Encrypted login | home

Program Information

Effect of Statistic Uncertainty On DVH and Dose Mass Histogram (DMH) Derived Dose Indices (DI) for Head and Neck IMRT-Monte Carlo(MC) Treatments

G Mora

G Mora1*, J Li2, J Fan3, C Ma4, (1) Lisbon University, Lisbon (2,3,4) Fox Chase Cancer Center, Philadelphia, PA

SU-E-T-518 Sunday 3:00PM - 6:00PM Room: Exhibit Hall

The use of DMH as an alternative method to validate the treatment plans of the head and neck region has been investigated in the past[1].For patient geometry contains large air cavities,high statistical uncertainties were found in dose distributions calculated using MC method in air cavities and tissues that surround air cavities.These previous studies reported that the shape of DMH is less affected by the statistical uncertainty than DVH. But the authors did not verify the differences between DIs derived from DVH and DMH.In the present work we compare the DIs derived from DVHs and DMHs based on dose distributions calculated using MC method and different levels of statistical precision.

Six IMRT plans were calculated using the EGS4 based MCSIM code.MCSHOW allows to built DVHs and DMHs from dose distributions calculated with different statistical uncertainty(from 0.5% to 10%).
The DIs were derived from the DVHs and DMHs for the target and critical structures (cord, chiasm, larynx,brainstem and parotid).

Comparing the values (averaged over 6 patients) of target DMH (0.5%) and DMH (2%) we found that all DIs (D05, D95, Mp, Dmax) do not differ in more than 0.7%. By comparison of target DVH (0.5%) and target DVH (2%) we observed differences of about 14% on the value of Dmax and V95 changes in more than 3%. We verified that the statistical effect is not significant (less than 1%) for the DVH of critical structures.

The DIs derived from DMHs are not strong dependent on the statistical uncertainty of the dose values as those derived from DVHs. Our results confirmed the previous idea that DMH is a better parameter than DVH for evaluating treatment plans calculated by MC simulations in low density regions.

[1]-G.Mora et al., MCTP 2009,Cardiff 2009.

Contact Email: