Abstract:
An MR-Linac combines magnetic resonance (MR) imaging and a linear accelerator
(linac) for the treatment of cancer. This enables image-guided radiotherapy with high
soft-tissue contrast, which can improve patient prognosis. To calculate irradiation plans
for the precise irradiation of patients, an accurate linac model must be available, which
is created using data of the individual linac measured with detectors. However, small
radiation fields and the magnetic field, which cannot be switched off in clinical routine,
can influence the detector's signal. This must be accounted by the output correction
factor k_B,Q,Clin. With correctly determined k_B,Q,Clin, smaller radiation fields can be
used for treatment of patients in the future.
Since no protocol for small field dosimetry in magnetic fields exists to date, this thesis
extended the established international and German dosimetry protocols for a situation
in a magnetic field, thereby enabling small field measurements in magnetic fields to be
traceable to primary standards for the absorbed dose to water (e.g. water calorimetry).
Furthermore, k_B,Q,Clin was determined for suitable detectors with measurements at
an experimental setup – consisting of a mobile electromagnet placed in front of a
clinical linear accelerator – as well as Monte Carlo simulations.
In the first step of this thesis, the dosimetry protocols were expanded. The influence of
the small field and the magnetic field can be combined into a single factor (k_B,Q,Clin).
Alternatively, the magnetic field’s influence can be isolated (k_B,Q,Ratio), which allows the use of output correction factors without a magnetic field, k_Q,Clin, from established
dosimetry protocol to determine k_B,Q,Clin. The latter method, newly developed in this
work, may be preferable to cope with the large uncertainty and substantial variability
in published k_Q,Clin and k_B,QmClin, since k_B,Q,Ratio is more comparable between
literature values.
In the second step, the general behaviour of the correction factors was examined. For
suitable small-field detectors (in all detector orientations), k_B,Q,Clin equals one for field sizes >3 × 3 cm2 for all perpendicular MR-linac systems. For these field sizes, the
established correction factor for the influence of the magnetic field at reference field size (usually 10 × 10 cm2), k_B,Q,msr, can be used including its determined uncertainty
arising from intra-type variation (e.g. for cross-calibrations). For smaller field sizes –
approximately when loss of lateral charged particle equilibrium occurs – the detectors’
k_B,Q,Clin is dependent on photon energy, magnetic flux density, detector orientation
and detector position.
In the third step, the influence of the detector position on the correction factors was
determined. For solid-state detectors, k_B,Q,Clin is the same at the central axis (CAX)
and at the lateral maximum of the beam profile (MAX), whereas for ionization
chambers, k_B,Q,Clin is lower at MAX. Contrary to a situation with no magnetic field, it
is therefore necessary to define a position for k_B,Q,Clin when using ionization
chambers. Furthermore, for dose distribution measurements no effective point of
measurement (EPOM) shift is needed for solid-state detectors, whereas for ionization
chambers, a lateral and longitudinal EPOM shift must be considered.
In a fourth step, the intra-type variation of k_B,Q,msr (max. ~0.5%), k_Q,Clin, k_B,Q,Clin and k_B,Q,Ratio (max. ~1%) was determined. This variation must be considered for the
uncertainty of tabulated factors in dosimetry protocols. The exact origin of the variation
is further unclear.
The comparison with a clinical MR-linac showed that the results determined in this
work are representative for a 1.5 T MR-linac. Therefore, the factors k_B,Q,Clin and
k_B,Q,Ratio determined in this work can be used alongside with literature data to
commission the treatment planning system (TPS) of MR-linacs for small fields. These
results, together with summarizing fits, are listed in the Appendix 9.7 and enable the
future use of smaller and more complex radiation fields at MR-linacs. Since the
magnetic field can influence the correction factor without a magnetic field (k_Q,Clin) for
field sizes below 2.5 × 2.5 cm2, current TPS commissioning data of MR-linacs may be
incorrect and should be corrected promptly.