognosis who receive surgery only. Surgery data are also based on 5-year survival and are taken from brief published summaries, but are derived from a large group of women. Although the mortality reductions are impressive in themselves, there may be problems in applying such figures as they stand.

Conversion factors to benefit/risk given in Table 2 vary considerably, but no more than might be expected from the different sources from which they are derived and the different end points for follow-up. Those derived from NPI data might be used on their own since, for reasons already stated, they are considered to be the most firmly based. They are also based on a large series. However, in view of those other sources of data, factors derived here from NPI may be slightly optimistic, and for the rest of this discussion the slightly lower mean values of Table 2 will be used.

Any numerical estimate of benefit due to screening will depend on the source of data chosen and on the number of years survival, but all the data imply that some mortality reduction has occurred since breast screening was introduced.

If a cancer is induced by radiation and later is successfully treated, it would be unfair to claim that this represented no detriment to the woman concerned, but quantification of such detriment is beyond the scope of this paper. It is not normally considered in papers that describe lives lost or saved, nor is it generally covered in discussions of years of life gained or lost.

The question of whether older women, e.g. 70–80 years, will be invited for screening in, for example, 10–15 years from now has a marked effect on the data from all four sources. That is likely to be the age group and time period in which any cancers induced by radiation now, and which do occur, will subsequently appear. The magnitude of this effect is at least comparable with the variation observed between results from the four sources of data. Even if it is assumed that no future screening is available for that age group (though self-referral of older women is increasing) a best estimate of 0.5 with a lower limit of 0.3 for the conversion factor seems reasonable. Nevertheless, screening of all older age groups would be very advantageous for the benefit/risk ratio, and could cause it to exceed the detection/induction ratio (Table 2).

If we consider the ratio of values of C with and without future screening of older women, its mean value, averaged over the four data sources, is 2.3±0.4. Thus, future screening of the whole of the older age group could more than double the benefit/risk ratio.

The current situation regarding screening of older age groups is already intermediate between the two extreme positions, all and none, which have been used for simplicity of calculation. Self-referral by women aged 65 years and over is currently only around 10% [11] with regional variations from 4% to 14%, but this is increasing. The NHSBSP has recently declared its intention of extending screening to the 65–70 years age group, which will bring at least 40% of the over 65 years age group into the population invited for screening. Assuming 70% compliance, this alone will result in nearly 30% of all women aged 65 years and over receiving screening. There will be some overlap of this 30% with the 10% who currently refer themselves, but the combined total for the future is likely to be around 40% of the over 65 years age group. Therefore the current best estimate for C, interpolating between the values of 0.5 and 1.3 shown in Table 2 and applicable when screening to age 70 years is fully introduced, will be approximately 0.8, with upper and lower limits probably around 1.1 and 0.5. It could therefore be said not to differ from unity with any statistical significance, leaving the detection/induction ratio fairly closely approaching the true benefit/risk ratio.

The value of C may well vary with the age group of those being screened, but treatment outcome data are not yet generally subdivided into age bands, so this point cannot be investigated at present.

Treatment outcome may be expected to improve with time for all kinds of treatment, though this improvement is likely to be slow and gradual. All such improvement should apply to some extent to both screen detected and symptomatically detected cases. Therefore the expression (A-B)% is expected to change little with time, if at all, while M will decrease with time and so C will increase. Hence, the benefit/risk ratio should increase in future in comparison with the ratio of cancer detections to inductions.

	Conclusions

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
All data used in this paper regarding treatment outcome and benefits arising from screening have some deficiencies. Nevertheless, they may still be useful as a reasonable guide and may be no poorer in the conclusions drawn from them than more rigorous methods, given the uncertainties attaching to the underlying data. Numerical values derived for the conversion factor C from any one source of data are not of the first importance and should only be used with considerable caution. Their average value may give some indication of the relation between detection/induction ratio and the balance of benefit and detriment. The approach adopted in this paper aims to assist those who choose to use the detection/induction ratio as a stepping stone or as an indicator in its own right, given the relative reliability with which it can be estimated. C may have general value, and is capable of retrospective application.

Perhaps the main point to emerge from these calculations is the relative importance of the question of screening women beyond the present upper age limit of the screening programme. This appears to have an effect at least as great as, if not greater than, the variation between different methods of estimating a benefit/risk ratio, and comparable with the factor of two uncertainty already mentioned for the cancer induction risk factor. The magnitude of that effect is likely to be relatively independent of which set of underlying data is used, and to that extent can probably be estimated more reliably than the benefit/risk ratio itself. The final outcome is that the benefit/risk ratio is now not much less than the simple detection/induction ratio as used in the past, which can be calculated with relative ease.

