AI-Guided Multi-OMIC analysis of endometrial samples for clinical decision making

Analysis of tissue samples from the womb mucosa, the endometrium, is critical for making clinical decisions and understanding disorders associated with reproductive health as well as diseases such cancer. For example, such analyses can be used for identifying risk factors, determining the success of pregnancies or determining the underlying causes of reproductive disorders.

Identifying what type of patterns in a tissue sample of a patient are associated with reproductive disorders is difficult due to inherent variability of the tissue over different phases in the reproductive cycle of a patient as well as the underlying disease state, genetic variations, age and other associated medical conditions. Integrating information from multiple "OMICs" technologies such as sequencing of genetic material of the tissue and its active components as well as microscopic visualisation of the tissue can enable more accurate assessment of a patient's samples for improving clinical decision making. However, it can be quite difficult to integrate different sources of information manually and discover novel patterns or markers of such diseases.
In this project, our aim is to utilise machine learning and artificial intelligence (AI) to help clinicians in discovering novel patterns and associating them with the outcome of different clinically important endpoints. We have previously used such approaches for detection of cancers from whole slide images of tissues as well as prediction of outcome to different treatments. We will build a secure and anonymized repository of endometrial/gynaecological data and develop new AI approaches for their analyses. The proposed project aligns with the MRC's focus area of investment and support in biomedical and health data science as well as MRC's strategy of supporting research using large datasets and developing data science skills to ensure UK's leading role in global health research.