Ethos

Performance Metrics

Clinical Evidence

The published record answers the implementation questions — can the AI contours be trusted as a starting point, does the adapted plan improve on the scheduled one, does the workflow fit a clinical timeslot — rather than the outcome question of whether adapting daily changes toxicity or control. Herlev Hospital in Copenhagen, the first clinical site worldwide, published the foundational implementation study (Sibolt et al., Physics and Imaging in Radiation Oncology 2021): thirty-nine pelvic cases across bladder, rectum, anal and prostate cancer, with one hundred adaptive sessions simulated on a pre-clinical software release before the first patients were treated. The most informative acceptance data come from an Australian multi-fraction analysis (Byrne et al., Journal of Applied Clinical Medical Physics 2022): across 366 prostate fractions — 182 simulated, 184 clinically delivered — 11% of the AI-generated influencer contours needed no change and a further 81% needed only minor edits, leaving under a tenth requiring substantial physician rework. A San Diego group (Moazzezi et al., JACMP 2021) probed the harder configuration of fully unedited auto-segmentation for prostate patients and mapped where edits genuinely change dose — a useful boundary study for departments deciding how much human editing the workflow needs. Three caveats frame the record. The evidence is concentrated in the pelvis, where deformation is largest and adaptation earns its keep; thoracic and abdominal use rests on later, thinner literature. Much of the foundational work ran on pre-clinical or early software releases, and the 2024 Ethos 2.0 generation (K232923) has no published clinical evaluation of its own that this review verified. And no randomised or controlled outcome study of Ethos-based adaptation was located — the dosimetric case for daily adaptation is consistent and plausible, but improved patient outcomes remain an inference, not a published result.

Clinical Pulse

Verified clinician sentiment

Verified clinician reviews launching soon · Apply to contribute →

Reviews are signed by clinicians verified against their regulator's public register.

Inside the Auris+ Listing

Regulatory Approvals

Safety Record

No recalls or field safety notices specific to Ethos were found in indexed public sources as of August 2026 — read as "none found", not an exhaustive audit, since the FDA recall and MAUDE databases could not be queried directly from this environment; Varian's wider treatment-planning portfolio (notably Eclipse) does carry Class 2 recall history, so the absence of an indexed Ethos entry should be re-checked against the register when access allows. The clinical risks are those of compressed on-couch decision-making: contour review under time pressure invites automation bias — the Byrne data showing most AI contours need little editing is reassuring for workflow and simultaneously the mechanism by which an occasional bad contour could pass; and the plan-selection step concentrates a consequential dosimetric decision into minutes. Departments publishing their implementations consistently pair the workflow with physicist plan checks and physician sign-off at every fraction, which is also how the device is labelled to be used.

Intended Use & Indications

Ethos is a radiotherapy delivery system built on the Halcyon ring gantry, cleared together with its Treatment Management and Treatment Planning software (K192377, February 2020). At each fraction the system acquires a cone-beam CT, and the software segments the influencer organs with deep-learning models, propagates the target volumes, and re-optimises the dose distribution against the anatomy of the day. The clinician reviews and edits the generated contours, then chooses between the scheduled plan and the adapted plan before delivery; the adaptive loop is decision support inside a physician-controlled workflow, not an autonomous treatment decision. Varian markets the end-to-end adaptive session as fitting a typical 15-minute timeslot — a vendor claim, though the published implementation studies confirm session times compatible with routine slots. The AI Segmentation component carries its own clearance chain (K203469, K211881), and the 2024 clearance of Ethos Treatment Management 3.0 with Treatment Planning 2.0 (K232923) marks the second product generation. The indication wording here is paraphrased from manufacturer materials and register extracts — the 510(k) letters could not be read directly from this environment.