Statistical Inference for Diffusion Measurements in Prostate MRI

Prostate cancer is the second most common cancer death in the UK. It is a general healthcare issue, and up to 1.5M men per year in Europe require testing of their prostate by Magnetic Resonance Imaging (MRI). The problem is that, because current MRI is really only taking images rather than making measurements, the images can vary from one scanner to another. It is difficult for doctors to provide a clear diagnosis for patients from MRI alone. The problem is so important that at present up to 40% of men will receive an unnecessary, painful and risky biopsy due to the lack of clarity from the MRI results. In addition, the fact that we cannot measure something by MRI consistently across different scanners makes it difficult to follow up patients over years suffering from this disease.

To address this problem, we have developed a product that can be scanned together with each patient, and that provides a reference measure, so that we can better compare the results between two scans or between two scanners, for example. But to do so, we need to solve some difficult mathematical problems. The National Physical Laboratory (NPL) can help us achieve this aim.

By solving this, we will be able to offer this unique medical device and service to make MRI much more reproducible and quantitative for each patient, reducing the costs and unnecessary interventions per patient, which we estimate is a market worth £12M per annum. Without it, what we have developed so far cannot be used to its full potential, and thus will not achieve this revolution in the diagnosis of prostate cancer.