Simulated online adaptive radiotherapy on C-arm linacs: Integrating high resolution cone-beam CT into abdominal stereotactic body radiation therapy.

BACKGROUND

Conventional cone-beam computed tomography (CBCT) systems, while providing adequate visualization for positioning, have historically been limited by image truncation, scatter contamination, image noise, geometric artifacts, and unreliable CT number accuracy, which preclude their routine use for accurate dose calculation and re-planning. Recent advances in CBCT detector technology and image reconstruction algorithms have led to the development of next-generation CBCT platforms with improved image quality and CT number stability.

PURPOSE: This study aimed to develop and evaluate an efficient daily adaptation workflow on a standard C-arm linear accelerator by integrating HyperSight CBCT and plan adaptation in RayStation.

METHODS

Image quality metrics and CT number accuracy were evaluated and compared with those of an institutional CT simulator and conventional CBCT. Dose calculation accuracy was validated through measurement. The HyperSight CBCT adaptive workflow used both deformable image registration and rigid image registration for contour propagation, followed by manual revision when necessary. The adapted plan was iteratively optimized in RayStation by using the fine-tune module. End-to-end test was done with a spine phantom and a pelvis phantom. Retrospective evaluation of four abdominal SBRT cases (12 fractions) compared estimated delivered and adapted plans with respect to target coverage, organs at risk (OARs) sparing, and monitor unit (MU) usage.

RESULTS

Improved image quality and CT number accuracy compared to conventional CBCT were observed. The adaptation process required 13 ± 3 min per fraction from HyperSight-CBCT acquisition to final plan evaluation, excluding time for contour revisions. Across all cases, adapted plans demonstrated improved dosimetric quality relative to delivered plans. Gross tumor volume V100% and planning target volume V100% were consistently higher in the adapted plans, with an average increase of 6.3% (P = 0.004). Doses to OARs, including the small bowel and ipsilateral kidney, were reduced or maintained. Adapted plan MU usage was comparable to reference plans, indicating no increase in treatment complexity.

CONCLUSIONS

Integrating HyperSight CBCT with RayStation's streamlined plan adaptation capability provided dosimetric advantages. It improved target coverage and OAR sparing without increasing treatment complexity and represents a practical alternative to Ethos or MR-Linac systems without the need for specialized delivery platforms.