Quality Assurance of 4D-CT Scan Techniques in Multicenter Phase III Trial of Surgery Versus Stereotactic Radiotherapy (Radiosurgery or Surgery for Operable Early Stage (Stage 1A) Non-Small-Cell Lung Cancer [ROSEL] Study)

Purpose: To determine the accuracy of four-dimensional computed tomography (4D-CT) scanning techniques in institutions participating in a Phase III trial of surgery vs. stereotactic radiotherapy (SBRT) for lung cancer. Methods and Materials: All 9 centers performed a 4D-CT scan of a motion phantom (Quasar, Modus Medical Devices) in accordance with their in-house imaging protocol for SBRT. A cylindrical cedar wood insert with plastic spheres of 15 mm (ø15) and 30 mm (ø30) diameter was moved in a cosine-based pattern, with an extended period in the exhale position to mimic the actual breathing motion. A range of motion of R = 15 and R = 25 mm and breathing period of T = 3 and T = 6 s were used. Positional and volumetric imaging accuracy was analyzed using Pinnacle version 8.1× at various breathing phases, including the mid-ventilation phase and maximal intensity projections of the spheres. Results: Imaging using eight CT scanners (Philips, Siemens, GE) and one positron emission tomography-CT scanner (Institution 3, Siemens) was investigated. The imaging protocols varied widely among the institutions. No strong correlation was found between the specific scan protocol parameters and the observed results. Deviations in the maximal intensity projection volumes averaged 1.9% (starting phase of the breathing cycle [ø]15, R = 15), 12.3% (ø15, R = 25), and -0.9% (ø30, R = 15). The end-expiration volume deviations (13.4%, ø15 and 2.5%, ø30), were, on average, smaller than the end-inspiration deviations (20.7%, ø15 and 4.5%, ø30), which, in turn, were smaller than the mid-ventilation deviations (32.6%, ø15 and 8.0%, ø30). A slightly larger variation in the mid-ventilation origin position was observed (mean, -0.2 mm; range, -3.6-4.2) than in the maximal intensity projection origin position (mean, -0.1 mm; range, -2.5-2.5). The range of motion was generally underestimated (mean, -1.5 mm; range, -5.5-1). Conclusions: Notable differences were seen in the 4D-CT imaging protocols for SBRT among centers. However, the observed deviations in target volumes were generally small. They were slightly larger for the mid-ventilation phases and smallest for the end-expiration phases. Steps to optimize and standardize the 4D-CT scanning protocols for SBRT are desirable.