TY - JOUR
AU1 - Moores, B., Michael
AB - Abstract A review of the role and relevance of the principles of radiation protection of the patient in diagnostic radiology as specified by ICRP has been undertaken when diagnostic risks arising from an examination are taken into account. The increase in population doses arising from diagnostic radiology over the past 20 years has been due to the widespread application of higher dose CT examinations that provide significantly more clinical information. Consequently, diagnostic risks as well as radiation risks need to be considered within the patient radiation protection framework. Justification and optimisation are discussed and the limitations imposed on patient protection by employing only a radiation risk framework is highlighted. The example of radiation protection of the patient in breast screening programmes employing mammography is used to highlight the importance of defined diagnostic outcomes in any effective radiation protection strategy. INTRODUCTION The fundamental principles of radiation protection derived by ICRP are applied universally to all activities involving the use of ionising radiation under the assumption that ‘one size fits all.’ Equally, the same principles are applied to all groups of potentially exposed individuals, workers, general public and patients. More recently, the need to protect all species has also been extended to include the environment in its broadest context(1). For workers and the general public it is always possible to separate the individual from the source of radiation for all applications. However, for medical applications where sources of ionising radiation are delivered to the patient this is not the case. Consequently, mechanisms that utilise ‘protective barrier mechanisms’ to reduce doses to workers and the general public cannot be applied to patients. This, therefore, places radiation protection of patients on a very different footing compared to workers and the general public thus generating its own unique set of protective requirements. Equally the application of ionising radiation in medical imaging practices is a multi-faceted undertaking involving all age groups and anatomical regions, while employing a wide variety of different techniques. Given the ongoing growth worldwide of the use of ionising radiation in medicine and associated population doses arising from it, it seems worthwhile to review the applicability of the basic principles of radiation protection to patients in diagnostic radiology. The fundamental principles of radiation protection have evolved over the past 90 years and no doubt will continue to do so in the future. The most recent enunciation derived by ICRP is stated to be(1): The principle of justification: Any decision that alters the radiation exposure situation should do more good than harm. This means that, by introducing a new radiation source, by reducing existing exposure, or by reducing the risk of potential exposure, one should achieve sufficient individual or societal benefit to offset the detriment it causes. The principle of optimisation of protection: The likelihood of incurring exposures, the numbers of people exposed and the magnitude of their individual doses should all be kept as low as reasonably achievable (ALARA), taking into account economic and societal factors. This means that the level of protection should be the best under the prevailing circumstances, maximising the margin of benefit over harm. In order to avoid severely inequitable outcomes of this optimisation procedure, there should be restrictions on the doses or risks to individuals from a particular source (dose or risk constraints and reference levels). The principal of application of dose limits: The total dose to any individual from regulated sources in planned exposure situations other than medical exposure of patients should not exceed the appropriate limits recommended by the Commission. The concepts of dose constraint and reference level are used in conjunction with the optimisation of protection to restrict individual doses...Diagnostic reference levels are already being used in medical diagnosis (i.e. planned exposure situations) to indicate whether, in routine conditions, the levels of patient dose or administered activity from a specified imaging procedure are unusually high or low for that procedure. All of these principles continue to be driven solely by radiation risk and they are aimed at minimising or containing risks that arise from the use of ionising radiation. Indeed ICRP was established specifically to deal with such risks when early pioneers, who helped establish the beneficial use of ionising radiation, were suffering the consequences of unprotected overexposure. More recently, following an indicated up to 20-fold increase in population doses arising from medical practices in developed countries over the past 50 years(2), reduction or elimination of unnecessary exposure to the patient has taken a pre-eminent position in radiation protection philosophy(3). Indeed the growth of high dose CT examinations that provide high levels of diagnostic information has played a major role in stimulating recent radiation protection activities. In order to reinforce the overriding need to ensure that the basic principles of radiation protection play a prominent role in medical diagnostic practice, the concepts of ethics have, recently been discussed within the framework of radiation protection(4). In support of this concept ICRP has established Task Group 94 on the Ethics of Radiological Protection(5). The inference from this initiative is that the promotion of ethical basis of radiation protection can make a specific contribution to the implementation of the basic principles of radiation protection in diagnostic radiology. Such an approach does not appear to have been discussed in relation to other applications of ionising radiation. Four ethical principles have been proposed that can help underpin the effective