THE EUROPEAN COMMISSION,

Whereas:

HAS ADOPTED THIS RECOMMENDATION:

This Recommendation concerns good practices with respect to the establishment, review and use of diagnostic reference levels for radiodiagnostic examinations and interventional radiology procedures. It invites the Member States to follow these good practices.

1.   National framework for DRLs

Member States should provide a national governance framework for the establishment, review and use of diagnostic reference levels (DRLs) and earmark appropriate resources to support and maintain such a framework.

Member States should promote a consistent national approach to the establishment, review and use of DRLs by clearly identifying the relevant stakeholders and their respective responsibilities in the process. Relevant stakeholders may vary but Member States should consider including competent authorities, health boards, professional societies, healthcare facilities, manufacturers and vendors of medical equipment. At a minimum, the establishment and review of DRLs, the collection, validation and analysis of data, and the production of guidance material for healthcare facilities on the use of DRLs should be included.

Member States should define the respective roles of the practitioner, the medical physics expert and those persons entitled to carry out practical aspects of medical radiological procedures in the collection, validation and analysis of data for establishment of DRLs as well as in the use of DRLs as a tool for optimisation in healthcare facilities.

Member States should put in place a flexible and dynamic process to establish and review DRLs. Flexibility is necessary for procedures where few data are available (e.g. interventional procedures in paediatric patients), or where data are available from a small number of healthcare facilities. A dynamic process is necessary to allow initial DRLs to be derived from these data while waiting for a more comprehensive data collection.

2.   Methodology for establishing and reviewing national DRLs

Member States should identify examinations and procedures for which national DRLs should be established, aiming for a sufficient coverage of the most common examinations and procedures or those where the collective dose to the population is significant. Where possible, given the different needs in image quality, national DRLs should be established for clinical indications. However, it is acknowledged that DRLs based on clinical indications are not suitable for all modalities, examinations and procedures, and hence DRLs based on anatomical regions could be used in these cases.

Member States should consider the nomenclature used to describe clinical indications, examinations and procedures when establishing DRLs. They should promote the use of standardised terminology and support the development of such terminology where not available. Consistent and precise description is important, especially for clinical indication-based DRLs.

For this purpose, Annexes 1 and 2 provide an overview of (i) terms used in this document; (ii) relevant dose quantities; and (iii) clinical indications, examinations and procedures for which European DRLs have been established. While differences in equipment stocks, together with varying frequency of examinations and procedures, may complicate a harmonised approach, Member States should refer to these annexes to promote consistency when establishing national DRLs.

Member States should seek to ensure that national DRLs reflect current national practices and associated patient doses, and should adopt a review frequency that fits continuous advances in imaging technologies and low-dose techniques. The relevance of the DRLs should be checked regularly by national surveys, at least every three to five years, except for dental examinations, for which a lower frequency can be adopted.

3.   Data collection and setting of the DRL values

National DRLs should be based on national dose surveys, collecting data from a representative range of healthcare facilities. To be considered representative, Member States should include both public and private facilities, and facilities of different sizes, from the entire country. To obtain national dose distributions suitable for setting DRL values, Member States should promote national DRLs based on a relevant number of typical values collected from an appropriate and representative range of healthcare facilities.

Typical values from healthcare facilities established from retrospective reviews of large numbers of patients, preferably obtained from dose management systems (DMS) or patient dose repositories, could be considered representative of the patient population even if the data include no information on the patient morphology. However, typical values could also be established for a smaller number of patients standardised by weight. In this case, data on the patient morphology are of great importance, and Member States should promote the quantification of the standard size (for example: weight, height, body mass index) for the patient population to guide this process. Care should be taken to ensure consistent data collection. Retakes and other irregularities should not be included in the definition of typical examinations used to derive typical values submitted to establish DRLs, because they are not considered part of a regular examination. Such data, however, are useful for national and local reviews. This is not the case for large amounts of data collected through DMS, as the impact of rejected images on typical dose will be negligible.

