As the recipient of a prestigious Avant Early Career Research Program scholarship, Dr Kevin Jang is applying the emerging method of radiomics to get a more detailed look at brain tumours and better direct the ways in which they’re treated.
Dr Jang's use of radiomics involves developing a machine learning model where a computer is trained to recognise thousands of patterns of a brain tumour from MRI data, resulting in more accurate predictions about its features.
By non-invasively determining more conclusive imaging results, this new research could solve a problem faced by many with brain cancer following their radiation therapy.
While high-dose radiotherapy treatment is often used successfully to extend the lives of people with brain cancer, brain tissue that has been irradiated can often show ambiguous features on follow-up scans.
As a result, clinicians have trouble distinguishing between radiation necrosis, an effect of the radiation treatment, and actual tumour. This can have negative effects for patients who may either undergo unnecessary interventions or have the accurate diagnosis of a tumour recurrence delayed.
“Differentiating radiation necrosis from tumour progression can be challenging, with substantial overlap of imaging features on conventional MRI. Accurate diagnosis is essential, as the way we manage these two entities is very different,” explains Dr Jang.
Employing AI as part of the radiomics method, Dr Jang aims to accurately distinguish between radiation necrosis and recurrent brain tumours, thereby identifying the best response sooner.
“By developing a reliable and non-invasive method to distinguish radiation necrosis from tumour progression, we can better target treatment strategies and monitor therapeutic response.”Dr Kevin Jang, researcher and radiation oncology registrar