management of radiation risks in diagnostic radiology(4): Respect for autonomy (of the individual) Non-maleficence (do not harm) Beneficence (do good) Justice (be fair) Together with three related values: Human dignity Prudence Honesty While the effective management of radiation risks is an important and integral part of radiation protection of the patient in diagnostic radiology it is by no means the only factor of importance. It has recently been shown that diagnostic risks are a significant and often overlooked aspect of radiation protection of the patient(6). By applying the cost–benefit analysis proposed by ICRP in 1973 to the first 2 levels of the National Council on Radiation Protection and Measurements (NCRP) efficacy model(7) it has been shown that the cost-risk-benefits associated with the diagnostic radiological process involving a patient population N, may be written in the form: K(E)=[NPT/P(E)+NNT/N(E)]–NPT/P(E)[(1−sens)/sens]+NNT/N(E)[(1–spec)/spec]}–NRX(E) (1)K(E) is the total costs of providing the radiological service, NPT/P(E) and NNT/N(E) are the true-positive and true-negative diagnostic outcomes, respectively, at the exposure level E. Where T/P(E) and T/N(E) are the conditional true-positive and true-negative probabilities, respectively, of a particular diagnostic outcome. NP and NN are the numbers of diseased and healthy patients, respectively, with prevalence nr = NP/(NP + NN). The terms sens and spec represent the sensitivity and specificity of the diagnostic process, respectively, so that the term in the second bracket represents the presence of both false-positive and false-negative outcomes or ‘diagnostic risk’. In Equation (1) the final term NRX(E) represents the radiation risk expressed as the number of cancers induced in a population of N patients where RX(E) is the product of the risk of cancer induction per Sv (5.7 × 10−2 Sv–1) multiplied by the average dose per examination within the group of N patients. There are three main causes of inaccurate or uncertain diagnostic outcomes: The pathology presented (size, extent, relevance to symptoms, etc.) The performance of the imaging system—image quality and patient dose The performance of the observer in forming a diagnosis At present radiation protection of the patient in diagnostic radiology concentrates almost exclusively on the second of these factors; imaging system performance (see section on Optimisation), however, the other two also have a major impact on diagnostic efficacy. In view of these recent developments it seems reasonable to review the application of the fundamental principles of radiation protection as formulated by ICRP to protection of the patient in diagnostic radiology. The purpose of this paper is to undertake such a review when both radiation and diagnostic, risks are taken into account. Equally, a discussion of the ethical basis of radiation protection when both types of risks are considered is also included. DISCUSSION Justification The principle of justification requires there to be sufficient individual or societal benefit to offset the detriment caused by a change in exposure conditions arising from the introduction of new applications/techniques or modifications of existing methods. For medical diagnostic applications a radiation risk driven radiation protection framework is applied with quantification of patient doses and their possible reduction as its main aims. This means that the benefits are generally assessed in relation to an assumed adequate true-positive [NPT/P(E)] and true-negative [NNT/N(E)] diagnostic outcomes as indicated by the first square bracket term in Equation (1) in comparison to the numbers of cancers induced, as represented by the final term NRX(E). This type of approach ignores any detriment arising from the diagnostic process. In medical applications of ionising radiation the clinical detriment arising from both therapeutic and diagnostic applications cannot be ignored. Poorly planned radiotherapy treatments and/or the irradiation of healthy tissues create detrimental outcomes, which are readily accepted and form an integral part of patient protection strategies in radiotherapy. Equally in diagnostic applications the effect of inaccurate diagnoses must also be considered within the overall patient protection framework. Merely comparing the numbers of correct diagnostic outcomes with the numbers of cancers induced is an inadequate risk-benefit framework. If a diagnostic imaging process provided a 50% T/P(E) and equal T/N(E) detection probabilities it would always indicate a numerically favourable benefit-risk outcome. Thus if 1 million patients are examined with 30% prevalence and 50% true detection outcomes (both T/P(E) and T/N(E)) then 150 000 true-positive and 350 000 true-negative outcomes (500 000 in total) would be predicted. Assuming that the imaging process employed an average dose of 1 mSv per patient, then assuming a radiation risk estimate of 5.7% per Sv(1), 57 cancers would be predicted corresponding to a benefit-risk ratio of roughly 8800. Such an approach would be meaningless since for 50% true detection probabilities [T/P(E) and T/N(E)] there would be an equal number of false detection outcomes due to false-positive [F/P(E)] and false-negative [F/N(E)] probabilities. The cost–benefit predicted by Equation (1) would in fact be zero since diagnostic outcomes would be completely uncertain. Therefore, merely assigning diagnostic outcomes on a random basis would be equally effective so that a diagnostic test employing ionising radiation would be superfluous. Consequently benefits (net benefit) arising from diagnostic radiological examinations must also take into account false or incorrect diagnoses expressed by the sensitivity and specificity of an examination since these degrade the diagnostic strength of any outcome and hence clinical benefit. Therefore, unproductive patient doses, which lead to false diagnostic outcomes and, therefore, do not provide clinical benefit, must be included