Poster Poster Program Therapy Physics

AI-Based Vertebral Body Auto-Contouring for Secondary Verification of Spine CBCT Registration

Abstract
Purpose

Automatic intensity-based CBCT registration for initial spine treatment setup is challenging due to the similar appearance of vertebral bodies (VBs), which can lead to misalignment. This study evaluated the feasibility of using AI-based VB auto-contouring as a secondary check to detect and alert clinical teams of potential setup errors.

Methods

Nineteen spine cases with target VBs between C3 and L2 were retrospectively analyzed. Limbus v1.8 software auto-generated individual VB contours on each patient’s helical planning CT (HCT). Contour accuracy was reviewed, and the target VB and/or two adjacent VBs were selected. For each CBCT acquired on an Elekta XVI system, a synthetic CT (sCT) was generated in Elekta Admire and imported into Limbus for VB auto-contouring. The center of mass (COM) of selected VBs was calculated relative to the treatment isocenter. The difference between COM positions from the HCT and sCT datasets was computed as the table shift for that registration and compared to clinical XVI registrations.

Results

The secondary alignment check took ~2.5 minutes to complete. Four patients were excluded because Limbus software failed to generate VB auto-contours on sCT. For the remaining 15 patients, superior-inferior shift discrepancies were <10 mm, with a mean difference between the clinical shift and the secondary check of −0.7 ± 3.0 mm (range: −1.8 to 7.0 mm). Review of auto-contoured VBs indicated discrepancies were primarily due to uncertainties in generating VB contours on the sCT. In three out of the 15 cases, initial shifts showed larger discrepancies (−20.4, 34.7, and 27.7mm in the superior-inferior direction), triggering alerts for clinicians.

Conclusion

AI-based VB auto-contouring on HCT and sCT is a feasible method for detecting potential spine mis-registrations. Based on auto-contouring uncertainties on sCT, a 10mm table shift discrepancy threshold appears appropriate for flagging possible spine misalignment.

People

Related

Similar sessions

Poster Poster Program
Jul 19 · 07:00
Python-Based Automation Framework for Annual Machine QA Data Archiving In Qatrack+

Annual water-tank measurements help ensure beam characteristics remain consistent with commissioning baselines. However, the lack of a standardized processing workflow and decentralized data storage makes it difficult to analyze...

Syed Bilal Ahmad, PhD
Therapy Physics 0 people interested
Poster Poster Program
Jul 19 · 07:00
User Expectations and Current Availability of HDR Brachytherapy Audits In Europe

The aim of this work was to evaluate the need to implement more dosimetric audits in high‐dose‐rate brachytherapy (HDR-BT) in Europe and to identify which characteristics such audits should meet according to users.

Javier Vijande, PhD Laura Oliver Cañamás
Therapy Physics 0 people interested