The method adopted in this paper can be criticized on many grounds. So, we believe, can every method in this field. This method is simple and can be applied to detection and induction data that are relatively firm and readily obtainable. The approach is new but seems worthy of some consideration.

	Appendix

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 
Calculation of conversion factor (C) from detection/induction to benefit/risk ratios from Nottingham Prognostic Indicator data in Table 1
Overall 15-year percentage survival before introduction of screening (from Table 1):

Overall 15-year percentage survival after introduction of screening (from Table 1):

Hence,

in the absence of subsequent screening, or

if screening continues for all subsequent years.

	Acknowledgments

 
We wish to thank Mr B F Wall of NRPB for discussion of breast cancer induction risk factors, Dr K C Young for discussion of dose data, Mr F Neilson for discussion of data on surgical outcome and Miss G R Kerr for provision and discussion of radiotherapy data.

	Footnotes

 
This research was partly funded by the European Commission's Radiation Protection Research Programme.

Received for publication August 23, 2001. Revision received January 29, 2002. Accepted for publication February 8, 2002.

	References

Top Abstract Introduction Sources of survival data General treatment of survival... Results Discussion Conclusions Appendix References

 

1. National Health Service Breast Screening Programme Reviews, 1994–5 to 1998–9 inclusive. Statistical Bulletin. London: Department of Health, 1996–2000.
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. Law J. Risks and benefits in mammography. In: K Faulkner, BF Wall (editors). Are X-rays safe enough? IPSM Report 55. York, UK: 1988.
4. Kerr GR, Kunkler IH, Langlands AO, Rodger A. (In)curability of breast cancer: a 30 year report of a series of 3933 cases. The Breast 1998;7:90–4.
5. Law J, Faulkner K. Cancers detected and induced, and associated risk and benefit, in a breast screening programme. Br J Radiol 2001;888:112–7.
6. Elston CW, Ellis IO, Goulding H, Pinder SE. Role of pathology in the prognosis and management of breast cancer. In: CW Elston, IO Ellis (editors). TheBreast (3rd edn). Edinburgh, UK: Churchill Livingston, 1998:419–22.
7. Quick Data Version 2. Northern and Yorkshire Cancer Registry and Information Service, Leeds.
8. Alexander FE, Anderson TJ, Brown HK, Forrest APM, Hepburn W, Kirkpatrick AE, et al. 14 years of follow-up from the Edinburgh randomised trial of breast cancer screening. Lancet 1999;353:1903–8.[Medline]
9. UK Trial of early detection of breast cancer group. 16 year mortality from breast cancer in the UK trial of early detection of breast cancer. Lancet 1999;353:1909–14.[Medline]
10. Audits of screen detected breast cancers for years 1997–8 to 1999–2000. Annual Reports of British Association of Surgical Oncology Breast Group. Sheffield, UK: NHSBSP, 1999–2001.
11. Reducing the risk. Annual Report of the NHS Breast Screening Programme, 2000. Sheffield, UK: NHSBSP, 2000.

This article has been cited by other articles:

				K. Faulkner, M.G Wallis, F Neilson, and C.J. Whitaker EVALUATION OF THE POPULATION DOSE TO THE UK POPULATION FROM THE NATIONAL HEALTH SERVICE BREAST SCREENING PROGRAMME Radiat Prot Dosimetry, May 14, 2008; (2008) ncn081v1. [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]

				J Law and K Faulkner Radiation benefit and risk at the assessment stage of the UK Breast Screening Programme. Br. J. Radiol., June 1, 2006; 79(942): 479 - 482. [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]

				K. Faulkner Mammographic screening: is the benefit worth the risk? Radiat Prot Dosimetry, December 1, 2005; 117(1-3): 318 - 320. [Abstract] [Full Text] [PDF]

				D. J. Brenner Radiation Risks Potentially Associated with Low-Dose CT Screening of Adult Smokers for Lung Cancer Radiology, May 1, 2004; 231(2): 440 - 445. [Abstract] [Full Text] [PDF]

				J Law and K Faulkner Two-view screening and extending the age range: the balance of benefit and risk Br. J. Radiol., November 1, 2002; 75(899): 889 - 894. [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