Vienna, Austria

ESTRO 2023

Session Item

Saturday
May 13
15:15 - 16:30
Lehar 1-3
Will automation of radiotherapy tasks solve more problems than it causes?
Eduard Gershkevitsh, Estonia;
Myriam Ayadi Zahra, France
1380
Pitch Session
Physics
15:40 - 15:55
Target volume contouring automation: Clinical perspective
Eleonor Rivin del Campo, France
SP-0200

Abstract

Target volume contouring automation: Clinical perspective
Authors:

Eleonor Rivin del Campo1

1Tenon University Hospital, Sorbonne University, Department of Radiation Oncology, Pars, France

Show Affiliations
Abstract Text

Automated target volume delineation is steadily integrating daily clinics. Not only is it used for contouring within the classic radiation therapy (RT) workflow, on the planning CT scan. It actually is key for enabling new adaptive RT (ART) workflows to be carried through within a more acceptable timeframe.
Besides reducing the time needed for contouring, this innovation also reduces interobserver contouring variability. Contouring has always been considered the weakest link in the RT planning workflow, due to interobserver variability. Thus, automated contouring may add robustness to the RT planning workflow. However, special attention must be used when applying it.
There are certain tumor sites, such as head and neck, or post-operative adjuvant treatments, where automated target volume delineations are less satisfactory. Furthermore, adaptation of the automated contours may be done differently depending on the RT workflow context. During the classic RT workflow, when contouring on a planning CT, the radiation oncologist has the time to review the complete medical history, other imaging, and leisurely look through and modify the automated target volume contours. In the case of the ART workflow, with the patient on the table, time efficiency is needed. This should be taken into consideration in this workflow, to avoid any possible contouring editing errors due to the time-crunch.
Additionally, to be able to adapt the automated target volume delineations accordingly, radiation oncologists must have adequate knowledge of contouring of the target volume beforehand. A major question at this time is how to best and effectively train radiation oncologists to contour using automated target volume delineation.
Another consideration are the legal implications with respect to the medical responsibility when these automated target volumes are evaluated and edited. Depending on how a medical error incurs within this workflow, the responsibilities may differ.