Quantifying uncertainties associated with reference dosimetry in an MR-Linac.

BACKGROUND: Magnetic resonance (MR)-guided radiation therapy provides capabilities to utilize high-resolution and real-time MR imaging before and during treatment, which is critical for adaptive radiotherapy. This emerging modality has been promptly adopted in the clinic settings in advance of adaptations to reference dosimetry formalism that are needed to account for the presence of strong magnetic fields. In particular, the influence of magnetic field on the uncertainty of parameters in the reference dosimetry equation needs to be determined in order to fully characterize the uncertainty budget for reference dosimetry in MR-guided radiation therapy systems. PURPOSE: To identify and quantify key sources of uncertainty in the reference dosimetry of external high energy radiotherapy beams in the presence of a strong magnetic field. METHODS: In the absence of a formalized Task Group report for reference dosimetry in MR-integrated linacs, the currently suggested formalism follows the TG-51 protocol with the addition of a quality conversion factor kBQ accounting for the effects of the magnetic field on ionization chamber response. In this work, we quantify various sources of uncertainty that impact each of the parameters in the formalism, and evaluate their overall contribution to the final dose. Measurements are done in a 1.5 T MR-Linac (Unity, Elekta AB, Stockholm, Sweden) which integrates a 1.5 T Philips MR scanner and a 7 MVFFF linac. The responses of several reference-class small volume ionization chambers (Exradin:A1SL, IBA:CC13, PTW:Semiflex-3D) and Farmer type ionization chambers (Exradin:A19, IBA:FC65-G) were evaluated throughout this process. Long-term reproducibility and stability of beam quality, TPR 10 20 ${\mathrm{TPR}}_{10}^{20}$ , was also measured with an in-house built phantom. RESULTS: Relative to the conventional external high energy linacs, the uncertainty on overall reference dose in MR-linac is more significantly affected by the chamber setup: A translational displacement along y-axis of ± 3 mm results in dose variation of < |0.20| ± 0.02% (k = 1), while rotation of ± 5° in horizontal and vertical parallel planes relative to relative to the direction of magnetic field, did not exceed variation of < |0.44| ± 0.02% for all 5 ionization chambers. We measured a larger dose variation for xy-plane (horizontal) rotations (< |0.44| ± 0.02% (k = 1)) than for yz-plane (vertical) rotations (< ||0.28| ± 0.02% (k = 1)), which we associate with the gradient of kB,Q as a function of chamber orientation with respect to direction of the B0 -field. Uncertainty in Pion (for two depths), Ppol (with various sub-studies including effects of cable length, cable looping in the MRgRT bore, connector type in magnetic environment), and Prp were determined. Combined conversion factor kQ × kB,Q was provided for two reference depths at four cardinal angle orientations. Over a two-year period, beam quality was quite stable with TPR 10 20 ${\mathrm{TPR}}_{10}^{20}$ being 0.669 ± 0.01%. The actual magnitude of TPR 10 20 ${\mathrm{TPR}}_{10}^{20}$ was measured using identical equipment and compared between two different Elekta Unity MR-Linacs with results agreeing to within 0.21%. CONCLUSION: In this work, the uncertainty of a number of parameters influencing reference dosimetry was quantified. The results of this work can be used to identify best practice guidelines for reference dosimetry in the presence of magnetic fields, and to evaluate an uncertainty budget for future reference dosimetry protocols for MR-linac.

Water calorimetry in MR-linac: Direct measurement of absorbed dose and determination of chamber k Q mag.

PURPOSE: To use a portable 4°C cooled MR-compatible water calorimeter to measure absorbed dose in a magnetic resonance-guided radiation therapy (MRgRT) system. Furthermore, to use the calorimetric dose results and direct cross-calibration to experimentally measure the combined beam quality and magnetic field correction factor ( k Q mag ) of a clinically used reference-class ionization chamber placed under the same radiation field. METHODS: An Elekta Unity MR-linac (7 MV FFF, B = 1.5 T) was used in this study. Measurements were taken using the in-house designed and built water calorimeter. Following preparation and cooling of the system, the MR-compatible calorimeter was positioned using a combination of MR and EPID imaging and the dose to water was measured by monitoring the radiation-induced temperature change. Immediately after the calorimetric measurements, an A1SL ionization chamber was placed inside the calorimeter for direct cross-calibration. The results allowed for a direct and absolute experimental measurement of k Q mag for this chamber and comparison against existing Monte Carlo values. RESULTS: The calorimeter was successfully positioned using imaging in under an hour. The 1-hour setup time is from the time the calorimeter leaves storage to the first calorimetric measurement. Absorbed dose was successfully measured with a relative combined standard uncertainty of 0.71 % (k = 1). Through a cross-calibration, the k Q mag for an Exradin A1SL ionization chamber, set up perpendicular to the incident photon beam and opposite to the direction of the Lorentz force, was directly determined in water in absolute terms to be 0.977 ± 0.010. The currently published k Q mag results, obtained via Monte Carlo calculations, agree with experimental measurements in this work within combined uncertainties. CONCLUSIONS: A novel design of an MR-compatible water calorimeter was successfully used to measure absorbed dose in an MR-linac and determine an experimental value of k Q mag for a clinically used ionization chamber.

Experimental measurement of ionization chamber angular response and associated magnetic field correction factors in MR-linac.

PURPOSE: To measure ionization chamber dose response as a function of the angle between magnetic field direction and ionization chamber orientation in magnetic resonance-guided radiation therapy (MRgRT) system, and to evaluate angular dependence of magnetic field correction factor for reference dosimetry. METHODS: Measurements were performed on an Elekta MR-linac that integrates a 1.5-T Philips MRI and a 7-MV FFF photon beam accelerator. The response of four reference class chambers (Exradin-A19, A1SL, IBA FC65-G, and CC13, paired with a PTW UE electrometer) was studied. An in-house built MR-compatible water tank and an accompanying cylindrical insert that allowed chamber rotation around the cylinder's axis was used. The EPID onboard imaging was used to center chamber at the MR-linac isocenter (143.5 cm, SAD), as well as to verify position at each datapoint. RESULTS: A clear angular dependence of dose response for all chambers has been measured. The most significant effect of magnetic field on relative chamber response in the presence of magnetic field was observed in the orientation when chamber axis is perpendicular to the direction of magnetic field with the tip pointing in the same direction as Lorentz force. This effect is more pronounced for larger volume chambers; the maximum relative variation in the chamber response (between the setup described above and the one where chamber and magnetic field are parallel) is a 5.3% and 4.6% increase for A19 and FC65-G, respectively, and only 2.0% and 1.9% for smaller volume A1SL and CC13 chamber, respectively. We measured the absolute magnitude of the magnetic field correction factor k Q mag for the Exradin-A19, A1SL, IBA FC65-G, and CC13 to be 0.938 ± 1.13%, 0.968 ± 0.99%, 0.950 ± 1.13%, and 0.975 ± 1.13%, respectively. The values are for perpendicular orientation of the chamber relative to magnetic field and parallel to the Lorentz force. CONCLUSIONS: Experimental measurements carried out in this study have verified the optimal orientation of ionization chamber in terms of minimizing effect of magnetic field on the chamber dose response. This study provides a detailed high-resolution measurement of absolute k Q mag values for four reference class chambers as a function of the angle between ionization chamber's central axis and the direction of strong magnetic field over a full 360° rotation. The experimental results of this study can further be used for optimization of the actual sensitive volume of the chamber (and analysis of dead volume) in future Monte Carlo chamber simulations in the presence of strong magnetic fields. In addition, it will provide some necessary data for future reference dosimetry protocols for MR-linac.