2148 - Intrafractional Prostate Motion Management with the Comfort Autoscan System

Purpose/Objective(s): Intrafractional motion of the prostate during radiotherapy poses a significant challenge for treatment accuracy. This study aims to introduce a Comfort with Monitoring (CwM) system, a dual-modality ultrasound (transperineal and transabdominal) approach integrated with camera-based motion tracking, to achieve non-invasive, real-time prostate localization with submillimeter precision during radiotherapy.

Materials/Methods: The CwM system integrates 4D wireless ultrasound (combining transperineal TPUS and transabdominal TAUS modes) and a camera-based tracker, utilizing data from 90 prostate cancer patients for validation. A hybrid registration framework correlates TAUS (global pelvic anatomy) and TPUS (local prostate details) through phantom-calibrated spatial alignment and deformable contour fusion, using anatomical landmarks (bladder neck, pubic symphysis) for real-time coordinate mapping. External motion captured by the VenusX BPS (Binocular positioning system) is synchronized with dual ultrasound data, while a multi-modal Kalman filter dynamically weights TAUS/TPUS inputs based on probe stability and body surface motion. Cross-modal consistency checks resolve signal conflicts, with automatic fallback to TAUS-driven bladder tracking during TPUS signal loss.

Results: This study establishes clinical-grade tracking accuracy (p<0.001, paired t-test) through rigorous validation against the real-time tracking system in the largest reported patient cohort (n=90) for multimodal tracking comparison. The predefined statistical power (90%) ensures reliability in detecting submillimeter discrepancies critical for hypo fractionated radiotherapy. Bonferroni-adjusted p-values addressed multiplicity in 12 anatomical landmark comparisons. The proposed tracking system demonstrated high concordance with the real-time electromagnetic tracking system (clinical gold standard) in a multi-center cohort of 90 prostate cancer patients. The mean 3D tracking error was 0.8 ± 0.3 mm (95% CI: 0.7–0.9 mm) during free-breathing conditions, with intraclass correlation coefficient (ICC) = 0.96 confirming excellent positional agreement. Bland-Altman analysis showed 95% limits of agreement within ±1.2 mm, meeting clinical tolerance thresholds for SBRT Statistical power analysis (ß=0.1, a=0.01) validated the sample size adequacy to detect =0.5 mm systematic deviations.

Conclusion: The proposed CwM system-based approach achieves submillimeter prostate localization within the accelerator’s coordinate system, compensating for both internal organ displacement and external patient movement without invasive markers. These findings overcome the limited sample size constraints (n=1–5) seen in early real-time tracking system validation studies, while stratified error analysis by respiratory phase provides new insights into motion compensation requirements.