Comparison Of Dose Distributions Calculated By Monte Carlo And Ray-Tracing Algorithms For The Treatment Of Lung Tumors With CyberKnife

Lung cancer is one of the most common cancers in the world, and some subtype of it, such as the early-stage non-small cell lung cancer(NSCLC), was not suitable for surgery therapy. Stereotactic ablative radiotherapy(SBAR) is another choice for the radical treatment to NSCLC besides the surgery. The Xsight? Lung Tracking System of CyberKnife could make real-time image guiding and dynamic tumor tracking without gold-marker implantation during the normal respiratory cycles. The accuracy of this treatment can be sub-millimeter scales. The dose of SBAR is decided by the treatment planning system(TPS). Ray-Tracing algorithm is equipped as standard in TPS of CyberKnife, while the Monte Carlo algorithm is just option for some machine. Ray-Tracing algorithm merely account for decreased attenuation of the primary photon beam in low density lung tissue, while increased electron range is not accounted for. Monte Carlo algorithm on the other hand explicitly accounts for these effects. It is the most accurate dose calculation algorithm currently available. The present study compared the dose distributions in CyberKnife treatment for peripheral lung tumors that were calculated by the Monte Carlo algorithm and by the Ray-Tracing algorithm respectively, and gave some guidelines in the stereotactic ablative radiotherapy for peripheral lung tumors to the users of CyberKnife without Monte Carlo algorithm equipment.For 70 patients with peripheral NSCLC, Cyber Knife treatment planning system with Monte Carlo and Ray-Tracing were preformed respectively with high resolution. The difference of dose distribution to target and organ at risk(OAR) were compared. The results showed the consistently overestimation of exposure does to tumor in Ray-Tracing algorithm group. In detail, compared with the results from Monte Carlo algorithm group, the planning target volume(PTV) D95 and gross tumor volume(GTV) D99 were overestimated by 8.59 Gy and 5.84 Gy in Ray-Tracing algorithm group. The homogeneity index(HI), conformity index(CI), and new conformity index(nCI) of PTV were overestimated by 0.11, 0.05, and 0.05 in Ray-Tracing algorithm group. The volume dose of spinal cord, esophagus, trachea, bronchus and heart were overestimated by 149%, 80%, 72%, 63%, 54% and 99% respectively. The mean dose and V5 dose volume of lung tissue were overestimated by 2.69 Gy and 7.52% higher respectively. Even if the prescription dose of Monte Carlo were rescale to 60 Gy, dose of OAR still lower than the result of Ray-Tracing algorithm.Seven parameters were concerned to analyze the difference of these two algorithms, including the PTV volume, the GTV volume, the TTV volume, the total monitor units, the number of beams, the horizontal range and vertical dimension between the tumor and the chest wall. The linear regression and nonlinear generalize additive model were used in the analysis. The results indicated that only the TTV volume could affect the PTV D95. The fitting function between the PTV D95 and the TTV volume is power function, and the R2 is 0.8626. The present study gave some guidelines for peripheral lung tumors radiotherapy to the users of CyberKnife without Monte Carlo algorithm.