A novel 121 pixel silicon diode array detector has been characterized by Monte Carlo simulation for its performance inside magnetic fields representative of current prototype and proposed MRI-linear accelerator systems. In the in-line orientation, the silicon dose is directly proportional to the water dose. In the perpendicular orientation, there is a shift in dose response relative to water in the highest dose gradient regions, at the edge of jaw-defined and single-segment MLC fields. The trend was not observed in-field for an IMRT beam. The array is expected to be a valuable tool in MRIgRT dosimetry.

Background Liver tumors are subject to motion with respiration, which is typically accounted for by increasing the target volume. The prescription dose is often reduced to keep the mean liver dose under a threshold level to limit the probability of radiation induced liver toxicity. A retrospective planning study was performed to determine the potential clinical gains of removal of respiratory motion from liver SABR treatment volumes, which may be achieved with gating or tumor tracking. Methods Twenty consecutive liver SABR patients were analysed. The treated PTV included the GTV in all phases of respiration (ITV) with a 5 mm margin. The goal prescription was 50Gy/5# (BED 100 Gy 10 ) but was reduced by 2.5 Gy increments to meet liver dose constraints. Elimination of motion was modelled by contouring the GTV in the expiration phase only, with a 5 mm PTV margin. All patients were replanned using the no-motion PTV and tumor dose was escalated to higher prescription levels where feasible given organ-at-risk constraints. For the cohort of patients with metastatic disease, BED gains were correlated to increases in tumour control probability (TCP). The effect of the gradient of the TCP curve on the magnitude of TCP increase was evaluated by repeating the study for an additional prescription structure, 54Gy/3# (BED 151 Gy 10 ). Results Correlation between PTV size and prescribed dose exists; PTVs encompassing < 10% of the liver could receive the highest prescription level. A monotonically increasing correlation (Spearman’s rho 0.771, p = 0.002) between the degree of PTV size reduction and motion vector magnitude was observed for GTV sizes <100cm 3 . For 11/13 patients initially planned to a decreased prescription, tumor dose escalation was possible (5.4Gy 10 –21.4Gy 10 BED) using the no-motion PTV. Dose escalation in excess of 20 Gy 10 increased the associated TCP by 5% or more. A comparison of TCP gains between the two fractionation schedules showed that, for the same patient geometry, the absolute increase in BED was the overarching factor rather than the gradient of the TCP curve. Conclusions In liver SABR treatments unable to be prescribed optimal dose due to exceeding mean liver thresholds, eliminating respiratory motion allowed dose escalation in the majority of patients studied and substantially increased TCP.

A monolithic silicon small-field array detector is proposed for relative dosimetry applications in hybrid MRI-linac systems. The detector has high sampling resolution, with 512 active elements arranged with 2 mm pitch over a 46 mm×46 mm detection area. Experimental measurements were performed in a custom-designed permanent magnet device that is compatible with a standard clinical linear accelerator. It can be configured in both inline and perpendicular magnetic-field-to-photonbeam orientations and produces magnetic field strengths 0.95 T and 1.20 T, respectively. Monte Carlo simulation data, obtained using the GEANT4 toolkit, are presented to supplement experimental data. Beam profiles show agreement to EBT3 film within 0.5 mm for FWHM and penumbral width measurement of small square fields (width ranging from 0.75 cm to 2.25 cm), in both inline and perpendicular magnetic field orientations. The detector can be used to accurately resolve normalised beam profiles in magnetic fields. The impact of electron return effects (ERE) in a small air gap surrounding the detector was also quantified. For the perpendicular orientation, a reduced profile intensity was observed for an increasing air gap width above the detector (10% at 2 mm) due to ERE. In the inline orientation, a very small increase in response relative to the zero field case was observed with an air gap above the detector (2% at 2 mm). Calibration of the device in a magnetic field will therefore be necessary; the zero field calibration is non-transferable. The MagicPlate-512 provides a high-resolution real time alternative to accurately measure normalised beam profiles in magnetic fields, and is expected to be a suitable array detector for use in magnetic field environments typical of MRI-linacs.

This experimental work has demonstrated how strong inline magnetic fields act to enhance the dose to lower density mediums such as lung tissue. Clinically, such scenarios will arise in inline MRI-linac systems for treatment of small lung tumours.