A Comparative Dosimetric Analysis of the Effect of Heterogeneity Corrections Used in Three Treatment Planning Algorithms

Successful treatment in radiation oncology relies on the evaluation of a plan for each individual patient based on delivering the maximum dose to the tumor while sparing the surrounding normal tissue (organs at risk) in the patient. Organs at risk (OAR) typically considered include the heart, the spinal cord, healthy lung tissue, and any other organ in the vicinity of the target that is not affected by the disease being treated. Depending on the location of the tumor and its proximity to these OARs, several plans may be created and evaluated in order to assess which "solution" most closely meets all of the specified criteria. In order to successfully review a treatment plan and take the correct course of action, a physician needs to rely on the computer model (treatment planning algorithm) of dose distribution to reconstruct CT scan data to proceed with the plan that best achieves all of the goals.;There are many available treatment planning systems from which a Radiation Oncology center can choose from. While the radiation interactions considered are identical among clinics, the way the chosen algorithm handles these interactions can vary immensely. The goal of this study was to provide a comparison between two commonly used treatment planning systems (Pinnacle and Eclipse) and their associated dose calculation algorithms. In order to this, heterogeneity correction models were evaluated via test plans, and the effects of going from heterogeneity uncorrected patient representation to a heterogeneity correction representation were studied.;The results of this study indicate that the actual dose delivered to the patient varies greatly between treatment planning algorithms in areas of low density tissue such as in the lungs. Although treatment planning algorithms are attempting to come to the same result with heterogeneity corrections, the reality is that the results depend strongly on the algorithm used in the situations studied. While the Anisotropic Analytic Method (AAA) and Collapsed Cone Convolution (CCC) have been shown to be much better than earlier heterogeneity correction algorithms (such as modified Batho), they still fail to agree to a significant degree to ensure agreement between the recommended planning and delivery of better than 2% (Loevinger & Loftus, 1977).;For lung plans, average minimum PTV dose was seen to be as much as 285 cGy (4.3% of the prescription dose) less than the same plan calculated without heterogeneity corrections applied. Average maximum PTV dose was seen to be as much as 680 cGy (10.3% of the prescription dose) higher than the same plan calculated with heterogeneity corrections turned off. When different heterogeneity correction algorithms are compared, average differences of up to 747 cGy in minimum PTV dose (11.3% of the prescription dose) can be seen. These results highlight the importance of careful considerations of the limitations of treatment planning algorithms under certain conditions of clinical use.