in any patient protection strategy. If we assume that 1 million patients undergo an X-ray examination with a diagnostic accuracy [T/P(E), T/N(E)] of 80% employing an average dose of 1mSv, then irrespective of the disease prevalence 800 000 correct diagnoses would result with 200 000 incorrect outcomes. The radiation risk would indicate that 57 cancers may be induced(1). This equates to a diagnostic benefit to radiation risk ratio 14 000 to 1 at a diagnostic detriment level of 20%. If the 200 000 patients who received an incorrect diagnosis had a further diagnostic X-ray examination with a 95% accuracy and employed an average dose of 10 mSv then a further 190 000 patients would be correctly diagnosed, 10 000 incorrectly and 114 cancers may be induced. Thus overall (both examinations) some 990 000 patients would receive a correct diagnosis, 10 000 an incorrect with 161 cancers indicated overall, ignoring single cancers arising from a double exposure. This corresponds to an overall diagnostic benefit to radiation risk ratio of 6000 to 1 for both examinations together at a diagnostic detriment (false outcome) level of 1%. The diagnostic benefit to radiation risk ratio for the second examination treated separately would be 1666 to 1 at a diagnostic detriment level of 5%.However, the 5% diagnostic detriment corresponding to 10 000 patients would still greatly exceed the radiation detriment represented by the possible 114 cancers. When diagnostic outcomes result from more than a single test, including ones that do not employ ionising radiation, how should the overall risk and benefits be described? Should they be defined for each individual test or also in terms of an overall ‘average’ diagnostic yield over a patient population that present with specific symptoms and follow a prescribed diagnostic pathway? This latter approach would then take cognizance of the fact that an indeterminate outcome from the first test prompted further investigation. However, patients referred for further investigation may also be subject to induced adverse effects such as a degree of emotional trauma, corresponding to intangible costs. Very little effort has been expended on attempts to quantify ‘emotional/psychological’ detriments but they are relevant to patient protection. Justification is dependent upon a full appraisal of diagnostic performance and must also take cognizance of both true and false diagnostic outcomes. Thus the overall strategy within which diagnostic radiology is employed, is a relevant consideration in any appraisal of justification. It is perhaps not surprising that high-dose/high-information content examinations have universally become a first useful port of call in modern healthcare systems. ICRP has indicated there are three levels of justification of a radiological practice in medicine: A first general level that accepts the proper use of radiation in medicine as doing more good than harm to society. A second level where a specified procedure with a specified objective is defined and justified. Aim is to judge whether the procedure will improve diagnosis or treatment, or will provide necessary information about the exposed individuals. A third level where the application of the procedure to an individual patient should be justified. Application should be judged to do more good than harm to the individual patient. Patients who receive an incorrect diagnosis, either NNF/P(E) or NPF/N(E) false-positive and false-negative respectively, cannot be deemed to fulfil the level 3 criteria. Consequently, the probability of good or harm for this group is of paramount importance and this will be discussed under ethical considerations. Since 100% correct diagnosis will never be possible, the statistical accuracy for a particular examination corresponding to a specific set of symptoms is required to make any meaningful judgement for justification at level 3. Ideally this would need to be tested over many different groups of patients with specific symptoms that undergo an examination at sufficiently many different locations in order to fully sample all the potential variations in the patient population, imaging process and reporting accuracy, which can affect diagnostic outcomes. Equally, this approach is required in order to ensure satisfaction at the second justification level. Without a clear indication of the acceptable levels of diagnostic accuracy that would be deemed reasonable, the applicability or otherwise of these justification criteria cannot be tested. This type of approach underpins attempts to develop evidence based practice, within which radiation protection should operate. The integration of diagnostic risks into an overall framework for radiation protection of the patient in diagnostic radiology does not in any way diminish the clinical impact of medical imaging on healthcare over the past 40 years. Significant improvements in diagnosis have and continue to be integrated into patient care through ongoing technological developments. Indeed comparing the performance of the three main modalities 40 years ago (film screen radiography, fluoroscopy and conventional tomography) with present day capability would be like comparing chalk and cheese! Modern developments include CT, digital fluoroscopy, dual energy and subtraction techniques, tomosynthesis, PET/CT, etc. as well as diagnostic CAD support tools for the observer. Awareness and discussions on false or unclear diagnostic outcomes and how to deal with them continues to help drive improvements in all aspects of breast cancer screening programmes. Optimisation According to ICRP(8), The optimisation of radiological protection for patients in medicine is usually applied at two levels: (1) the design, appropriate selection and construction of equipment and installations; (2) the day-to-day methods of working (i.e. the working procedures). Thus optimisation as applied to medical radiation protection is largely centred on the evaluation of imaging performance (image quality and