Computing Publications

Publications Home » Kinetic-spatial filters for PET d...

Kinetic-spatial filters for PET dynamic studies in oncology

Katherine Gray

Master's Thesis
Imperial College London
August, 2008

A novel kinetic filtering technique has been developed, which enhances the visualisation of tumours imaged using [18 F]Fluorothymidine Positron Emission Tomography. [18F]Fluorothymidine is a cell proliferation marker which undergoes a predominantly hepatic metabolism, and therefore shows a high level of accumulation in the liver, as well as in rapidly proliferating cancerous tissue.

A linear classification algorithm to isolate cancerous tissue against such healthy organs was thus developed and validated using 28 images of patients with locally advanced or metastatic breast cancer. A reduction in signal from the liver and heart of over 80% is observed following application of the filter, whilst the majority of signal from both primary tumours and metastases is retained. The algorithm can be applied to images reconstructed using either analytic or iterative methods, and a scan acquisition time of 60 minutes is thought to be sufficient to obtain the necessary kinetic data.

The technique could have direct relevance in clinical oncology, as well as being applicable to a variety of tracers, and to the removal of other organs, including the bladder.


Supervised by Lula Rosso, Federico Turkheimer and Eric Aboagye

Available at

BibTEX file for the publication built & maintained by Ashok Argent-Katwala.