ML

Real-time respiratory motion modelling during image-guided radiotherapy

We have developed the first method for simultaneous tumour and organ-at-risk tracking during image-guided radiotherapy.

schedule Date & time
Date/time
7 Dec 2023 11:00am - 7 Dec 2023 12:00pm
person Speaker

Speakers

Nicholas Hindley
next_week Event series

Content navigation

Description

Title: Real-time respiratory motion modelling during image-guided radiotherapy

Abstract: Cancer is one of the leading causes of death worldwide and approximately 1 in every 2 cancer patients can benefit from radiotherapy. The bulk of lesions leading to cancer-related death reside in the thorax and abdomen, where every structure is constantly moving due to respiration. This motion hinders the effective delivery of radiation as every millimetre of motion that goes unchecked leads to unnecessary healthy tissue damage. We have developed the first method for simultaneous tumour and organ-at-risk tracking during image-guided radiotherapy. Leveraging deep learning, a neural network is trained to learn patient-specific breathing patterns given the data routinely acquired for thoracoabdominal targets. We recently established proof-of-principle for this method, achieving submillimetre accuracy on data acquired for two lung cancer patients. By providing an accurate and accessible way of modelling respiratory motion in real-time, our technology has the potential to drastically reduce treatment uncertainties in image-guided radiotherapy.

 

Location

Robertson Building #46

DNA Room S104
46 Sullivans Creek Road,
The Australian National University,
Canberra, ACT 2600
Australia