dose) applicable to Level 1—technical efficacy of the NCRP efficacy model(7). The aim then is to ‘optimise’ image quality at a particular level of dose, under the assumption that this will translate directly into optimised diagnostic outcomes. Because the ICRP principles of radiation protection are applied equally to all uses of ionising radiations, optimisation expressed in terms of diagnostic outcomes would represent an anomaly. However, the need to include diagnostic outcomes in radiation protection of the patient results directly from the use of ionising radiation as the vehicle for deriving clinical information. The administration/utilisation of ionising radiation and the information gained cannot be separated. In reality, when viewed from a diagnostic (clinical) risk perspective there would appear to be a number of potentially desirable and valid optimisation strategies: Lower the patient dose while maintaining the same level of diagnostic performance (reduce radiation risk for the same level of diagnostic risk). This may result from an improvement in technique or technology. Increase the diagnostic performance for the same level of patient dose (reduce diagnostic risk for the same level of radiation risk). This may result in an improvement in reporting accuracy through training, computer aided diagnosis or the use of contrast agents. Increase the diagnostic performance proportionately to an increase in the patient dose (reducing diagnostic risk in proportion to increasing radiation risk). This may result from an improved delineation of relevant pathology either through technical improvements such as the use of dual energy or subtraction techniques. Increase the diagnostic performance while lowering the patient dose (lower diagnostic risk while lowering radiation risk). This could be due to a combination of technical developments/new methods and associated improvements in reporting—classical ALARA optimisation. This variety of optimisation strategies is ignored within the universally applied concept of optimisation with its bias towards ALARA and dose limitation. When diagnostic outcomes are considered the limitations imposed by ALARA become quite obvious. Paragraph (217) of ICRP 2007(1) states: The optimisation of protection is a forward – looking iterative process aimed at preventing or reducing future exposures... In medical applications, developments do not always follow an iterative process but may involve a significant increase in patient doses when new techniques are developed as seen with developments in CT. However following such developments, ALARA based strategies continue to drive radiation protection in diagnostic radiology without any balancing influence from the marked improvements in diagnostic quality. Paragraph (219) states: Optimisation of protection is not minimisation of dose. Optimised protection is the result of an evaluation, which carefully balances the detriment from the exposure and the resources available for the protection of individuals. Thus the best option is not necessarily the one with the lower dose. In medical practice a major source of detriment may occur if an examination is not performed or is poorly performed so that incorrect diagnostic outcomes occur. The concept of optimisation of protection for patients in medical exposures is expressed in paragraph (70) of ICRP publication 105:(8) The optimisation of radiological protection means keeping the doses ‘as low as reasonably achievable, economic and social factors being taken into account,’ and is best described as management of the radiation dose to the patient to be commensurate with the medical purpose. When viewed from the perspective of patient care, optimisation should primarily be aimed at maximising the diagnostic yield from an examination and then exploring the patient dose options available. This would, of course, need to consider the relative diagnostic benefit to radiation risk ratios that apply. For most diagnostic radiological examinations (dose range 1–10 mSv) the diagnostic benefits exceed the radiation risks by a factor of 1000 to 1 or more and this fact helps establish the justification of practices (see section on Justification). However, diagnostic risks expressed in terms of numbers of incorrect diagnoses (false-positive and false-negative) also significantly exceed the radiation risks(6). Doubling or tripling the average patient dose to achieve a 1% increase in diagnostic accuracy (i.e. lowering diagnostic risks) could lead to an increased diagnostic benefit that may still greatly exceed any potential increase in radiation detriment. Thus for a group of 1 million patients, a 1% increase in diagnostic accuracy would beneficially affect the outcomes for 10 000 patients. Whereas an increase in average doses from 1 to 3 mSv may lead to roughly 100 extra cancers induced. Any cancers induced may not present during a patient's lifetime (hence the patient's age would be an important consideration in any risk–benefit) whereas the improved diagnostic results would be of immediate and direct benefit. Thus any increase in diagnostic accuracy may fully justify a significant increase in patient dose. The risks associated with false or incorrect diagnosis merit ongoing consideration within a developing radiation protection framework for a fuller understanding of optimisation and justification. Radiation protection needs to be actively involved in developing strategies and mechanisms for assessing the optimisation of diagnostic outcomes. For patients who undergo medical diagnosis or treatment specific dose limits do not apply. However, diagnostic reference levels (DRLs) are now the accepted mechanism by which the concept of dose constraints may be applied to patients in diagnostic radiology and they are promoted as a patient dose management tool(8). From a clinical perspective, a fundamental assumption underpinning the use of DRLs as a tool for optimisation assumes that the net diagnostic benefit for a specific type of