MSOR09  Presentation Time: 9:10 AM: Improving Treatment Quality in Multi-Catheter Breast Brachytherapy by Decision-Tree Based Treatment Adaption and Implementation of Patient Immobilization

Purpose Interstitial multi-catheter breast brachytherapy (ibBT) performed as accelerated partial breast irradiation (APBI) represents a well-established therapy modality for early breast cancer. Despite good clinical outcomes, treatment quality assurance appears important. In this work, we present for the first time a retrospective assessment of geometric and dosimetric implant stability with an identification of underlying causes of implant variations. Based on our results, we suggest a decision-tree aiming to identify patients requiring treatment adaption. Furthermore, we investigated the effects of patient immobilization on geometric and dosimetric implant stability. Materials and Methods For the first part, we considered 100 patients receiving ibBT as APBI (9x3.8 Gy or 7x4.3 Gy). We compared planning computed tomography scans (pCT) to control-CTs (cCT) acquired halfway through the treatment period. For assessing geometric stability, Fréchet-distance and button-to-button distance changes of all catheters as well as variations of Euclidean distances of all dwell positions were determined. The location-dependency of dwell position variations was investigated. Dosimetric effects were evaluated by isodose volume calculations, target volume transfers and re-contouring of organs at risk (skin, ribs). Based on our results and particularly correlations found between individual geometric and dosimetric parameters, we established a decision-tree (Fig.1) for identifying patients requiring re-planning. Moreover, we implemented the use of a vacuum cushion for patient immobilization during treatment for 25 consecutive patients and investigated the corresponding effects on implant stability. Results In the first part, Fréchet-distance and dwell position deviations >2.5mm as well as button-to-button distance changes >5mm were detected for 5%, 2%, and 6.3% of catheters, but for 32%, 17%, and 37% of patients, respectively. Variations were predominantly observed in the lateral breast and close to the ribs, e.g. due to different arm positions. We observed small dosimetric effects to the target volume in general (coverage index variation: (1.4±1.8)%), but larger skin dose variations. Based on our results, we established a decision-tree focusing on geometric stability and organ at risk (OAR) dosimetry. This is, since OARs were much more sensitive to dosimetric variations than the actual target, and geometric implant variations and corresponding isodose volume changes were significantly correlated also with dosimetric variations of the target volume. If the decision-tree is applied to our patient cohort, 14 and 1 patients would have required treatment adaption due to skin dose variations and target coverage loss, respectively. After the implementation of patient immobilization, none of the 25 examined patients required treatment adaption, and substantial improvements of geometric and dosimetric stability were observed. Fréchet-distance and dwell position deviations >2.5mm as well as button-to-button distance changes >5mm were detected for only 3%, 0.2%, and 1.1% of catheters, and 24%, 4%, and 8% of patients, respectively. Conclusions ibBT shows a high implant stability in general, but considering particularly skin dose changes is important. Our work revealed partly significant correlations of geometric and dosimetric implant variations, forming the basis for a decision-tree concept for treatment adaption. Implant stability can be substantially improved by patient immobilization. Interstitial multi-catheter breast brachytherapy (ibBT) performed as accelerated partial breast irradiation (APBI) represents a well-established therapy modality for early breast cancer. Despite good clinical outcomes, treatment quality assurance appears important. In this work, we present for the first time a retrospective assessment of geometric and dosimetric implant stability with an identification of underlying causes of implant variations. Based on our results, we suggest a decision-tree aiming to identify patients requiring treatment adaption. Furthermore, we investigated the effects of patient immobilization on geometric and dosimetric implant stability. For the first part, we considered 100 patients receiving ibBT as APBI (9x3.8 Gy or 7x4.3 Gy). We compared planning computed tomography scans (pCT) to control-CTs (cCT) acquired halfway through the treatment period. For assessing geometric stability, Fréchet-distance and button-to-button distance changes of all catheters as well as variations of Euclidean distances of all dwell positions were determined. The location-dependency of dwell position variations was investigated. Dosimetric effects were evaluated by isodose volume calculations, target volume transfers and re-contouring of organs at risk (skin, ribs). Based on our results and particularly correlations found between individual geometric and dosimetric parameters, we established a decision-tree (Fig.1) for identifying patients requiring re-planning. Moreover, we implemented the use of a vacuum cushion for patient immobilization during treatment for 25 consecutive patients and investigated the corresponding effects on implant stability. In the first part, Fréchet-distance and dwell position deviations >2.5mm as well as button-to-button distance changes >5mm were detected for 5%, 2%, and 6.3% of catheters, but for 32%, 17%, and 37% of patients, respectively. Variations were predominantly observed in the lateral breast and close to the ribs, e.g. due to different arm positions. We observed small dosimetric effects to the target volume in general (coverage index variation: (1.4±1.8)%), but larger skin dose variations. Based on our results, we established a decision-tree focusing on geometric stability and organ at risk (OAR) dosimetry. This is, since OARs were much more sensitive to dosimetric variations than the actual target, and geometric implant variations and corresponding isodose volume changes were significantly correlated also with dosimetric variations of the target volume. If the decision-tree is applied to our patient cohort, 14 and 1 patients would have required treatment adaption due to skin dose variations and target coverage loss, respectively. After the implementation of patient immobilization, none of the 25 examined patients required treatment adaption, and substantial improvements of geometric and dosimetric stability were observed. Fréchet-distance and dwell position deviations >2.5mm as well as button-to-button distance changes >5mm were detected for only 3%, 0.2%, and 1.1% of catheters, and 24%, 4%, and 8% of patients, respectively. ibBT shows a high implant stability in general, but considering particularly skin dose changes is important. Our work revealed partly significant correlations of geometric and dosimetric implant variations, forming the basis for a decision-tree concept for treatment adaption. Implant stability can be substantially improved by patient immobilization.