Main Session
Sep 29
PQA 04 - Gynecological Cancer, Head and Neck Cancer

2880 - Study of Artificial Intelligence-Assisted 3D Printed Vaginal Templates in 3D Brachytherapy for Cervical Cancer

10:45am - 12:00pm PT
Hall F
Screen: 7
POSTER

Presenter(s)

Xiaoyan Wu, - Shenzhen Hospital, Cancer Hospital, Chinese Academy of Medical Sciences, Shenzhen,

X. Wu, H. Song, and Y. Zhang; Shenzhen Hospital, Cancer Hospital, Chinese Academy of Medical Sciences, Shenzhen, China

Purpose/Objective(s): To analyze the application value of AI-assisted 3D printed vaginal templates in three-dimensional brachytherapy for cervical cancer.

Materials/Methods: This study included 40 patients with advanced cervical cancer, treated between August 2022 and February 2024. Twenty patients were treated using 3D printed vaginal templates with needle insertion paths designed by artificial intelligence, while the remaining 20 patients were treated with a three-tube applicator (Fletch). CT localization was performed for each patient, and treatment plans were developed using a reverse dose optimization algorithm.The D90 (the minimum absorbed dose received by 90% of the volume) of the target area HR-CTV, and the D2cc values for the bladder, sigmoid colon, rectum, and small intestine were calculated for the three-dimensional brachytherapy plan.Independent sample t-tests were used to compare the doses of CTV (clinical target volume) and OAR (organs at risk) between the two applicators in three-dimensional brachytherapy for cervical cancer.

Results: The D90 values of the target area HR-CTV for the three-tube applicator and the AI-assisted 3D printed vaginal template were (626.64 ± 11.92) cGy and (621.39± 9.43) cGy, respectively (P=0.13).The D2cc values for the bladder, sigmoid colon, rectum, and small intestine in the three-tube applicator were (490.89 ±58.34) cGy, (332.46 ± 93.99) cGy, (348.17 ± 48.23) cGy, and (327.35 ± 92.30) cGy, respectively.The D2cc values for the bladder, sigmoid colon, rectum, and small intestine in the AI-assisted 3D printed vaginal template were (420.15 ± 40.97) cGy, (246.68 ± 58.05) cGy, (334.73 ± 47.08) cGy, and (221.13 ± 100.67) cGy, respectively (P=0.000, P=0.002, P=0.378, P=0.001).

Conclusion: In three-dimensional brachytherapy for cervical cancer, the AI-assisted 3D printed vaginal template can reduce the operational difficulty for doctors and provides better target area conformity. Compared with the three-tube applicator, it better protects the bladder, sigmoid colon, and small intestine, offering certain therapeutic advantages.