Purpose: To evaluate the precision and accuracy of patient localization using a cone‐beam CT (CBCT) and an orthogonal x‐ray pair with infrared markers on the Novalis Tx treatment unit. Method and Materials: Twenty stereotactic radiosurgery/therapy patients with a total of twenty‐five lesions over ninety sessions (ranging from 1−5 fractions per patient) were localized daily using both ExacTrac (ETX, BrainLab) and on‐board CBCT (Varian) coupled to the Novalis Tx treatment unit. Each patient was first positioned using the ETX system accounting for variances in all six dimensions using a robotic couch top. Following these shifts, a CBCT was performed and further translations were made (x, y, z, table rotation) based on image fusion between the CBCT and simulation CT. A phantom study was also performed, mimicking the patient set‐up method to assess the reproducibility of each system and to determine any systematic differences between the ETX and CBCT localization approaches. Results: Patient positioning between ETX and CBCT was consistent in all four dimensions within 1.1mm and 0.1°. The average discrepancy between each system across all sessions was 1.1±1.2mm A/P, 1.0±1.2mm S/I, 0.1±1.4mm M/L, and 0.1°±0.5° couch rotation. Phantom testing showed that both systems were reproducible within 1.5mm and 0.5° in all dimensions. A systematic discrepancy of 0.3mm A/P, 1.2mm S/I, 0.8mm M/L, and 0.6° rotation was found between the two systems; however, this difference was deemed to be within the calibration tolerance of both systems. Conclusions: The ETX and on‐board CBCT systems were found to agree on tumor localization within 1.1mm (all dimensions) and 0.1° (couch rotation). Phantom studies showed the reproducibility of each system to be acceptable for stereotactic treatments. The Novalis Tx treatment unit incorporates both fiducial marker‐based, and volume‐based localization for reproducible and accurate SRS/SBRT patient treatments.
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging