Poster Poster Program Therapy Physics

Evaluating the Geometric Accuracy of Automatic Applicator Anchor Point Placement for Cervical Cancer HDR Brachytherapy Planning

Abstract
Purpose

To quantify the geometric accuracy of an automatic applicator anchor point placement framework for cervical cancer (CC) high-dose-rate (HDR) brachytherapy implemented in Oncentra Brachy (Elekta).

Methods

Planning MRIs acquired with 3D T1 Axial VIBE from patients with CC receiving HDR brachytherapy between December 2023 and December 2025 were used to develop and test an applicator-based anchor point placement framework. All MRIs were acquired on the day of treatment with the Venezia applicator in place, using intrauterine tandem lengths of 40-70 mm and semi-lunar ovoid diameters of 22-30 mm. Applicator components were manually contoured on each MRI to serve as ground truth for model development. A deep learning model was trained to reconstruct applicator geometry from the MRI. Predicted contours were used as input to an analytic, geometry-based algorithm to determine three clinically defined anchor point locations (Intrauterine Tandem Tip, Ovoid 1: Centre, and Ovoid 1: 1st Marker). Automatically placed anchor points were compared against clinical anchor points in Oncentra Brachy for each treatment plan in the testing dataset.

Results

The retrospective dataset included 44 patients (37 patients, 106 MRIs for development; 7 patients, 21 MRIs for testing). The best model used a UNet++ architecture and demonstrated good geometric agreement for all applicator components (validation/testing DSCs ranging from 0.73-0.92 and 0.79-0.93, respectively), confirming accurate reconstruction of applicator geometry and supporting evaluation of clinically relevant anchor point accuracy. Median 3D distances between clinical and automatic anchor points were 1.1 mm (max 8.7 mm), 1.0 mm (max 1.9 mm), and 2.1 mm (max 4.9 mm) for the Intrauterine Tandem Tip, Ovoid 1: Centre, and Ovoid 1: 1st Marker, respectively.

Conclusion

Automatic anchor point placement achieved median 3D errors of approximately 1-2 mm, supporting future clinical integration to improve HDR brachytherapy planning efficiency for patients with CC.

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