Single cell spatial transcriptomics and radiomics profiling of non-small cell lung cancer for precision PD-1/PD-L1 checkpoint blockade immunotherapy

Context Non-small cell lung cancer affects many people and is the most common cause of cancer deaths worldwide. Most of its cases are diagnosed late, where the disease has already reached important structures or spread to other parts in the body, making it harder to treat. A new treatment for these difficult cases is now available, called immunotherapy, which works by helping the body’s own system to attack the cancer cells and slow down their spread. There is however a great deal that we don’t yet know about how it treats the disease at a deeper level. Over the years, scientists have invented new ways of looking at cancer cells and how the body’s own immune system works around them. Recently, they have invented a new technology, called single-cell spatial transcriptomics, which allows a closer look at the genetic makeup and function of the cells without taking them out of the place they are found, therefore telling us where this information about the cells is in and around the cancer. Patients routinely receive image scans to look at their cancer. There is a great amount of information about their cancer in these scans that we can’t see with our naked eye. With the help of computers, scientists are now able to get this information, called radiomics, and turn it into something useful, called an imaging biomarker, that could help with cancer treatment. The challenge the project addresses Not everyone’s cancer is the same. To best choose a treatment, doctors need to have a good understanding of the disease first, which often requires getting some tissue, by passing a needle through skin and lung, a thin tube down the throat into the lung, or surgery. In the case of immunotherapy, choosing the right patient to treat is important because the treatment can sometimes cause extra harm to the body, which can result in death in the most serious cases. Currently, to help making their call, doctors use a special lab test performed on the tissue taken, called PD-L1 expression. However, this test is not very reliable, and cannot predict how the cancer responds to treatment in many cases. We therefore need a better test to help us choose the right patients to treat. Aims and objectives This study aims to use two cutting edge technologies, single cell spatial transcriptomics and radiomics, to better understand non-small cell lung cancer and discover new ways of testing the disease for its suitability for immunotherapy. Using artificial intelligence, we will also look for a way to use medical images to predict the useful information from single cell spatial transcriptomics, which is more difficult and expensive to get in real life. Potential applications and benefits We will find new ways of testing non-small cell lung cancer for its suitability for immunotherapy and help making a bespoke treatment plan best suited for each individual patient. We will also get to know more about how the body's immune cells work around the cancer, which could help us develop new drugs to treat it.