Encrypted login | home

Program Information

Central Moments to Analyze Dose Distributions in Planning Target Volumes

no image available
Q Zhang

Q Zhang*, Y Cao , A Jamshidi , P Vigneri , l Potters , E Klein , Long Island Jewish Medical Center, Lake Success, NY


SU-F-T-602 (Sunday, July 31, 2016) 3:00 PM - 6:00 PM Room: Exhibit Hall


To quantify the dose distribution inside planning target volumes (PTV), the coefficient of variation, skewness (third central moment) and kurtosis (fourth central moment) were used to analyze the dose distributions of PTVs for lung and prostate cases. Those variables can be used to measure the shape of the dose distribution.


The mean dose, coefficient of variation, skewness, and kurtosis were used to analyze the dose distributions inside PTVs for 24 randomly selected patients. The bin width of the relative dose distribution is 0.1 percent. The coefficient of variation (ratio of the standard deviation to the mean), skewness (third standardized moment), kurtosis (fourth standardized moment) were calculated by using the calculated dose distribution. The coefficient of variation measures the ratio of noise to signal. Skewness measures the asymmetry of the distribution, and negative values indicate that data is skewed left (more low dose). Kurtosis measures the peak of the distribution, where a value of 3 is for a standard normal distribution.


The average mean dose was found to be (101.17±5.26)% of the prescribed dose. The averaged standard deviation is (2.07±0.68)%. The average coefficient of variation is 0.020±0.007. The average skewness is -1.17±0.50, with range from -2.12 to -0.24. Because the values of the skewness are negative for all patients, that means more voxels were in lower dose regions. The average kurtosis is 5.89±2.29 with a range from 3.58 to 12.06. The dose distributions are wider than standard normal distributions.

The coefficient of variation, skewness and kurtosis were used to analyze PTV dose distributions. It was found that these variables can be used to quantify the dose distributions inside PTVs, and maybe a preferred way to assess PTV coverage

Contact Email: