UCLA Journal of Radiation Oncology December First Edition - Flipbook - Page 10
UCLA RADIATION ONCOLOGY JOURNAL
AGE OF A.I.: ETHOS
Ethos uses AI to automatically contour certain organs in order to
expedite the adaptive planning process
T
he UCLA Department of Radiation
Oncology will be installing an Ethos
treatment machine at its Santa
Clarita facility. Ethos, Varian’s newest
offering, leverages high quality conebeam CT imaging, artificial intelligence
driven organ auto-segmentation, and a
simplified planning workflow to make
online adaptive radiation therapy more
accessible. Until now, only MRI-guided
systems such as our ViewRay MRIdian
have been able to offer online adaptive
therapy. While UCLA has been able to
routinely provide this service to patients,
not every clinic has the resources required
for an MR-guided radiotherapy system.
Importantly, online adaptive therapy
doesn’t necessarily require MR images –
in order to adapt, we just need in-room
imaging that is high enough quality to use
for planning. Although the CBCT images
on our TrueBeams and the MVCT on our
TomoTherapy are adequate for patient
alignment, neither are up to this standard.
In contrast, Ethos uses an iterative image
reconstruction technique to provide CBCT
imaging where targets and organs at risk
can clearly be identified and delineated.
The enclosed gantry allows for a rotation
speed of 4 RPM, which facilitates
acquisition of an image in 17 seconds and
therefore greatly limits patient motion
artifacts.
In addition to higher quality imaging,
Ethos uses AI to automatically contour
certain organs in order to expedite the
adaptive planning process. Contouring is
often the most time consuming part of the
workflow, and online adaptive needs to
be rapid in order to be accurate. Ethos is
able to reliably identify several “influencer
structures” for multiple treatment sites.
These influencer structure contours are
verified by clinicians and are then used to
guide the generation of a new plan that is
better suited to the patient’s anatomy on
that day. Ethos promises to make online
adaptive therapy better and available to
a greater number of patients, and we
look forward to being one of the earliest
adopters of this technology.
Contributed by:
Dylan O'Connell, PhD
Assistant Professor, Department of Radiation Oncology
Dr. O'Connell received his bachelor's degree in Physics
from Tufts University in 2013, and his Ph.D in Biomedical
Physics from UCLA in 2018. Subsequently, he completed
the medical physics residency program at UCLA before
joining the faculty in 2020. His research interests include
improving 4DCT reconstruction using a respiratory
motion model, motion compensated cone-beam CT
reconstruction, online adaptive therapy, and in-house
clinical software safety.
10