Using digital technology to improve our understanding of rhabdomyosarcoma

Project title: Deep learning: An integrated approach to define clinical significance to components of the tumour microenvironment of rhabdomyosarcomas

Lead investigator: Dr Yinyin Yuan, The Institute of Cancer Research
Funded by The Little Princess Trust and administered by CCLG
Awarded January 2019
Award: £98,292.50

Rhabdomyosarcomas are rare cancers but a major cause of death from cancer in children. Looking at cancer cells down a microscope is traditionally used for diagnoses and contributes to deciding treatment. Molecular data is starting to augment this process. However, many biological features are not currently objectively assessed that are likely to impact on the clinical course of rhabdomyosarcoma including:

  1. Resemblance of rhabdomyosarcoma cancer cells to different stages of muscle cell development.
  2. The mixture of normal cells with cancer cells, such as cells from the immune system.
  3. The abnormal microenvironment that develops inside cancers (e.g. lack of oxygen that can inhibit response to therapies).

Our proposal takes microscopic analyses of rhabdomyosarcomas to a new level to quantitativey assess how these features alone and in combination impact on clinical behaviour. We will:

  1. Develop a tailored state-of-the-art digital mapping approach using artificial intelligence to characterise biological features of cells in a large number of rhabdomyosarcoma samples taken from well-characterised patients before their treatment on previous clinical trials.
  2. Identify correlations between the defined features and clinical data to determine factors that predict clinical behaviour and outcome.
  3. Compare the digital image mapping results in patients with those in experimental 3D models we have developed that show a range of features related to stages of muscle development and different levels of oxygen. This may link results from patients to the models that will support use of the models to develop new treatment strategies based on targeting the clinically relevant molecular/cellular features we define.

This approach will comprehensively determine the cellular architecture of rhabdomyosarcomas. Key features correlated with disease progression are expected to predict patients that will benefit from new therapeutic strategies through our close links with a new international clinical trial for rhabdomyosarcoma patients.