On the feasibility of absolute 3D dosimetry using LiF thermoluminescence detectors and polymer gels on a 0.35T MR-LINAC.

Background And Purpose: As shown in our previous study, highly accurate absolute dosimetry in 3D is feasible by combining polymer gels (PG) with thermoluminescence dosimetry (TLD). In this setup, the thermoluminescence (TL)-based point dose information is used to renormalize the PG. This new PG-TLD reference system is now extended to measurements in magnetic fields.

Materials And Methods: Experiments were carried out on a conventional 6 MV linear accelerator (LINAC) and a 6 MV 0.35 T magnetic resonance (MR)-LINAC. Signal stability of TLD600 and TLD700 was examined without and with magnetic field. Afterwards, the combination of PAGAT PG and TL detectors was employed within a cylindrical phantom in presence of the magnetic field. Two scenarios were tested: (I) an air-filled phantom and (II) a water-filled phantom. For each scenario, two plans were irradiated: (a) opposed beams with a field size of 10 × 10 cm and (b) a 3D conformal plan assuring homogeneous target coverage using three equally distributed coplanar beams.

Results: Mean relative uncertainty of TL calibration reproducibility for TLD600/TLD700 was 0.49%/0.85% at the MR-LINAC and 0.48%/0.83% for the conventional LINAC. Individual TL calibration coefficients of TLD600 and TLD700 behaved differently in the presence of the magnetic field. An average difference of (3.29 ± 0.89)% occurred for all TLD600, whereas the result for TLD700 is not quite as clear with (1.09 ± 0.89)% after excluding some outliers. Using the TL dose information for PG renormalization, high 3D gamma passing rates were achieved using the 3%/2 mm criteria: 91.0% (Ia), 92.6% (Ib), 94.3% (IIa), 97.4% (IIb).

Conclusion: This study shows that TL signal reproducibility is not affected by a low magnetic field. Nevertheless, absolute calibration coefficients of the individual detectors indicate a dependency on the magnetic field. Hence, a calibration at the appropriate LINAC type is recommended. Furthermore, the previously established renormalization method for PG was applied to measurements at a MR-LINAC and was verified as suitable for evaluations of homogeneous dose distribution in the target volume.