PHYSICS CONTRIBUTION DIGITALTOMOSYNTHESIS FOR RESPIRATORY GATED LIVER TREATMENT: CLINICAL FEASIBILITY FOR DAILY IMAGE GUIDANCE Q. JACKIE WU,PH.D., JEFFREY MEYER, M.D., JESSICA FULLER, M.S., DEVON GODFREY,PH.D., ZHIHENG WANG,PH.D., JUNAN ZHANG,PH.D., AND FANG-FANG YIN,PH.D. Department of Radiation Oncology, Duke University Medical Center, Durham, NC Purpose: Breath-hold (BH) treatment minimizes internal target volumes (ITV) when treating sites prone to mo- tion. Digital tomosynthesis (DTS) imaging has advantages over cone-beam CT (CBCT) for BH imaging: BH- DTS scan can be completed during a single breath-hold, whereas BH-CBCT is usually acquired by parsing the gan- try rotation into multiple BH segments. This study evaluates the localization accuracy of DTS for BH treatment of liver tumors. Methods: Both planning CT and on-board DTS/CBCT images were acquired under BH, using the planning CT BH window as reference. Onboard imaging data sets included two independent DTS orientations (coronal and sagit- tal), and CBCT images. Soft tissue target positioning was measured by each imaging modality and translated into couch shifts. Performance of the two DTS orientations was evaluated by comparing target positioning with the CBCT benchmark, determined by two observers. Results: Image data sets were collected from thirty-eight treatment fractions (14 patients). Mean differences be- tween the two DTS methods and the CBCT method were <1 mm in all directions (except the lateral direction with sagittal-DTS: 1.2 mm); the standard deviation was in the range of 2.1–3.5 mm for all techniques. The Pearson correlation showed good interobserver agreement for the coronal-DTS (0.72–0.78). The interobserver agreement for the sagittal-DTS was good for the in-plane directions (0.70–0.82), but poor in the out-of-plane direction (lateral, 0.26). Conclusions: BH-DTS may be a simpler alternative to BH-CBCT for onboard soft tissue localization of the liver, although the precision of DTS localization appears to be somewhat lower because of the presence of subtle out- of-plane blur. Ó 2011 Elsevier Inc. Digital tomosynthesis, Image guidance, Respiratory motion, Breath hold, liver. INTRODUCTION One of the challenges in irradiating sites such as the liver is organ motion. Studies have shown that liver motion associ- ated with breathing can be several centimeters (1–5). To minimize the motion, treatment using abdominal compression, under deep inspiration breath-hold (BH) or ac- tive breathing control (ABC) has been used (6–16). These motion-management techniques can reduce large planning target volumes required to account for liver motion, allowing higher doses to be delivered without increasing normal liver toxicity (9, 17). The reproducibility of organ position using ABC or BH has been studied for lung and liver cancer patients (7, 10, 11, 18). Dawson et al. (7) reported 2.5-mm intrafraction re- producibility and 4.4-mm interfraction reproducibility of liver position relative to the vertebral bodies for six patients using daily orthogonal kilovoltage (kV) imaging. Eccles et al. (18) reported the interfraction reproducibility ranges from 1.5 to 7.7 mm in the superior–inferior direction, with average difference of diaphragm positions relative to verte- bral body position per patient in the range of 0.1–12.0 mm. Kimura et al. (11) also found that the BH reproducibility was 4.0  3.5 mm (intrafraction) and 5.1  4.8 mm at the end-inspiration phase (interfraction). Similarly, Kim et al. (10) found the standard deviation of diaphragmatic position for BH ranged from 0.13 to 2.57 mm, with an average of 0.97 mm. These findings indicate a change in diaphragm po- sition from day to day despite using ABC or BH in the same planned position. Therefore, quick daily imaging is preferred to verify target position before treatment (6, 17, 19). For ABC or deep-inspiration BH treatment, the ideal local- ization imaging technology should have fast acquisition (i.e., within about 20 s or single BH equivalent) and provide suf- ficient three-dimensional (3D) visualization of soft tissues. Reprint requests to: Q. Jackie Wu, Ph.D., Box 3295, Department of Radiation Oncology, Duke University Medical Center, Durham, NC 27710. Tel: (919) 681-1841; Fax: (919) 681-7183; E-mail: Jackie.Wu@duke.edu The research is in part supported by research grants from National Institutes of Health (Grant No. R21 CA128368), General Electric Medical Systems, and Varian Medical Systems. Conflict of interest: none. Received May 14, 2009, and in revised form Jan 29, 2010. Accepted for publication Jan 31, 2010. 289 Int. J. Radiation Oncology Biol. Phys., Vol. 79, No. 1, pp. 289–296, 2011 Copyright Ó 2011 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/$–see front matter doi:10.1016/j.ijrobp.2010.01.047