1157 - Evaluation of Single-Energy (SE) and Synthetic Dual-Energy (sDE) Imaging to Validate Tumor Position for Respiratory Gating Using 4D-CBCT Projection Data
Presenter(s)
H. Nguyen1, J. Luce1, Y. Abdelal2, A. Keeler1, L. Zhu3, M. Lehmann4, H. Kang1, M. M. Harkenrider5, and J. C. Roeske1; 1Department of Radiation Oncology, Stritch School of Medicine, Cardinal Bernardin Cancer Center, Loyola University Chicago, Maywood, IL, 2Loyola University Medical Center, Maywood, IL, United States, 3Varian Medical Systems, Palo Alto, CA, 4Varian Medical Systems, Baden-Dattwil, Switzerland, 5Loyola University Medical Center, Maywood, IL
Purpose/Objective(s): Respiratory gating is commonly used to minimize radiation exposure to normal tissues during lung cancer treatment, relying on an external surrogate signal that assumes a correlation with tumor position. This study aims to validate tumor position by evaluating markerless tumor tracking (MTT) using single-energy (SE) and synthetic dual-energy (sDE) images generated from 4D-CBCT projection data.
Materials/Methods: A 4D-CBCT scan (125 kVp) was performed on a dynamic thorax phantom containing a 15-mm-diameter simulated tumor, using the onboard imager (OBI) of a commercial linear accelerator. A marker block was placed on a platform that moved in sync with the lung’s longitudinal position, monitored by an infrared camera, and used as a motion surrogate. After data acquisition, sDE images were generated from SE projection data using a previously trained U-net neural network, which removes the bony structures to enhance tumor visibility. A template matching algorithm tracked tumor motion in both SE and sDE sequences. The images were subsequently sorted into 10 respiratory phases based on the external surrogate. For each phase, the median tumor location was determined and used to calculate the tracking success rate (TSR)—the percentage of images in which the tracked location was within ±2 mm from the median tumor location. Missing frame rate (MFR), defined as the percentage of images where tracking failed, was also calculated. TSR and MFR were used to compare SE and sDE performance.
Results: The MFR was similar for both SE and sDE (0.9% vs. 1.7%, p = 0.38). The TSR was comparable between SE and sDE in the 0-20% and 70-90% phases, with average TSR of 53.9% ± 17.4% and 56.7% ± 17.1%, respectively (p = 0.61). However, for the 30-60% phases which are commonly used in respiratory gating, the TSR was significantly higher for sDE (86.3% ± 2.3%) compared to SE (72.6% ± 5.3%, p < 0.05).
Conclusion: This study demonstrates that 4D-CBCT projection data can be used to validate tumor position during respiratory gating. The use of sDE imaging enhances tumor tracking performance by improving tumor visibility through synthetic bone removal, particularly during the 30-60% respiratory phases, supporting the potential for more accurate respiratory gating during treatment.