Investigation of Heuristic Checks for Treatment Beam Data
J Chang*, R Heaton, M Islam, Princess Margaret Hospital, Toronto, ONSU-E-T-238 Sunday 3:00PM - 6:00PM Room: Exhibit Hall
While radiotherapy treatment is accurately delivered in most cases, errors do occur with potentially fatal consequences. Many procedures have been developed over the years to minimize radiation therapy errors. However most of these procedures involve costly time consuming tasks and yet are not realistically fool-proof. Errors can still percolate through the process due to human errors or system malfunctions. We have investigated the development of an automated heuristic check of the beam data prior to transfer to the delivery system. The primary objective of this check is to identify any gross error without significantly impacting costs.
The total energy fluence in a treatment plan contributes mainly to the integral dose to the targeted volumes. To a first approximation, a simple correlation should exist between the integral dose to the target volume and the total energy fluence of the treatment beams. This correlation was investigated for clinical IMRT plans targeting the prostate and head & neck regions. Correlations between the summation of (Monitor Unit(MU)*Segment area) and integral CTV dose per fraction were investigated.
The sum, ΣMU*Segment area, used as a surrogate for the total energy fluence, was found to correlate with the integral dose, approximated as the CTV*prescribed dose. A linear fit of total energy fluence vs. integral dose for prostate and head & neck plans yielded a correlation coefficient of 0.99, with a slope of 3.3*10⁻⁵J*cm⁻²*MU⁻¹. Of the plans investigated, 90% of these were contained with in a region defined by ±5% of this slope.
A useful correlation between the ΣMU*Segment area and integral target dose was identified. This correlation can be used to automatically identify gross treatment errors and provide a low cost barrier to some types of catastrophic treatment errors.