Congress:
EuroSafe Imaging 2017
Keywords:
Radiation safety, Digital radiography, CT, Cone beam CT, Radioprotection / Radiation dose, Action 5 - Performance indicators for radiation protection management, Action 9 - Facilitation of research in advanced topics of radiation protection, Action 2 - Clinical diagnostic reference levels (DRLs), Action 9 - Development of criteria for safe imaging procedures, Action 7 - Medical radiation protection research: MELODI, EURAMED, Action 3 - Optimisation, diagnostic reference levels, image quality, Technical aspects, Education and training, Quality assurance
Authors:
W. Stiller, R. W. Loose, C. Hoeschen, K. Katsari, L. Oleaga Zufiría, D. Tsetis, C. Granata, S. Schindera, J. Damilakis, D. Tack, L. Rainford
DOI:
10.1594/esi2017/ESI-0054
Conclusion and Recommendations
Conclusion:
The level of image quality that is deemed “appropriate” for answering a particular clinical question with high diagnostic accuracy and confidence needs to be specific for each clinical indication.
‘Appropriateness’ of the image quality of diagnostic procedures should therefore only depend on clinical indication,
while theoretically being independent of factors such as patient characteristics or the imaging hard- and software used for the examination.
For imaging modalities relying on the use of ionizing radiation,
e.g.
CT,
achieving “appropriate image quality” for a particular examination implies having minimized associated radiation exposure.
Retrospective analysis of acquired image data in view of its “appropriateness” directly after completion of each imaging examination is therefore desirable,
but requires task-based standardized “appropriate image quality” criteria derived from and driven by clinical practice.
However,
at the moment no such criteria are available,
and their development is a complex task since qualitative observer perceptions and quantitative image quality metrics have to be accounted for and need to linked to each other[2,
4,
5].
Recommendations:
For a successful and sustainable approach to the complex problem of enabling standardized measurement of the “appropriateness” of image quality of radiological examinations,
the following stepwise approach seems advisable:
-
First,
qualitative clinical criteria that unambiguously determine the image quality for a certain indication need to be identified and standardized by a group of experts representing all radiological subspecialties.
-
Then,
measurable quantitative image quality criteria linked to these qualitative clinical criteria are to be sought and validated.
Ideally,
already existing quantitative metrics can be applied or adapted for this purpose [5].
-
In the next step,
standardized sets of quantitative image quality criteria allowing for analysis of acquired image data in view of “appropriateness” for reliable and reproducible diagnosis with regard to clinical indication need to be identified.
-
Finally,
similar to model observers enabling objective assessment of image quality for clinically relevant tasks,
e.g.
as described in ICRU Report No.
54 [3,
6],
algorithms for the automated application of image quality appropriateness criteria to image data post-acquisition will need to be developed in order allow for successful use in clinical practice.
It is to be expected that the stepwise approach outlined above will require a mid- to long-term commitment for successful completion.
These efforts should be part of future research in medical radiation protection,
since reproducibly achieving appropriate image quality in clinical routine should be the ultimate goal of every optimization of imaging procedures [1}.
With regard to imaging modalities relying on the use of ionizing radiation,
e.g.
CT,
work on the “appropriateness” of image quality should also account for clinical diagnostic reference levels (DRL) currently being defined specific to particular clinical indication.
To this end,
a joint position paper together with the EuroSafe Imaging Subgroup on Clinical Diagnostic Reference Levels is foreseen to define the basis for future work on “appropriate image quality”.