Search

• Overlap volume histograms can be used for prediction models of D 2cm3 values. • Prediction models are robust to interstitial needle use and applicator types. • Prediction models can be used in a multi-center setting. • Suboptimal plans can be detected when models are trained on high-quality data. Image-guided adaptive brachytherapy (IGABT) is a key component in the treatment of cervical cancer, but the nature of the clinical workflow makes it vulnerable to suboptimal plans, as the theoretical optimal plan depends heavily on organ configuration. Patient anatomy-based quality-assurance (QA) with overlap volume histograms (OVHs) is a promising tool to detect such suboptimal plans, and in this analysis its suitability as a multi-institutional clinical QA tool is investigated. A total of 223 plans of 145 patients treated in accordance with the current state-of-the-art IGABT protocols from UMC Utrecht (UMCU) and Erasmus MC (EMC) were included. Machine-learning models were trained to predict dose D 2cm3 to bladder, rectum, sigmoid and small bowel with the help of OVHs. For this strategy, points are sampled on the organs-at-risk (OARs), and the distances of the sampled points to the target are computed and combined in a histogram. Machine-learning models can then be trained to predict dose-volume histograms (DVHs) for unseen data. Single-center model robustness to needle use and applicator type and multi-center model translatability were investigated. Performance of models was assessed by the difference between planned (clinical) and predicted D 2cm3 values. Intra-validation of UMCU data demonstrated OVH model robustness to needle use and applicator type. The model trained on UMCU data was found to be robust within the same protocol on EMC data, for all investigated OARs. Mean squared error between planned and predicted D 2cm3 values of OARs ranged between 0.13 and 0.40 Gy within the same protocol, indicating model translatability. For the former protocol cohort of Erasmus MC large deviations were found between the planned and predicted D 2cm3 values, indicating plan deviation from protocol. Mean squared error for this cohort ranged from 0.84 to 4.71 Gy. OVH-based models can provide a solid basis for multi-institutional QA when trained on a sufficiently strict protocol. Further research will quantify the model's impact as a QA tool. [ABSTRACT FROM AUTHOR]

• Overlap volume histograms can be used for prediction models of D 2cm3 values. • Prediction models can be used in a multi-center setting to detect suboptimal plans. • Plan quality improvement is possible for suboptimal plans detected by the prediction model. Image-guided adaptive brachytherapy (IGABT) is an important modality in the cervical cancer treatment, and plan quality is sensitive to time pressure in the workflow. Patient anatomy-based quality-assurance (QA) with overlap volume histograms (OVHs) has been demonstrated to detect suboptimal plans (outliers). This analysis quantifies the possible improvement of plans detected as outliers, and investigates its suitability as a clinical QA tool in a multi-center setting. In previous work OVH-based models were investigated for the use of QA. In this work a total of 160 plans of 68 patients treated in accordance with the current state-of-the-art IGABT protocol from Erasmus MC (EMC) were analyzed, with a model based on 120 plans (60 patients) from UMC Utrecht (UMCU). Machine-learning models were trained to define QA thresholds, and to predict dose D 2cm3 to bladder, rectum, sigmoid and small bowel with the help of OVHs of the EMC cohort. Plans out of set thresholds (outliers) were investigated and retrospectively replanned based on predicted D 2cm3 values. Analysis of replanned plans demonstrated a median improvement of 0.62 Gy for all Organs At Risk (OARs) combined and an improvement for 96 % of all replanned plans. Outlier status was resolved for 36 % of the replanned plans. The majority of the plans that could not be replanned were reported having implantation complications or insufficient coverage due to tumor geometry. OVH-based QA models can detect suboptimal plans, including both unproblematic BT applications and suboptimal planning circumstances in general. OVH-based QA models demonstrate potential for clinical use in terms of performance and user-friendliness, and could be used for knowledge transfer between institutes. Further research is necessary to differentiate between (sub)optimal planning circumstances. [ABSTRACT FROM AUTHOR]