Evaluation of MRI-only based online adaptive radiotherapy of abdominal region on MR-linac

Abstract

Purpose: A hybrid magnetic resonance linear accelerator (MRL) can perform magnetic resonance imaging (MRI) with high soft-tissue contrast to be used for online adaptive radiotherapy (oART). To obtain electron densities needed for the oART dose calculation, a computed tomography (CT) is often deformably registered to MRI. Our aim was to evaluate an MRI-only based synthetic CT (sCT) generation as an alternative to the deformed CT (dCT)-based oART in the abdominal region.

Methods: The study data consisted of 57 patients who were treated on a 0.35 T MRL system mainly for abdominal tumors. Simulation MRI-CT pairs of 43 patients were used for training and validation of a prototype convolutional neural network sCT-generation algorithm, based on HighRes3DNet, for the abdominal region. For remaining test patients, sCT images were produced from simulation MRIs and daily MRIs. The dCT-based plans were re-calculated on sCT with identical calculation parameters. The sCT and dCT were compared in terms of geometric agreement and calculated dose.

Results: The mean and one standard deviation of the geometric agreement metrics over dCT-sCT-pairs were: mean error of 8 ± 10 HU, mean absolute error of 49 ± 10 HU, and Dice similarity coefficient of 55 ± 12%, 60 ± 5%, and 82 ± 15% for bone, fat, and lung tissues, respectively. The dose differences between the sCT and dCT-based dose for planning target volumes were 0.5 ± 0.9%, 0.6 ± 0.8%, and 0.5 ± 0.8% at D_2% , D_50% , and D_98% in physical dose and 0.8 ± 1.4%, 0.8 ± 1.2%, and 0.6 ± 1.1% in biologically effective dose (BED). For organs-at-risk, the dose differences of all evaluated dose-volume histogram points were within [-4.5%, 7.8%] and [-1.1 Gy, 3.5 Gy] in both physical dose and BED.

Conclusions: The geometric agreement metrics were within typically reported values and most average relative dose differences were within 1%. Thus, an MRI-only sCT-based approach is a promising alternative to the current clinical practice of the abdominal oART on MRL.

Keywords: MR-linac; MRI-only; online adaptive radiotherapy; synthetic CT.

FIGURE 1

Examples of simulation magnetic resonance imaging (MRI), deformed computed tomography (CT), and synthetic CT. Corresponding axial slices at the isocenter level of two patients with (a and d) the simulation MRI, (b and e) the deformed CT, and (c and f) the synthetic CT produced with the prototype abdominal. The overall geometric agreement was among the worst for patient 1 and among the best for patient 2

FIGURE 2

The mean absolute error (MAE) between deformed computed tomography (dCT) and synthetic CT (sCT) in 20 HU intervals. The MAE is averaged over (a) the abdominal cases, (b) the pelvic cases, and (c) all the cases. The gray area defines one standard deviation (1SD). The drop at –500 HU shows the threshold where air pockets were set to 0 HU

FIGURE 3

Accumulated dose distributions on deformed computed tomography (dCT) and synthetic CT (sCT). The accumulated physical dose distributions calculated for three patients on dCT (a, c, and e) and sCT (b, d, and f). The prescribed doses are 50 Gy in five fractions to adrenal gland for patient 1, 24 Gy in three fractions to pancreas for patient 2, and 75 Gy in five fractions to liver for patient 3. The dose distributions are thresholded at 5, 3, and 15 Gy for the patients 1–3, respectively. The overall dose differences were among the best for patients 1 and 2, and among the worst for patient 3.