Vienna, Austria

ESTRO 2023

Session Item

Monday
May 15
15:00 - 16:15
Strauss 1
What can EBRT learn from brachytherapy and vice versa?
Kari Tanderup, Denmark;
Shirin Abbasinejad Enger, Canada
Symposium
Physics
15:36 - 15:54
Image-guidance, intensity-modulation and AI in brachytherapy
Christian Kirisits, Austria
SP-0861

Abstract

Image-guidance, intensity-modulation and AI in brachytherapy
Authors:

Christian Kirisits1

1Medical University of Vienna, Radiation Oncology, Vienna, Austria

Show Affiliations
Abstract Text

Long lasting experience with image-guidance in clinical practice of brachytherapy resulted in a new state of the art for this treatment modality. It is based on the clinical needs for an image showing the location of the dose delivery device (source path defined by an implant), target volumes and organs at risk in one image, considered as the relevant “room” of brachytherapy dose delivery. Such imaging data serve as a basis for an online planning concept. Variations and uncertainties exists intra-fractional as well as between fractions in terms of one implant for more HDR fractions or PDR pulses. Also for fractionated treatments several workflows have been presented integrating repetitive imaging for verification and potential replan or plan of the day approaches. The use of MRI in definitive brachytherapy treatment is based on more than 20 decades including the clinical application of MRI only workflows. Also the use of ultrasound (US), with a soft tissue contrast that allows in several disease sites similar target volume definition accuracy as with MRI has its place in brachytherapy. The combination of imaging modalities as MRI, US, CT and new developed cone-beam CT devices is beneficial for some workflows. However, the registration between imaging techniques is fundamentally different to external beam techniques, related to the brachytherapy set-up “room” showing the depicted applicators fixed to the surrounding anatomy.

Intensity modulation is an intrinsic characteristic of brachytherapy especially applying the stepping source technology of afterloaders. In combination with manual forward planning based on risk adapted target volumes or inverse planning approaches highly heterogeneous dose distributions can be achieved. Brachytherapy dose painting benefits from high dose gradients for focal treatment as well as risk adapted target dose variation and organ sparing. Modern approaches combine TCP/NTCP findings but also source loading patterns analyzed from clinical trials with intensity modulation. However, daily clinical application often remains cumbersome. Here automated planning workflows could results in clinical relevant benefits, both in terms of accuracy and resources.

The use of AI is increasingly applied in brachytherapy, by now within research settings. AI can assist in the before mentioned automation of the treatment planning process, but also in several other areas. While dose prediction, automated contouring, outcome analysis and decision support might be similar as in radiation oncology in general, image registration, applicator reconstruction and the online planning approach is different for brachytherapy. Here AI developments could results in a huge benefit, but would also need joint efforts from centers, researchers and the industry.