Predictivity and optimisation in medical rediation protection

• Research unit: EU RFP
• Project number: 95.0117

Key words (English)

Optimisiation; radiography; image quality

Alternative project number (English)

EU project number: FI4P-CT95-0005

Research programs (English)

EU-programme: 4. Frame Research Programme - 5.2 Nuclear fission safety

Short description (English)

See abstract

Partners and International Organizations (English)

M. Moores (UK), S. Mattson (S). W. Panzer (D), D. Hoenschen (D), D. Dance (UK), G.A. Carlsson (S)

Abstract (English)

In order to get an optimum between dose and image quality, a global characterisation of the radiological procedure, including all the steps from the x-ray generator to the visualisation of the film, is needed. The aim of this project was to derive quantitative methods to parameterise the radiological information content provided by two x-ray examinations, i.e. chest and lumbar spine. The final goal was to demonstrate that predictivity was possible when dealing with radiographic examinations in a framework of patient dose and image quality optimisation.

To evaluate at once the whole radiological chain, several methods can be employed. The most common way is to use clinical images and to rely on radiologists judgement. This solution is however quite inappropriate for optimisation since it includes the assessment of the radiological information content mixed with the radiologists training and personal preferences. This methodology leads to inconsistent and non reproducible results. To solve this problem, test-objects have been developed in order to eliminate patient variability and to simplify the evaluation by incorporating simple targets. Their main drawback is that they assess properties which are generally far from the clinical applications. Our group developed an image quality index in the field of mammography which represents the size of the smallest microcalcification detectable on a film. This index is completely objective since it is based on the statistical decision theory. Moreover, it has the advantage of being directly linked to the clinical information content of mammograms

The first part of the work has been dedicated to the complete characterisation of conventional screen-film system. A traceability between PTB (which is the primary laboratory for such measurements in Germany) and our institute has been established. A series of measurements has then been performed to feed the simulations prepared by the other members of the project. A full characterisation, and a quality assurance programme has then been prepared with PTB concerning the film digitiser required digitise all the films involved in the project. This work has shown that medical radiographs should be digitized with appropriate systems in order to reduce the loss of relevant information.

Finally, a test object has been designed and tested to subjectively and objectively assess image quality in the field of chest and lumbar spine radiographs. Three strustures have been chosen : high contrast spheres, low contrast paraboloidal objects and filaments. The measurements have then been compared with Monte Carlo simulations performed by other partners to demonstrate that predictivity was possible in the field of radiography.

The project has been concluded by an ERPET training course and a scientific meeting organised in Malmš in June 99 where most of the results have been presented.

References in databases (English)

Swiss Database: Euro-DB of the
State Secretariat for Education and Research
Hallwylstrasse 4
CH-3003 Berne, Switzerland
Tel. [phone]
Swiss Project-Number: 95.0117