INFERENCE
IdentificatioN oF biomarkErs foR Early diagNosis of CancEr
Early diagnosis is central to improving outcomes for patients with cancer. For cancers without specific risk factors, or with no screening programs are difficult to diagnose and patients often present with non-specific symptoms. Unfortunately this means that these patients are often diagnosed late on in the development of the disease and treatment options are reduced.
Using Fourier transform infrared (FTIR) spectroscopy, we are aiming to create a blood test that would help identify patients who have a high likelihood of cancer. To do so we are collecting blood samples from: i) participants for whom the assessing clinician has identified a clinical need to undergo a diagnostic investigation to exclude a new brain, chest, abdominal or pelvic abnormality, where the differential diagnosis includes cancer, or ii) participants with a recent new cancer diagnosis, before surgery, chemotherapy or radiotherapy. We will use the data from anlaysed samples, and machine learning, to train an algorithm to ascertain whether a new cancer diagnosis can be accurately predicted at the time of presentation with non-specific symptoms. This would permit prioritisation for diagnostic imaging, support clinical decision making and hopefully earlier diagnosis.