examination is a universally consistent outcome applicable to the three ICRP levels of justification. The only relevant optimisation variable necessary is then the patient dose. The detriment associated with the diagnostic risks inherent in any false or unclear diagnostic outcomes is deemed to fall outside the scope of radiation protection. In order to support the application of DRLs, ICRP has produced draft guidance on DRLs in Medical Imaging(9). Section 1.4 of this draft guidance discusses the effectiveness of DRLs quoting the example of the US Breast Exposure: Nationwide Trends (BENT) mammographic quality assurance (QA) programme as an early demonstration of the effectiveness of the DRL approach(10). Surprisingly, mammographic breast screening programmes (BSP) are an ideal example of the effective application of diagnostic outcomes within the overall framework of patient protection in order to define the acceptability and optimisation of practices. The UK NHS Breast Screening Programme (NHSBSP) Quality Assurance Guidelines for Breast Cancer Screening Radiology provides quality standards for this activity(11). Table 1 presents the core radiological quality standards for the UK BSP expressed in terms of specific clinical outcomes required for justification. This includes the number of cancers detected as well as the number of small invasive cancers detected. Table 1. Core radiological standards for UK BSP. Objective Criteria Minimum standard Achieveable standard 1. To maximise the number of cancers detected The rate of invasive cancers in eligible women invited and screened Prevalent screen ≥3.6 per 1000 Prevalent screen ≥5.1 per 1000 Incident screen ≥5.7 per 1000 Incident screen ≥ 4.1 per 1000 2. To maximise the number of small invasive cancers detected The rate of invasive cancers <15 mm in diameter detected in eligible women invited and screened Prevalent screen ≥2.0 per 1000 Prevalent screen ≥2.8 per 1000 Incident screen ≥3.1 per 1000 Incident screen ≥2.3 per 1000 Objective Criteria Minimum standard Achieveable standard 1. To maximise the number of cancers detected The rate of invasive cancers in eligible women invited and screened Prevalent screen ≥3.6 per 1000 Prevalent screen ≥5.1 per 1000 Incident screen ≥5.7 per 1000 Incident screen ≥ 4.1 per 1000 2. To maximise the number of small invasive cancers detected The rate of invasive cancers <15 mm in diameter detected in eligible women invited and screened Prevalent screen ≥2.0 per 1000 Prevalent screen ≥2.8 per 1000 Incident screen ≥3.1 per 1000 Incident screen ≥2.3 per 1000 Table 1. Core radiological standards for UK BSP. Objective Criteria Minimum standard Achieveable standard 1. To maximise the number of cancers detected The rate of invasive cancers in eligible women invited and screened Prevalent screen ≥3.6 per 1000 Prevalent screen ≥5.1 per 1000 Incident screen ≥5.7 per 1000 Incident screen ≥ 4.1 per 1000 2. To maximise the number of small invasive cancers detected The rate of invasive cancers <15 mm in diameter detected in eligible women invited and screened Prevalent screen ≥2.0 per 1000 Prevalent screen ≥2.8 per 1000 Incident screen ≥3.1 per 1000 Incident screen ≥2.3 per 1000 Objective Criteria Minimum standard Achieveable standard 1. To maximise the number of cancers detected The rate of invasive cancers in eligible women invited and screened Prevalent screen ≥3.6 per 1000 Prevalent screen ≥5.1 per 1000 Incident screen ≥5.7 per 1000 Incident screen ≥ 4.1 per 1000 2. To maximise the number of small invasive cancers detected The rate of invasive cancers <15 mm in diameter detected in eligible women invited and screened Prevalent screen ≥2.0 per 1000 Prevalent screen ≥2.8 per 1000 Incident screen ≥3.1 per 1000 Incident screen ≥2.3 per 1000 Similarly the general radiological quality standards are shown in Table 2. This includes the number of women screened who are referred for further assessment due to inconclusive findings. All of these standards define the framework for optimised radiation protection of patients who undergo X-ray examinations as part of breast cancer screening. The performance of individual screening programmes can easily be compared against these standards. Indeed it would be meaningless to try and optimise mammographic screening programmes merely from a patient dose (ALARA) perspective. Table 2. General radiological standards for the UK BSP. Objective Criteria Minimum standard Achievable standard 1. To minimise the number of women screened who are referred for further tests (a) The percentage of women who are referred for assessment (a) Prevalent screen <10% (a) Prevalent screen <7% Incident screen <5% (b) The percentage of women screened who are placed on early recall (b) <0.25% (b) <0.12% 2. To ensure that the majority of cancers, both palpable and impalpable receive a non-operative tissue diagnosis of cancer (a) The percentage of women who have a non-operative diagnosis of invasive cancer by needle histology after a maximum of two attempts (a) ≥90% (a) ≥95% 3. To minimise the number of unnecessary operative procedures The rate of benign biopsies Prevalent screen <1.5 per 1000 Prevalent screen <1.0 per 1000 Incident screen <1.0 per 1000 Incident screen <0.75 per 1000 Objective Criteria Minimum standard Achievable standard 1. To minimise the number of women screened who are referred for further tests (a) The percentage of women who are referred for assessment (a) Prevalent screen <10% (a) Prevalent screen <7% Incident screen <5% (b) The percentage of women screened who are placed on early recall (b) <0.25% (b) <0.12% 2. To ensure that the majority of cancers, both palpable and impalpable receive a non-operative tissue diagnosis of cancer (a) The percentage of women who have a non-operative diagnosis of invasive cancer by needle histology after a maximum of two attempts (a) ≥90% (a) ≥95% 3. To minimise the number of unnecessary operative procedures The rate of benign biopsies Prevalent screen <1.5 per 1000 Prevalent screen <1.0 per 1000 Incident