Phantomless patient-specific TomoTherapy QA via delivery performance monitoring and a secondary Monte Carlo dose calculation

PURPOSE

To describe the validation and implementation of a novel quality assurance (QA) system for TomoTherapy using a Monte Carlo (MC)-based secondary dose calculation and CT detector-based multileaf collimator (MLC) leaf opening time measurement QA verification. This system is capable of detecting plan transfer and delivery errors and evaluating the dosimetric impact of those errors.

METHODS

The authors' QA process, MCLogQA, utilizes an independent pretreatment MC secondary dose calculation and postdelivery TomoTherapy exit detector-based MLC sinogram comparison and log file examination to confirm accurate dose calculation, accurate dose delivery, and to verify machine performance. MC radiation transport simulations are performed to estimate patient dose utilizing prestored treatment machine-specific phase-space information, the patient's planning CT, and MLC sinogram data. Sinogram data are generated from both the treatment planning system (MC_TPS) and from beam delivery log files (MC_Log). TomoTherapy treatment planning dose (DTPS) is compared with DMC_TPS and DMC_Log via dose-volume metrics and mean region of interest dose statistics. For validation, in-phantom ionization chamber dose measurements (DIC) for ten sample patient plans are compared with the computed values.

RESULTS

Dose comparisons to in-phantom ion chamber measurements validate the capability of the MCLogQA method to detect delivery errors. DMC_Log agreed with DIC within 1%, while DTPS values varied by 2%-5% compared to DIC. The authors demonstrated that TomoTherapy treatments can be vulnerable to MLC deviations and interfraction output variations during treatment delivery. Interfractional Linac output variations for each patient were approximately 2% and average output was 1%-1.5% below the gold standard. While average MLC leaf opening time error from patient to patient varied from -0.6% to 1.6%, the MLC leaf errors varied little between fractions for the same patient plan, excluding one patient.

CONCLUSIONS

MCLogQA is a new TomoTherapy QA process that validates the planned dose before delivery and analyzes the delivered dose using the treatment exit detector and log file data. The MCLogQA procedure is an effective and efficient alternative to traditional phantom-based TomoTherapy plan-specific QA because it allows for comprehensive 3D dose verification, accounts for tissue heterogeneity, uses patient CT density tables, reduces total QA time, and provides for a comprehensive QA methodology for each treatment fraction.