The Feasibility Of Evaluating Radiation Dose To The Heart By Integrating Kilovoltage-cone Beam Computed Tomography

OBJECTIVE:To research the feasibility of contouring the planning risk organ volume (PRV) for theheart, and to determinethe achievability of evaluating radiation dosage to the heart using kilovoltage-cone beam computed tomography (KV-CBCT) in thoracic patients.METHODS:40thoracic patients between November2010and December2012were enrolled in thisstudy. The patient population included nine males and eight females. Each patient wasimmobilized by a vacuum bag, and a pre-treatment computed tomography (CT) scan at a slicethickness of3mm was performed. The images were transferred to an Eclipse radiotherapytreatment planning system (TPS, version8.6, Varian Medical Systems, Palo Alto, CA, USA)for contouring and planning, according to ICRU recommendations. The Institutional ReviewBoard approved this study, and each patient was given informed consent.CBCT image acquisitionCBCT images were acquired in the treatment room using a scanner attached to the gantry of the linear Varian Trilogy accelerator (Varian Medical Systems). Ten series of CBCT imageswere acquired to verify setup accuracy. CBCT acquisition time was75~90s. The rest of theacquisition parameters were120kVp, with various exposures ranging from0.16to0.64mAsper frame.Contouring the heart on CBCT imagesThree seasoned physicians, a radiologist, radiation therapist, and physicist, wereinvolved in contouring the heart on the CBCT image, according to a uniform standard. Thecontouring criteria were as follows:1) based on anatomical landmarks, the boundary betweenthe cardiac posterior and the esophagus was profiled according to the thickness of theesophagus that filled with gas;2) the boundaries between the cardiac anterior edge, and thesternum and walls of the chest were distinguished by their continuities of original contours,and the CT value;3) the boundary between the cardiac superior edge and aorta started fromthe bifurcation layer of the pulmonary artery. Then, the boundary between the lower edge ofthe heart and liver was generated through the dropping method of CT value. We firstmeasured the CT values of several spots at the junction between the liver and heart (spotswere chosen based on experience on the edge of upper heart, and on the edge of the junctionsbetween the lower liver and heart), and then marked those spots with the largest CT value.Finally, we connected these spots to profile the border of the heart and liver.Calculating heart volume, heart coincidence, and mean heart margins from planning CTand CBCT and dose-volume indicesThe heart volumes from both CBCT and CT images were determined using Eclipse TPS(Varian). The coinciding regions of the hearts on the same anatomical section between thefirst series of CBCT images and other images were computed using MATLAB2012a software.The mean heart margins from planning CT and CBCT images were obtained afterrectification of setup errors for both CBCT and CT scans. Dose-volume indices were matchedby two-way scanning of the dose volume histogram (DVH). Relative dose-volume indices(VxGy-R) were: V5Gy-R, V10Gy-R, V15Gy-R, V20Gy-R, V25Gy-R, V30Gy-R,and V35Gy-R. Absolute dose-volume indices (VxGy-A) were: V5Gy-A, V10Gy-A, V15Gy-A, V20Gy-A, V25Gy-A, V30Gy-A, andV35Gy-A. Dose index for fixed heart volume was: D50cm3, D100cm3, D150cm3, D200cm3, D250cm3, andDmean(mean dose), Dmax(exposure dose per1cm3).RESULTS:Heart volumeThis summarizes the reproducibility of heart volume obtained from different CBCTscans, as well as a comparison of heart volumes between CBCT and CT images. Heartvolumes for each patient were largely similar among CBCT images (F=1.00, p=0.44), andwere larger than those from planning CT scans. The heart volume between the first CBCT andCT images was significantly different (t=4.63, p=0.001).Heart reproducibility from a series of different CBCT imagesThe reproducibility of heart outlines from different series of CBCT images of the sameanatomical section (about40sections) reached up to0.985±0.020, and the margins of errorwere not significantly different between each set of images (p>0.05).The external margins in different axial directionsThe margins in different axial directions from the heart on planning CT and CBCTimages were10.5±2.8mm in the left direction,5.9±2.8mm in the right direction,2.2±1.6mm in the direction of the head,3.3±2.2mm in the direction of the foot,6.7±1.1mm in theanterior direction, and4.5mm±2.5mm less in the posterior direction.Dose-volume indicesRelative and absolute dose-volume indices from CBCT images were all significantlylarger than those from CT images, with the exception of V5Gy-R(p<0.05). In addition, thedoses in the fixed heart volumes for CBCT were all significantly larger than those for CT(p<0.05).CONCLUSIONS:The PRV of heart contouring based on KV-CBCT is feasible with good reproducibility. Thus, more accurate and objective dose-volume indices could be obtained from applyingKV-CBCT in thoracic patients.