Dosimetry and optimisation in high dose fluoroscopic and fluorographic procedures
This thesis describes the search for a practical skin dosimetry method for cardiac catheterization procedures, and the application of an optimisation strategy in barium enema imaging. Kodak EDR2 film was characterised across the range of exposure conditions used in the cardiac catheterization laboratory. Its dose-response curve was modelled using a novel equation, and overall uncertainty in film response was estimated. The film saturated at 1 Gy, limiting its usefulness for skin dosimetry. Its performance was found to be strongly dependent on beam filtration, an aspect that had not previously been studied. The film was then used to measure skin doses to patients undergoing coronary angiograms and angioplasties. For angiograms, all skin doses were well below 1 Gy. For angioplasties, 23% of films showed localised saturation, indicating peak skin doses of at least 1 Gy. Dose-area-product was shown to be a poor predictor of high peak skin dose. A mathematical model was developed and software written, to calculate patient skin dose maps from exposure and projection data stored in the image files. This offered a practical method for assessing the magnitude and approximate location of the peak skin dose. Accuracy was limited by a lack of information regarding fluoroscopic exposures, couch position and beam limitation. After including an estimated contribution from fluoroscopy, the model successfully identified those patients whose skin doses exceeded 1 Gy. Following a baseline survey of local barium enema practice, several dose reduction methods were considered. It was decided to introduce copper filtration. 0.1 mm copper reduced mean patient DAP by 37%, without any measurable difference in contrast detail detectability. A detailed phantom study determined the optimal copper thickness as 0.3 mm. This reduced mean patient DAP by 55%, relative to the baseline survey. A visual grading analysis study showed no significant difference in clinical image quality.