screen <1.0 per 1000 Incident screen <0.75 per 1000 Table 2. General radiological standards for the UK BSP. Objective Criteria Minimum standard Achievable standard 1. To minimise the number of women screened who are referred for further tests (a) The percentage of women who are referred for assessment (a) Prevalent screen <10% (a) Prevalent screen <7% Incident screen <5% (b) The percentage of women screened who are placed on early recall (b) <0.25% (b) <0.12% 2. To ensure that the majority of cancers, both palpable and impalpable receive a non-operative tissue diagnosis of cancer (a) The percentage of women who have a non-operative diagnosis of invasive cancer by needle histology after a maximum of two attempts (a) ≥90% (a) ≥95% 3. To minimise the number of unnecessary operative procedures The rate of benign biopsies Prevalent screen <1.5 per 1000 Prevalent screen <1.0 per 1000 Incident screen <1.0 per 1000 Incident screen <0.75 per 1000 Objective Criteria Minimum standard Achievable standard 1. To minimise the number of women screened who are referred for further tests (a) The percentage of women who are referred for assessment (a) Prevalent screen <10% (a) Prevalent screen <7% Incident screen <5% (b) The percentage of women screened who are placed on early recall (b) <0.25% (b) <0.12% 2. To ensure that the majority of cancers, both palpable and impalpable receive a non-operative tissue diagnosis of cancer (a) The percentage of women who have a non-operative diagnosis of invasive cancer by needle histology after a maximum of two attempts (a) ≥90% (a) ≥95% 3. To minimise the number of unnecessary operative procedures The rate of benign biopsies Prevalent screen <1.5 per 1000 Prevalent screen <1.0 per 1000 Incident screen <1.0 per 1000 Incident screen <0.75 per 1000 Because NHSBSP provides a comprehensive framework for radiation protection of the patient that includes clinical outcomes, the introduction of new techniques such as tomosynthesis or CT can be properly evaluated in respect of both justification and optimisation. Equally the performance of different screening programmes may be compared in terms of clinical effectiveness. For example a comparison of screening programmes in the US and Denmark has highlighted the much higher recall rate (21 recalls to detect 1 cancer—lower specificity 83.2%) in the US compared to a recall rate of 3 to detect 1 cancer in Denmark (specificity of 96.6%)(12). High recall rates represent higher costs as well as emotional impact and associated cost detriment, which lower the efficacy, optimisation and justification of practices all integral to radiation protection of the patient. The justification for an examination with higher sensitivity and specificity must be deemed to be greater. If such a rigorous framework for protection of the patient is employed within a breast cancer screening programme, then why has not the same approach been adopted more generally in diagnostic radiology where much higher dose techniques are employed? The ability to grade a diagnostic outcome from definite abnormality through to definitely normal for specific symptoms for particular referral criteria and to link these to subsequent patient management outcomes is now eminently possible. Employing ‘Big Data’ analytical techniques could then provide an indication of the diagnostic outcomes for specific types of examinations locally, regionally and nationally. Such quantitative outcomes may then be used to assess expected referrals vs diagnostic outcome criteria, which could be used to indicate investigations when a high proportion of unhelpful examinations were being requested. Ethical basis for radiation protection in diagnostic radiology It has recently been stated that(4) ‘…for the most part scholarship in medical ethics does not attend to the problems in radiation protection. Indeed there appears to be an unwritten assumption that matters relating to radiation are dealt with in a separate system and medical/general ethicists have not engaged with it.....The counterpoint is that the ethical issues in radiation protection have low recognition in the medical field.’ The primary reason for the separation of radiation protection from the main medical ethical principles is largely due to the fact that the principles of radiation protection as formulated by ICRP completely ignore the clinical/diagnostic outcomes from the use of ionising radiation. These are the elements that justify medical applications and through which medical practitioners demonstrate their professional (and ethical) competency. Consequently this cultural dichotomy will continue while only radiation related factors are considered to be relevant for patient protection with ALARA its driving force. In the case of patient protection in BSPs the divergence between ethical considerations in radiation protection and general medical ethics do not exist. Protection of the patient operates within a clinical framework in which diagnostic outcomes are fully integrated and provide a quantifiable basis for justification. Patient dose and QA considerations then act as underlying components that help to support the optimisation of practices. Attempts to lower patient dose would be balanced against any impact on diagnostic accuracy and any new techniques can be assessed against clearly defined clinical criteria. In the case of general diagnostic radiology a pragmatic ethical basis for radiation protection has been proposed based upon the primacy of radiation risks. This approach refers to the four basic principles outlined in Introduction together with three related values(4). In some respects ethical considerations represent supporting ethical criteria for justification whereby radiation protection experts can assess the degree of compliance of an examination against the four basic principles and two additional values of ethics. A number of scenarios have been presented in which aspects of clinical governance are assessed against the ethical principles and values in order to assess the suitability of the radiological process from an ethical viewpoint. Radiation risks associated with an examination are the driving force behind this proposed pragmatic ethical approach. In turn these considerations have arisen because of the increase in population doses arising from the growing utilisation of high dose examinations that provide higher levels of diagnostic information. Unfortunately, the assessment of patient dose and its reduction seems to have taken precedence over the corresponding assessment of diagnostic efficacy. For example in a good practice guide for clinical radiologists published by the Royal College of Radiologists in the UK it was noted that(13): ‘There are still few nationally agreed figures for diagnostic accuracy for any imaging investigation.’ Nonetheless, it has been shown that the most common general cause of medico-legal claims against radiologists is error in diagnosis(14, 15). Consequently by concentrating on dose reduction strategies alone, even for examinations that are justified, there is a real risk that we might unknowingly also reduce diagnostic efficacy. For an examination that involves a dose of 10 mSv the indicated probability of inducing a cancer would be 1 in 2000(1). If we assume that the examination provides a diagnostic accuracy of 95% , either true positive or true negative, the diagnostic risk would be 5% or 1 in 20, approximately 100X greater than that due to radiation. This is not to say that 95% diagnostic accuracy is not acceptable, but merely to indicate that more scientific effort in quantification of diagnostic outcomes and methods/tools for their assessment would seem to be desirable from a relative risk perspective. In the case of CT examinations it is probable that the increase in patient dose for each examination may well be balanced by a proportionate increase in diagnostic efficacy. The problem associated with growth in population doses would then be due to a growth in unjustifiable high dose examinations, which provide unhelpful diagnoses. If the symptomatic diagnostic yield (examinations that indicated an abnormality) per patient referral was assessed routinely for each practice, then higher than normal referrals would be self evident and investigated. Concentrating on lowering doses to each individual patient may help to reduce population doses, however, it does not tackle the fundamental problem of over-utilisation. A simple three level diagnostic outcome grading system (probably abnormal, indeterminate, probably normal) would provide an initial useful measure of diagnostic performance as a function of referral patterns and would be ethically desirable. Such an assessment could be included in any radiologist's report. Alternatively it is now feasible to apply ‘google-type’ search engines to existing reports in order to assess the frequency of characteristic words that could be assigned to one of the suggested simple three tier class of diagnostic outcomes indicated above(16). Such outcomes could then be analysed in relation to the associated referral criteria. The optimisation of diagnostic outcomes is of primary importance and techniques that can maximise them are ethically desirable. Diagnostic tests are of value only if they result in improved outcomes for patients. In fact optimisation of diagnostic outcomes is an overarching ethical requirement that sits above any radiation risk related framework. It is possible to have an examination that ticks all the ethical boxes from a radiation risk related perspective but if the probability of a satisfactory diagnostic outcome does not meet an acceptable level then the examination may not be justified. To ignore the diagnostic risks inherent in patient protection seems illogical. Particularly when diagnostic outcomes have for some time now been successfully integrated into the fabric of radiation protection of the patient in breast cancer screening programmes. In fact, if diagnostic outcomes were an integral part of medical radiation protection, quantitative ethics would automatically apply. Any high proportion of unnecessary or unhelpful examinations (high negative or indeterminate yields with symptoms not clearly defined) associated with unnecessary exposure of the patients could be assessed. All of the discussions that have been presented to highlight ethical considerations in relation to radiation risks apply unequivocally to diagnostic risks(4). From a medico-legal perspective it may seem reasonable to advocate discussing radiation risk issues with patients, when it would be very difficult to link a medical exposure to any adverse effect decades later. However, adopting the same principle to diagnostic risks would mean discussing the significantly greater probability of a false diagnosis with patients, which could have significant legal consequences in the actual event of an incorrect diagnosis. This creates a moral dilemma that cannot be ignored unless a completely open and transparent framework for informed consent is accepted by both patients and professional radiological communities(14,17). Consequently, such a framework would need to include accepted and verifiable levels of diagnostic accuracy for routine examinations. This aspect falls naturally within the field of radiation protection as amply demonstrated in the field of breast screening. In BSPs, whether or not an acceptable dose was employed becomes a quantitative adjunct to the fulfilment of defined primary clinical objectives that are fundamental to clinical radiation protection of the patient. Quantification of desirable diagnostic outcomes for specific referral criteria (symptoms, age group, lifestyle, etc.) and the necessary tools for routine application are needed in diagnostic radiology. As indicated previously the capability now exists for this approach given the role of IT in healthcare and the use of ‘Big Data’ analytics and the overall costs and importance of radiological services worldwide should be sufficient