The DRL value set at the 75th percentile of the national distribution of typical values (i.e. median values) for a given DRL quantity should normally be adopted as the national DRL (10).

Member States may consider establishing an achievable dose level (11), to be set at the median of the national dose distribution. This achievable dose level provides a further opportunity for optimisation for healthcare facilities whose typical values lie under the national DRL.

Member States may consider establishing the 25th percentile as a level for identifying examinations where sufficient diagnostic image quality should be verified.

Member States may also consider analysing the national dose distribution to obtain valuable information about the level of optimisation at national level. For instance, a wide dose distribution may indicate that additional optimisation efforts would be needed.

4.   Data validation and data quality

Member States should support healthcare facilities in ensuring appropriate quality control and validation of data before submission for inclusion in national dose distributions.

It should be recognised that the use of DMS aids the establishment of DRLs and their frequent updating. However, when large amounts of data obtained from different IT systems are used, the quality and inherent uncertainty of the data should be carefully considered.

Vendors should make efforts that the relevant Digital Imaging and Communications in Medicine (DICOM) data comply with the concept of Radiation Dose Structured Report (RDSR) (12). For the implementation of these standards, integration profiles such as those from the IHE (Integrating the Healthcare Enterprise) initiative could be used, as detailed in the IHE Radiation Exposure Monitoring (REM) profile (13). Vendors and equipment manufacturers also play an important role in the harmonisation of dose quantities and units displayed by radiological equipment to improve the consistent display of DRL quantities (14), which can assist staff and foster optimisation.

Mechanisms to assess the quality of the data and the consistency of the values should be checked in relation to the examination under consideration. Member States should consider possible limitations of automated systems used for the establishment and review of DRLs. Member States should work together with manufacturers to address the need for harmonised examination nomenclature, access to structured data on clinical indications and the harmonisation of dose quantities and units displayed by radiological equipment. Access to good-quality digital data plays an important role in the generation and presentation of large amounts of data used for establishing DRLs.

The use of artificial intelligence for data verification and analysis may be considered by Member States, based on appropriate validation of the AI tools available. A coordinated effort could be supported by the Member States already advanced in this area.

5.   Image quality

Member States should support the inclusion of image quality assessments when applying the concept of DRLs. Evaluation of patient radiation dose based on DRL quantities alone, without taking image quality criteria into account, should not be seen as appropriate. Otherwise, the DRL quantity could reach a level that compromises the image quality. Healthcare facilities should always place emphasis on image quality being adequate before collecting dose data to set national DRLs. It should be recognised that the 25th percentile value could be used as a tool to identify situations where image quality needs to be evaluated. However, the impact of image post-processing tools should be considered in image quality assessments, and the complexity of the relationship between dose and image quality is further increased by introduction of AI to enhance image quality.

6.   DRLs in specific clinical settings

Member States should promote the establishment of national DRLs to sufficiently cover common examinations and procedures across all modalities, including dental cone beam computed tomography (CBCT), mammography, hybrid imaging and image-guided radiotherapy (IGRT), for both paediatric and adult patients.

Methodological differences and challenges in data collection, especially in paediatric radiology, nuclear medicine and IGRT, indicate the need for further development and guidance. In circumstances of limited availability of resources, establishment of DRLs should be prioritised for frequent and high-dose examinations/procedures and examinations/procedures of populations at higher risks (e.g. paediatric patients).

6.1.   Paediatric DRLs

Establishment of paediatric DRLs should consider the individual size differences in this population, which can make it difficult to collect sufficient data to establish DRLs. For examinations/procedures of the body, it is recommended to differentiate size by weight groups instead of age groups, while differentiation by age groups is preferred for head examinations/procedures. Since weight may not yet routinely be registered and/or accessible from IT systems of healthcare facilities, Member States may still have their paediatric DRLs based on age. Actions to shift from age- to weight-based DRLs should be encouraged by the Member States. Further recommendations on weight and age groups can be found in Annex 3. In addition, surveys to establish DRLs may focus primarily on specialised paediatric facilities with high volumes of paediatric imaging.

