Science Spin July 2008 : Sponsored Content - Science Fouindation Ireland

Maths applied to benefit patients

By Sean Duke

In many developed nations such as the USA, it is almost routine now that expensive PET (Positron Emission Tomography) scanners are used as a doctor's aid in cases of stroke, cancer and Alzheimer's disease. The powerful scans produced from PET can help doctors define the precise nature of the patient's illness, and from that starting point to chart the best course of treatment for a patient, and improve patient outcomes. In Ireland PET is currently only available in a limited number of locations such as the Blackrock Clinic and in Belfast. However, the application of mathematics to MRI (Magnetic Resonance Imaging) and CAT (Computerised Axial Tomography) scanners - scanners that are widely available in Ireland - can help doctors improve patients' diagnosis and treatment without PET.

It is a case of making the most of the medical technology we have, to help patients today, rather than waiting for the day when PET scanners are installed in centres all over Ireland. Until that day, however, we can use mathematical methods to get the best from the technology we have. In this case, we are talking about the statistical analysis of medical imagery. SFI thought that this was of value to Ireland when they funded the 'Statistical Methods in Biomedical Imaging Project', headed by Professor Finbarr O'Sullivan, based at UCC, to the tune of €990,818 over a four-year period.

A unique element of this project is a study to develop, implement and test perfusion scanning with CT and MRI scanners available in Ireland. Perfusion scanning is where doctors track blood flow, following, for example, a stroke, or in early diagnosis of Alzheimer's. Understanding blood flow in such cases is of crucial importance. If this can be done with the scanners we have available it will represent a major advance. This work will also lay the groundwork for developing the mathematical expertise that will be required when PET becomes more widely available in Ireland in future.

Molecular imaging

Molecular imaging in medicine is done to try and better understand the molecular mechanisms of disease 'in vivo'. These techniques are based on the injection of radioactive tracers into the body, and watching where various 'labels' go. MR and CAT scans generate such images, but the more powerful technique is PET. This latter technique provides an understanding how tissues use glucose - which is used for energy - by introducing a radio-labelled form of glucose into the body and watching where it goes. This provides important information on the metabolism of a disease.

Professor Finbarr O'Sullivan is Head of the Department of Statistics at UCC. He has been working on the statistical analysis of PET images in disease for at least 20 years, in collaboration with a group in Seattle USA who are interested primarily in using PET for cancer applications. In cancer, CAT and MR scans can determine where a cancer is, but they can't describe the exact metabolic state a cancer is in - something a PET scan can achieve. The determination of cancer metabolism post-chemotherapy can, for example, help doctors decide what they should do next. They can determine which patients would benefit from more treatment and those who wouldn't benefit.

The state of oxygenation of a tumour, for instance, can be a critical determinant of its metabolic state. Those tumours that have low oxygen levels may have grown so fast that they cut off their own oxygen supply. In these cases, the tumour can hibernate, and be quite resistant to chemotherapy. Using PET, doctors have the information to decide to use more aggressive radiation treatment to deal with a resistant tumour.

In Ireland, the imaging technology that is available is MR and CAT scanners. The goal is, said Professor O'Sullivan, is to set up statistical analytical methods which will enable doctors to extract the maximum amount of useful data from these machines as is possible. In order to do this, a cohort of patients was required, and Prof O'Sullivan turned to the geriatric population, as many of these people coming into a clinic would already be getting MR or CAT scans in order to check for Alzheimer's disease. For the study, the patients are required to remain for an extra two minutes in the scanner.

The aim with these geriatric patients is to determine 'blood flow' - a key parameter in many diseases - in patients with Alzheimer's, stroke or cancer compared to 'normal' patients, in order to improve diagnostics for the disease and better information to make decisions on the care of these patients. A dynamic data set can be analysed to produce a 'perfusion scan'. This is something akin to what a PET scanner might achieve, but using our available machines. It is the best we can do, for the moment.

Prof O'Sullivan said that he was attracted to this project, not in order to get access to a particular data set, as he could get such data elsewhere, but to make a contribution in Ireland to improving 'real' patient outcomes, and to pave the way for the introduction of PET scanners here. The two centres that are participating in the statistical study are the Cork University Hospital and the South Infirmary, the latter is an acute care unit.