A new plan quality objective function for determining optimal collimator combinations in prostate cancer treatment with stereotactic body radiation therapy using CyberKnife.

Stereotactic body radiation therapy with CyberKnife for prostate cancer has long treatment times compared with conventional radiotherapy. This arises the need for designing treatment plans with short execution times. We propose an objective function for plan quality evaluation, which was used to determine an optimal combination between small and large collimators based on short treatment times and clinically acceptable dose distributions. Data from 11 prostate cancer patients were used. For each patient, 20 plans were created based on all combinations between one small (⌀ 10-25 mm) and one large (⌀ 35-60 mm) Iris collimator size. The objective function was assigned to each combination as a penalty, such that plans with low penalties were considered superior. This function considered the achievement of dosimetric planning goals, tumor control probability, normal tissue complication probability, relative seriality parameter, and treatment time. Two methods were used to determine the optimal combination. First, we constructed heat maps representing the mean penalty values and standard deviations of the plans created for each collimator combination. The combination giving a plan with the smallest mean penalty and standard deviation was considered optimal. Second, we created two groups of superior plans: group A plans were selected by histogram analysis and group B plans were selected by choosing the plan with the lowest penalty from each patient. In both groups, the most used small and large collimators were assumed to represent the optimal combination. The optimal combinations obtained from the heat maps included the 25 mm as a small collimator, giving small/large collimator sizes of 25/35, 25/40, 25/50, and 25/60 mm. The superior-group analysis indicated that 25/50 mm was the optimal combination. The optimal Iris combination for prostate cancer treatment using CyberKnife was determined to be a collimator size between 25 mm (small) and 50 mm (large)