The Royal Marsden Diagnosis and treatment

Colour-coding sarcomas

A multidisciplinary team at The Royal Marsden has produced software that can colour-code the different components of a tumour

Altered image: 3D MRIs of sarcomas before (left) and after (right) treatment

A group comprising radiologists, physicists and the sarcoma team at The Royal Marsden has come together to produce software that can colour-code the different components of a tumour from 3D magnetic resonance images (MRI).

Soft-tissue sarcomas are made up of several components, which respond differently to various treatments. A biopsy won’t detect all the different components within the tumour.

The 3D rotating maps are produced using artificial intelligence techniques to help interpret the complexities within MRI scans that measure tumour blood supply and how densely packed the tumour cells are.

By showing the different areas of the sarcoma in different colours, clinicians can see what part of the tumour is the most aggressive and which areas are responding to treatment.

The system was developed with the support of Royal Marsden physicist Matthew Blackledge, who wrote the software for it.

This is exciting because it gives clinicians the ability to develop personalised and adaptive treatment plans

Dr Christina Messiou, Consultant Radiologist

It was a partnership between the NIHR Imaging Clinical Research Facility at The Royal Marsden, The Institute of Cancer Research and Cancer Research UK.

Dr Christina Messiou, Consultant Radiologist, said: “When sarcomas are treated with radiotherapy or some drugs, they can swell and it is very difficult to tell if they are responding. The new imaging system allows us to see which parts of the tumour are responding by looking at how the tumour is functioning, without having to perform repeated biopsies.

“This is exciting because it gives clinicians the ability to develop personalised and adaptive treatment plans. These developments have only been possible by collaborating with researchers from very different backgrounds.

"Matthew’s ‘machine-learning’ expertise – where computers are taught how to do things that would be considered too difficult or time-consuming for a human to do – has taken this work to another level.

“The new imaging system allows us to see which parts of the tumour are responding.”