The impact of viewing conditions on image quality assessment tasks in mammographic quality control

Papathanasiou, S

Abstract

Objective:
Mammography is a technically demanding x-ray examination requiring optimal image quality to enable the early detection of breast cancer. Quality control (QC) processes ensure that the equipment in the diagnostic imaging process is functioning and maintained at the highest level. High image quality is fundamental to the accuracy of the quality procedures, technical evaluation, and diagnosis. This research investigates the impact of display monitor specification and environmental conditions on quality control procedures within the mammographic departments in both screening and symptomatic services.
Methods:
The research described in this thesis was conducted in four phases to understand the different environmental conditions and monitor specifications that exist in mammography departments and to understand the impact of these different conditions during the QC processes. This research is focused predominantly on the technical and environmental conditions found in acquisition rooms due to a lack of guidelines for the viewing conditions in these areas during the QC procedures. This knowledge gap was identified by a systematic review considering monitor specifications and viewing conditions in mammography.
Each phase of this research has its own methods with a purpose of achieving a specific set of aims. The first phase was a survey of the clinical environments of the mammographic departments in North West England aiming to (i) investigate the illumination levels that exist in the clinical environments, (ii) evaluate the monitors, (iii) evaluate the room layout of the acquisition and reporting rooms, and finally (iv) evaluate test object images in these environments in order to provide a baseline assessment of QC image assessment under different environmental and technical conditions.
During the second phase of this research, nine TORMAM test object images with different technical characteristics, simulating a range of images with different qualities from clinical environments, were obtained. The images were acquired with different target / filter combinations and PMMA thicknesses. The quality of the images was assessed via objective image evaluation tools (AutoPIA and ImageJ software). The next phase of this research was the main experiment that was divided into two sub-phases. During the first sub-phase 16 participants evaluated the nine images under 12 different environmental conditions (low, medium, and high illumination level, white and grey wall colour and 2 monitors with high and low technical characteristics). The aim of this phase was to investigate the number of visible structures (groups of multi directional filaments, microcalcifications and low contrast detail discs) under the 12 conditions. During the second sub-phase the participants evaluated the number of low contrast detail discs separately under the same 12 conditions in order to investigate how the visualisation of the different discs can be affected under these conditions. Finally, during the last step of the experiment, the coefficient of reflection and the presence of reflections on the surface of the monitors was also investigated for the 12 experimental conditions.
Results:
The results from the first phase of the research revealed great variances in the illumination levels and monitor characteristics in the acquisition rooms. On the contrary, consistency was noticed in the ambient light levels and monitor specifications in the reporting rooms in the mammographic departments. This phase demonstrated the need for further specialized research for the acquisition rooms, focused on the environmental conditions.
The second phase of the research categorised the images according to their quality based on their target filter combination and PMMA thickness. CNR indices of the low contrast detail discs were reported and variance ratio measurements for the groups of multi directional filaments and groups of microcalcifications were noted. The results from these evaluations were compared with the results obtained from the observer study during the next phase of the research.
The observer study demonstrated great variation in the number of visible structures in the different environmental conditions. The illumination level and the wall colour have a statistically significant effect on the number of the visible structures by the participants during the experiment. A preference for grey wall colour, lower ambient light and a monitor with high specifications was identified by the majority of the participants. Additionally, it was statistically proved that the grey wall colour had positively effect on the number of visible low contrast detail discs highlighted by the participants of the experiment.
Furthermore, further analysis of the results from the last phase of the quality assurance observation study and the evaluation of the coefficient of reflection revealed that white wall colour increased the presence of reflections on the faceplate of the monitors. The white wall colour added a rise on the illumination level in front of the monitor up to 19.94 % ± 7.45 and the grey wall colour added an illuminance rise up to 11.27 % ± 9.56 within the room. Moreover, the location of the light bulb in relation to the monitor’s screen proved of a great importance. It was revealed that structures that should be appear visible to the participants according to their physical indices, were not noted as visible due to their location in the image and the light emitted on the faceplate of the monitor. Finally, 12/16 participants reported that the reflections within the room with the white wall colour made the TORMAM test object observations difficult and characterized the environment as ‘uncomfortable’ and ‘unpleasant’.
Conclusion:
The wall colour, the ambient light level around the monitors and the location of the natural or artificial light sources can have a significant impact on quality control processes. Grey wall colour, low ambient light and high specification monitors allow greatest structure visibility. On the other hand, high ambient light and a white wall colour contributed to increased reflections on the surface of the monitors. This has a negative impact on the technical evaluation of images during QC procedures. This may lead to a detrimental impact on service delivery leading to the need of a better standardization of the acquisition rooms including careful position of the monitors, use of low reflectance wall colours and low illumination level during QC procedures.

Citation

Papathanasiou, S. The impact of viewing conditions on image quality assessment tasks in mammographic quality control. (Thesis). University of Salford

• Thesis Type: Thesis
• Deposit Date: Oct 5, 2022
• Award Date: Nov 1, 2021