MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

What Knowledge-Based Dose Prediction Models Tell Us About Ovoid Vs. Ring Based Brachytherapy Applicators

K Kallis*, B Covele, A Simon, D Brown, D Scanderbeg, K Kisling, C Yashar, J Einck, L Mell, J Mayadev, K Moore, S Meyers, UC San Diego, La Jolla, CA

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: Currently there is a lack of patient-specific tools to guide brachytherapy planning and applicator choice, and little data comparing tandem-and-ring (T&R) to tandem-and-ovoids (T&O). The purpose of this study is to evaluate the accuracy of organ-at-risk (OAR) dose prediction using knowledge-based intracavitary models, and to use these models and clinical data to determine the dose differences of common applicators.

Methods: Knowledge-based models, which use target-OAR distances to predict OAR DVHs, were developed for T&R and T&O applicators. T&O (T&R) models were trained on 356 (62) cases, and validated on 100 (48) cases. Model performance was quantified using ?D2cc=D2cc,actual–D2cc,predicted, with standard deviation (s(?D2cc)) representing model precision. In order to estimate the dose difference of the two applicators, the T&O model was applied to T&R clinical cases, and vice versa. Model-predicted applicator differences were compared to clinically achieved D2cc values for these cases. T-tests were used to compare groups.

Results: For the training dataset, T&O(T&R) model precision was 0.61Gy(0.53Gy), 0.57Gy(0.41Gy), and 0.52Gy(0.54Gy) for bladder, rectum and sigmoid, respectively. T&O(T&R) model performance was similar in the validation, with mean?D2cc±s of 0.02±0.61Gy (-0.03±0.61Gy), 0.00±0.46Gy (0.19±0.42Gy) and -0.02±047Gy (-0.24±0.56Gy) for bladder, rectum and sigmoid. When applying the T&O(T&R) model to T&R(T&O) cases, bladder, rectum and sigmoid D2cc were on average 0.74Gy(0.61Gy), 1.2Gy(0.82Gy) and 0.64Gy(0.10Gy) lower for T&R. Clinical D2cc in EQD2 were significantly lower for T&R (mean deviation=2.67Gy(p<0.05), 10.58Gy(p<0.01) and 7.42Gy(p<0.01) for bladder, rectum and sigmoid), which is similar to recently published results from EMBRACE I (mean deviation=7.7Gy and 3.2Gy for bladder and rectum).

Conclusion: Accurate knowledge-based dose prediction models were developed for two common intracavitary applicators. These models could be beneficial for standardizing and improving the quality of brachytherapy plans. Both models and clinical data suggest that significant OAR sparing can be achieved with T&R over T&O applicators, particularly for the rectum.

Funding Support, Disclosures, and Conflict of Interest: This work was funded by a research grant from Padres Pedal the Cause

Keywords

Brachytherapy, Modeling, Treatment Planning

Taxonomy

TH- Brachytherapy: Treatment planning using machine learning/automation

Contact Email