An MLC calibration method using a detector array.

PURPOSE: The authors have developed a quantitative calibration method for a multileaf collimator (MLC) which measures individual leaf positions relative to the MLC backup jaw on an Elekta Synergy linear accelerator. METHODS: The method utilizes a commercially available two-axis detector array (Profiler 2; Sun Nuclear Corporation, Melbourne, FL). To calibrate the MLC bank, its backup jaw is positioned at the central axis and the opposing jaw is retracted to create a half-beam configuration. The position of the backup jaws field edge is then measured with the array to obtain what is termed the radiation defined reference line. The positions of the individual leaf ends relative to this reference line are then inferred by the detector response in the leaf end penumbra. Iteratively adjusting and remeasuring the leaf end positions to within specifications completes the calibration. Using the backup jaw as a reference for the leaf end positions is based on three assumptions: (1) The leading edge of an MLC leaf bank is parallel to its backup jaw's leading edge, (2) the backup jaw position is reproducible, and (3) the measured radiation field edge created by each leaf end is representative of that leaf's position. Data from an electronic portal imaging device (EPID) were used in a similar analysis to check the results obtained with the array. RESULTS: The relative leaf end positions measured with the array differed from those measured with the EPID by an average of 0.11+/-0.09 mm per leaf. The maximum leaf positional change measured with the Profiler 2 over a 3 month period was 0.51 mm. A leaf positional accuracy of +/-0.4 mm is easily attainable through the iterative calibration process. The method requires an average of 40 min to measure both leaf banks. CONCLUSIONS: This work demonstrates that the Profiler 2 is an effective tool for efficient and quantitative MLC quality assurance and calibration.

Multileaf collimator characteristics and reliability requirements for IMRT Elekta system.

Understanding the characteristics of a multileaf collimator (MLC) system, modeling MLC in a treatment planning system, and maintaining the mechanical accuracy of the linear accelerator gantry head system are important factors in the safe implementation of an intensity-modulated radiotherapy program. We review the characteristics of an Elekta MLC system, discuss the necessary MLC modeling parameters for a treatment planning system, and provide a novel method to establish an MLC leaf position quality assurance program. To perform quality assurance on 40 pairs of individual MLC leaves is a time-consuming and difficult task. In this report, an effective routine MLC quality assurance method based on the field edge of a backup jaw as referenced in conjunction with a diode array as a radiation detector system is discussed. The sensitivity of this test for determining the relative leaf positions was observed to be better than 0.1 mm. The Elekta MLC leaf position accuracy measured with this system has been better than 0.3 mm.

Comparison of analytic source models for head scatter factor calculation and planar dose calculation for IMRT.

The purpose of this study was to choose an appropriate head scatter source model for the fast and accurate independent planar dose calculation for intensity-modulated radiation therapy (IMRT) with MLC. The performance of three different head scatter source models regarding their ability to model head scatter and facilitate planar dose calculation was evaluated. A three-source model, a two-source model and a single-source model were compared in this study. In the planar dose calculation algorithm, in-air fluence distribution was derived from each of the head scatter source models while considering the combination of Jaw and MLC opening. Fluence perturbations due to tongue-and-groove effect, rounded leaf end and leaf transmission were taken into account explicitly. The dose distribution was calculated by convolving the in-air fluence distribution with an experimentally determined pencil-beam kernel. The results were compared with measurements using a diode array and passing rates with 2%/2 mm and 3%/3 mm criteria were reported. It was found that the two-source model achieved the best agreement on head scatter factor calculation. The three-source model and single-source model underestimated head scatter factors for certain symmetric rectangular fields and asymmetric fields, but similar good agreement could be achieved when monitor back scatter effect was incorporated explicitly. All the three source models resulted in comparable average passing rates (>97%) when the 3%/3 mm criterion was selected. The calculation with the single-source model and two-source model was slightly faster than the three-source model due to their simplicity.

Matching IMRT fields with static photon field in the treatment of head-and-neck cancer.