driving forces. This would enable the principles of radiation protection in medical practice to be fully integrated into clinical practice as effective quality standards. For progress to be made, it requires an acceptance by the radiation protection community of the fact that a single set of ALARA driven radiation protection principles, cannot be applied equally to all applications of ionising radiation, especially in respect of diagnostic radiology. SUMMARY AND CONCLUSIONS For the past 50 years medical applications of ionising radiation has been recognised as an important consideration in the overall framework of radiation protection of mankind. Although patients are exposed directly as part of their treatment or diagnosis, ICRP has always attempted to establish a single framework for radiation protection that covers all applications and all exposed groups, workers, general public and patients with a primary aim of keeping doses as low as reasonably achievable. Unfortunately, in diagnostic radiology the quality of clinical information gained is often dose dependent and without any consideration of the diagnostic yield from an examination neither justification nor optimisation of practices is verifiable in respect of the their efficacy. The situation that exists in general diagnostic radiology is in contra-distinction to that which has been applied in BSPs employing mammography. For this specific practice clearly defined clinical outcomes have been established in order to underpin radiation protection of the patient through verifiable justification and optimisation criteria. Such an approach is now long overdue in the field of general radiology. Simple measures of diagnostic outcomes arising from an examination for specific referral criteria are quite feasible and the necessary technology for managing this type of information exists. As one of the most prevalent and effective diagnostic techniques the ability to quantify diagnostic outcomes is a basic requirement of the fundamental principles of radiation protection of the patient in diagnostic radiology. REFERENCES 1 International Commission on Radiological Protection The 2007 Recommendations of the International Commission on Radiological Protection. ICRP Report 103 . Ann. ICRP 37 , 2 – 4 ( 2007 ). WorldCat 2 Moores , B. M. and Regulla , D. A review of the scientific basis for radiation protection of the patient in diagnostic radiology . Radiat. Prot. Dosim. 147 , 22 – 29 ( 2011 ). Google Scholar Crossref Search ADS WorldCat 3 Council of the European Union . Council Directive laying down basis safety standards for protection against the dangers arising from exposure to ionising radiation. Council Directive 2013/59/EURATOM ( 2013 ). 4 Malone , J. and Zoler , F. A pragmatic ethical basis for radiation protection in diagnostic radiology . Br. J. Radiol. 89 ( 1059 ), 20150713 ( 2016 ). Google Scholar Crossref Search ADS PubMed WorldCat 5 International Commission on Radiological Protection . Code of ethics of radiological protection. ICRP Ref. 4842-2907-2666 ( 2014 ). 6 Moores , B. M. Cost-risk-benefit analysis in diagnostic radiology: a theoretical and economic basis for radiation protection of the patient. Proceedings of the Fourth Malmo Conference . Radiat. Prot. Dosim. 169 , 2 – 10 ( 2016 ). Google Scholar Crossref Search ADS WorldCat 7 National Council on Radiation Protection and Measurements . An introduction to efficacy in diagnostic radiology and nuclear medicine (justification of medical radiation exposure). NCRP Commentary No. 13, NCRP ( 1995 ). 8 International Commission on Radiological Protection Radiological protection in medicine. In: ICRP Publication 105 (Volume 37, Page 6) ( 2007 ). Google Preview WorldCat COPAC 9 International Commission on Radiological Protection . ICRP 201x. Diagnostic Reference Levels in Medical Imaging. Consultaion Document. ICRP, Publication 1XX Ann. ( 2016 ). 10 Jensen , J.E. and Butler , P.F . Breast exposure: nationwide trends; a mammographic quality assurance program – results to date . Radiol. Technol 50 , 251 – 257 ( 1978 ). Google Scholar PubMed WorldCat 11 NHSBSP . Quality Assurance Guidelines for Breast Cancer Screening Radiology. Second edn. NHS Cancer Screening Programmes ( 2011 ). 12 Jacobsen , K. K. , O'Meara , E. S. , Key , D. , Buist , D. , Kerlikowske , K. , Veiborg , I. , Sprague , B. L. and von Euler- Chelpin , M. Comparing sensitivity and specificity of screening mammography in the United States and Denmark . Int. J. Cancer 137 , 198 – 207 ( 2015 ). WorldCat 13 Royal College of Radiologists . Good Practice Guide for Clinical Radiologists . The Royal College of Radiologists , 38 Portland Place, London W1B 1JQ , ( 2012 ). 14 Halpin , S. F. S. Medico-legal claims against English radiologists: 1995–2006 . Br. J. Radiol. 82 , 982 – 988 ( 2009 ). Google Scholar Crossref Search ADS PubMed WorldCat 15 Whang , J. S. , Baker , S. R. , Patel , R. , Luk , L . and Castro , A. The causes of medical malpractice suits against radiologists in the United States . Radiology 266 , 548 – 554 ( 2013 ). Google Scholar Crossref Search ADS PubMed WorldCat 16 Shi , L. , Ling , T. and Zhang , J. Semantic information extracting system for classification of radiological reports in radiology information system (RIS). In: Proceedings of SPIE 9789, Medical Imaging 2016: PACS and Imaging Informatics: Next Generation and Innovations 97890 T March 25 ( 2016 ). 17 European Society of Radiology . Risk management in radiology in Europe, Section 2.2.2; Consent. ESR/EAR Office, Neutograsse 9/2a/AT-1010 Vienna, Austria ( 2004 ). © The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com
TI - A REVIEW OF THE FUNDAMENTAL PRINCIPLES OF RADIATION PROTECTION WHEN APPLIED TO THE PATIENT IN DIAGNOSTIC RADIOLOGY
JF - Radiation Protection Dosimetry
DO - 10.1093/rpd/ncw259
DA - 2017-06-01
UR - https://www.deepdyve.com/lp/oxford-university-press/a-review-of-the-fundamental-principles-of-radiation-protection-when-n4Ut4ndG8Z
SP - 1
VL - 175
IS - 1
DP - DeepDyve
ER -