• Dennis M Duggan, George X Ding, Charles W Coffey, Wyndee Kirby, Dennis E Hallahan, Arnold Malcolm, and Bo Lu.
• Vanderbilt University, Department of Radiation Oncology, B1003 Preston Research Building, Nashville, TN 37232-5671, United States. Dennis.Duggan@Vanderbilt.edu
• Lung Cancer. 2007 Apr 1; 56 (1): 77-88.

AbstractWe report our initial experience with deep-inspiration breath-hold (DIBH) cone-beam CT (CBCT) on the treatment table, using the kilovoltage imager integrated into our linear accelerator, for setting up patients for DIBH stereotactic body radiation therapy (SBRT) for lung tumors. Nine patients with non-small cell lung cancer (seven stage I), were given 60Gy in three fractions. All nine patients could perform a DIBH for 35s. For each patient we used a diagnostic reference CT volume image acquired during a DIBH to design an SBRT plan consisting of 7-10 noncoplanar conformal beams. Four patients were setup by registering DIBH kilovoltage projection radiographs or megavoltage portal images on the treatment table to digitally reconstructed radiographs from the reference CT. Each of the last 14 fractions out of a total of 27 was setup by acquiring a CBCT volume image on the treatment table in three breath-holds. The CBCT and reference CT volume images were directly registered and the shift was calculated from the registration. The CBCT volume images contained excellent detail on soft tissue and bony anatomy for matching to the reference CT. Most importantly, the tumor was always clearly visible in the CBCT images, even when it was difficult or impossible to see in the radiographs or portal images. The accuracy of the CBCT method was confirmed by DIBH megavoltage portal imaging and each treatment beam was delivered during a DIBH. CBCT acquisition typically required five more minutes than radiograph acquisition but the overall setup time was often shorter using CBCT because repeat imaging was minimized. We conclude that for setting up SBRT treatments of lung tumors, DIBH CBCT is feasible, fast and may result in less variation among observers than using bony anatomy in orthogonal radiographs.

22 keywords...

hide…