Improving comparability of medical images

Using QA and calibration data from scanners to improve the comparability of medical images 

Device-specific variations in medical image characteristics impact image analysis, treatment planning, dosimetry and prognosis. To account for these variations, regular calibration procedures are undertaken to track the changes and ensure the stability of the imaging system. While calibration data is vital for meaningful image analysis and other calculations, there are no mechanisms in place to link it to the patient images.

Within this case study we aim to establish:

• The feasibility of linkage of the calibration data and patient images.
• The range of inter- and intra-device variations.
• Methods to apply the scanner calibration data to correct for inter- and intra-device variations, to provide the traceability of image values to tissue electron density and to prepare the images for big data applications such as machine learning.

This case study used 23,000 megavoltage CT images from 800 patient cases acquired for the VoxTox study between 2007 and 2017 (Burnet et al. 2017) that were linked to the calibration data from two TomoTherapy machines provided by Cambridge University Hospitals.

We are analysing image uniformity and noise in calibration data and investigating the use of 'calibration proxies' in patient images that should be predictive of daily fluctuations in the image properties. Future work will investigate the feasibility of applying process control techniques to predict machine breakdowns in order to better plan maintenance and minimise downtime.               

This work is ongoing and we are keen to work with as many people as possible. Please contact us if you would like to know more or want to discuss the challenges in your work environment.

Contact us

Partners