A clinical evaluation of HyperSight CBCT enhanced offline adaptive breast radiotherapy
TL;DR
It is suggested that HyperSight provides HU accuracy closer to that of simulation CT, improved image quality, and more accurate CBCT-based dose recalculation than Halcyon 3.1, thereby supporting the identification of clinically meaningful anatomical changes.
Abstract
Introduction Anatomical changes such as seroma shrinkage, soft-tissue deformation, and tumor-bed displacement frequently occur during breast radiotherapy and may compromise target coverage or increase unnecessary dose to surrounding tissues when the initial plan is maintained throughout treatment. Offline adaptive radiotherapy (ART) requires CBCT systems with CT-equivalent HU accuracy and robust dose-recalculation accuracy to reliably quantify these changes. This study evaluated whether the newly integrated HyperSight CBCT on the Halcyon platform provides high image quality, HU-to-electron density accuracy, and dosimetric reliability to support offline ART in breast cancer. Methods Twelve patients who underwent treatment with both Halcyon 3.1 and Halcyon-HyperSight systems were analyzed. Image quality was assessed using structural similarity metrics (SSIM, RMSE, NRMSE), uniformity, signal and noise characteristics, contrast-to-noise ratio (CNR), and edge-based indices. HU-to-ED curves were generated using a tissue-equivalent electron-density phantom and applied for CBCT-based dose recalculation. Results Compared with Halcyon 3.1, HyperSight demonstrated significantly improved structural similarity to simulation CT, with higher SSIM and lower RMSE/NRMSE, as well as significantly improved uniformity and edge sharpness. HU-to-ED linearity for HyperSight closely matched simulation CT across soft-tissue and bone-equivalent materials, whereas Halcyon 3.1 exhibited HU bias and large variability. Dosimetrically, HyperSight accurately reproduced CT-based CI_Paddick and HI values, showing markedly reduced deviations relative to Halcyon 3.1. Clinical evaluation further showed that HyperSight identified volumetric and positional changes in both the whole breast and tumor bed, identifying regions susceptible to high-dose spill or under-coverage when no plan adaptation was performed. Discussion These findings suggest that HyperSight provides HU accuracy closer to that of simulation CT, improved image quality, and more accurate CBCT-based dose recalculation than Halcyon 3.1, thereby supporting the identification of clinically meaningful anatomical changes. HyperSight may therefore serve as a practical imaging option for offline ART in breast radiotherapy and may help reduce reliance on repeat simulation CT in selected cases while supporting adaptive treatment assessment.