Radiation treatment with intensity-modulated radiation therapy (IMRT) for head-and-neck cancer usually involves treating the superior aspects of the target volume with intensity-modulated (IM) fields, and the inferior portion of the target volume (the low neck nodes) with a static anterior-posterior field (commonly known as the low anterior neck, or LAN field). A match line between the IM and the LAN fields is created with possibly large dose inhomogeneities, which are clinically undesirable. We propose a practical method to properly match these fields with minimal dependence on patient setup errors. The method requires mono-isocentric setup of the IM and LAN fields with half-beam blocks as defined by the asymmetric jaws. The inferior jaws of the IM fields, which extend approximately 1 cm inferiorly past the isocenter, are changed manually before patient treatment, so that they match the superior jaw of the LAN field at the isocenter. The matching of these fields therefore does not depend on the particular treatment plan of IMRT and depends only on the matching of the asymmetric jaws. Measurements in solid water phantom were performed to verify the field-matching technique. Dose inhomogeneities of less than 5% were obtained in the match-line region. Feathering of the match line is done twice during the course of a treatment by changing the matching jaw positions superiorly at 3-mm increments each time, which further reduces the dose inhomogeneity. Compared to the method of including the lower neck nodes in the IMRT fields, the field-matching technique increases the delivery efficiency and significantly reduces the total treatment time.

An immobilization system for claustrophobic patients in head-and-neck intensity-modulated radiation therapy.

PURPOSE: To evaluate the effectiveness of an immobilization treatment system used for claustrophobic patients in head-and-neck intensity-modulated radiation therapy (IMRT). METHODS AND MATERIALS: Instead of the thermoplastic facemask, the Vac Fix (S & S Par Scientific, Odense, Denmark) mold is used for immobilization of claustrophobic patients at the University of Florida in head-and-neck IMRT. The immobilization procedure combines the use of commercial stereotactic infrared (IR) ExacTrac camera system (BrainLAB, Inc., Westchester, IL) for patient setup and monitoring. The Vac Fix mold is placed on the headrest and folded up as needed to provide support before the mold is hardened. For the camera system, a frame referred to as a "tattoo-free immobilization accessory" is fabricated, on which the IR markers can be placed. A patient-specific dental impression is made with the bite tray. The movement of the markers, connected through the dental impression of the patient, accurately represents the overall patient motion. Patient movement is continuously monitored and repositioning is performed whenever patient movement exceeds the predefined tolerance limit. Monitored patient movements are recorded at a certain frequency. Recorded data are analyzed and compared with those of patients immobilized with the thermoplastic facemask plus the camera system that is the standard immobilization system in our clinic. RESULTS: For three patients treated with the Vac Fix mold plus the camera system, on average, the histogram-based uncertainties, U(95)(5), U(95)(20), and mean displacement, R(mean) (mm) were 1.03, 1.08, and 0.60, respectively. These values are close to those obtained with the mask plus the camera system. The Vac Fix mold plus the camera system often requires more beam interruptions because of repositioning than the mask plus the camera system (on average, the Vac Fix mold plus the camera system required repositioning 7.7 times and the mask plus the camera system required repositioning 1.8 times during 20 treatments). CONCLUSION: The Vac Fix mold immobilization procedure plus the camera monitoring system has been set up for patients who are claustrophobic or cannot tolerate a mask during head-and-neck IMRT. Although this system causes more frequent beam delivery interruptions, it is as effective as the mask plus the camera system in immobilizing patients within the tolerance limit.

Generalized monitor unit calculation for the Varian enhanced dynamic wedge field.

The generalized monitor unit (MU) calculation equation for the Varian enhanced dynamic wedge (EDW) is derived. The assumption of this MU calculation method is that the wedge factor of the EDW at the center of the field is a function of field size, the position of the center of the field in the wedge direction, and the final position of the moving jaw. The wedge factors at the center of the field in both symmetric and asymmetric fields are examined. The difference between calculated and measured wedge factors is within 1.0%. The method developed here is easy to implement. The only datum required in addition to the standard set of conventional physical wedge implementation data is the off-axis output factor for the open field in the reference condition. The off-center point calculation is also examined. For the off-center point calculation, the dose profile in the wedge direction for the largest EDW field is used to obtain the relative off-center ratio in any smaller wedge field. The accuracy of the off-center point calculation decreases when the point of calculation is too close to the field edge.