Detecting artefacts in Pulmonary Embolism CT images
Lung iodine maps have the potential to add value in the assessment of acute and chronic pulmonary embolic disease. However, a variety of patient and protocol related factors can result in image artefacts that may impact the clinical confidence of interpreting the images.Indeed, patients with chronic thromboembolic disease are particularly challenging to diagnose, resulting in this treatable diagnosis often being missed –local published data suggests that Chronic Thromboembolic Pulmonary Hypertension (CTEPH) may be missed in as many as 50% of patients. It is thus important to have a tool to aid fast and accurate detection of pulmonary embolisms.
Using data from the Royal United Hospital, this project aims to provide doctors with a robust algorithm to assess the trustworthiness of artefacts with the hope of improving patient diagnosis. Advanced methods from Convex Optimisation and Bayesian Statistics will be used to build a robust uncertainty quantification tool for artefacts identified by doctors as shown below.
The uncertainty quantification tool will aid doctors in making decisions on the type of treatment to use on patients.
The images on this web pages are taken from:
 https://www.cancerimagingarchive.net/ (top image),  Repetti, Audrey, Marcelo Pereyra, and Yves Wiaux. “Scalable Bayesian uncertainty quantification in imaging inverse problems via convex optimization.” SIAM Journal on Imaging Sciences 12, no. 1 (2019): 87-118.