To overcome the general paucity of dose data in paediatric examinations/procedures within each size group, the DRLs can be presented as a function of the parameter used for patient grouping (15). The DRL-curve approach should only be applied if data from the patient dose surveys indicate a clear relationship between the DRL quantity and the patient grouping parameter, for example patient weight (DRL quantity-weight curve). This approach enables healthcare facilities to verify compliance with DRLs based on data collected from a limited number of patients (e.g. ten consecutive patients) regardless of their size, by indicating these data points on the DRL curve. If the majority of the data points are beneath the DRL curve, or if a curve fitted to the data lies below the DRL curve, then the DRL value has not been exceeded (16).

6.2.   Nuclear medicine DRLs

Observed differences in methodology for DRLs in nuclear medicine indicate a need for harmonisation across the EU. Standard methodology should be followed, meaning that DRLs should be set at the 75th percentile of the national distribution of typical values. However, it is acknowledged that Member States may continue to use existing methodologies for calculating DRL values to maintain their ability to monitor national trends. A transition may be performed by setting DRLs based on both approaches for a transition period, over one or a few data collection cycles.

To promote the use of DRLs as a tool for optimisation in nuclear medicine, Member States should establish DRLs based on measured administered activity per body weight and not on fixed nominal activity. While it is acknowledged that weight-based activity is not always in clinical use, Member States should strive to implement this practice where reasonably applicable.

For hybrid imaging, the concept of DRLs should be extended to include the CT component of the examination. The various purposes of the CT scan (attenuation correction, localisation and diagnostics) should be considered when setting the DRL value.

6.3.   Image-guided radiotherapy DRLs

Directive 2013/59/Euratom requires that all doses due to medical exposure for planning, guiding and verification purposes are optimised. However, the establishment of DRLs for image-guided radiotherapy (IGRT) has not been explicitly included in the concept of DRLs. With reference to the steady increase in the use of IGRT in treatment planning as well as for positioning and verification purposes, Member States may consider establishing DRLs (17) also for these exposure situations. Easy access to relevant dose quantities is still limited, and Member States should engage with vendors and manufacturers of equipment to facilitate efficient access to dose quantities for establishment of DRLs in IGRT.

7.   Use of DRLs as a tool for optimisation of radiation protection in healthcare facilities

Member States should encourage healthcare facilities to implement the concept of DRLs as a tool for optimisation in their clinical practice to enhance the quality and safety of medical applications using ionising radiation. The role and use of DRLs in the optimisation process should be described in the facilities’ management systems.

DRLs are used to identify and evaluate whether, in routine circumstances, the amount of ionising radiation used in medical imaging or interventional procedures at a local healthcare facility is systematically unusually high or low. This is done by comparing their typical values for given examinations/procedures against established national DRL values. Appropriate local review should be undertaken whenever typical doses exceed national DRLs, and appropriate corrective actions should be taken without undue delay. If typical values are significantly below the DRL values, an investigation should be performed to ensure that the image quality is sufficiently high for the clinical needs.

It is acknowledged that optimisation is a task that requires a multidisciplinary team and must consider all factors affecting radiation dose and image quality. In addition to the DRL methodology used, factors like equipment type, quality control, calibration, exposure parameters and procedure complexity, as well as training of the practitioner and personnel who conduct the practical aspects of the procedure should be investigated. Healthcare facilities should aim for representative typical values, which implies that these values should be updated after changes have been made to the standardised protocols and following equipment replacement or software upgrades that may affect the patient dose.

Member States should provide guidance and educational support to healthcare professionals enabling them to fully integrate the concept and use of DRLs in clinical practice.

8.   Establishment of local DRLs

Establishment of local DRLs is recommended to extend the concept of DRLs even further. Local DRLs should be established where national DRLs are missing or outdated, or where local circumstances make national DRLs not suitable for optimisation purposes. Local needs should be considered when setting local DRLs, for example when a new technology is put into clinical use. Establishment and review of local DRLs should follow the same methodology as for the establishment and review of national DRLs.