Comparison of Tumor Bed Delineation Using a Novel Radiopaque Filament Marker Versus Surgical Clips for Targeting Breast Cancer Radiotherapy

Abstract

Background: Accuracy of tumor bed (TB) delineation is essential for targeting boost doses or partial breast irradiation. Multiple studies have shown high interobserver variability with standardly used surgical clip markers (CMs). We hypothesize that a radiopaque filament marker (FM) woven along the TB will improve TB delineation consistency. Methods: An FDA-approved FM was intraoperatively used to outline the TB of patients undergoing lumpectomy. Between January 2020 and January 2022, consecutive patients with FM placed after either (1) lumpectomy or (2) lumpectomy with oncoplastic reconstruction were identified and compared with those with CM. Six “experts” (radiation oncologists specializing in breast cancer) across 2 institutions independently defined all TBs. Three metrics (volume variance, dice coefficient, and center of mass [COM] deviation). Two-tailed paired samples t tests were performed to compare FM and CM cohorts. Results: Twenty-eight total patients were evaluated (14 FM and 14 CM). In aggregate, differences in volume between expert contours were 29.7% (SD ± 58.8%) with FM and 55.4% (SD ± 105.9%) with CM (P < 0.001). The average dice coefficient in patients with FM was 0.54 (SD ± 0.15), and with CM was 0.44 (SD ± 0.22) (P < 0.001). The average COM deviation was 0.63 cm (SD ± 0.53 cm) for FM and 1.05 cm (SD ± 0.93 cm) for CM; (P < 0.001). In the subset of patients who underwent lumpectomy with oncoplastic reconstruction, the difference in average volume was 21.8% (SD ± 20.4%) with FM and 52.2% (SD ± 64.5%) with CM (P<0.001). The average dice coefficient was 0.53 (SD ± 0.12) for FM versus 0.39 (SD ± 0.24) for CM (P < 0.001). The average COM difference was 0.53 cm (SD ± 0.29 cm) with FM versus 1.25 cm (SD ± 1.08 cm) with CM (P < 0.001). Conclusion: FM consistently outperformed CM in the setting of both standard lumpectomy and complex oncoplastic reconstruction. These data suggest the superiority of FM in TB delineation.