Feasibility of artificial-intelligence-based synthetic computed tomography in a magnetic resonance-only radiotherapy workflow for brain radiotherapy: Two-way dose validation and 2D/2D kV-image-based positioning

Abstract

Background and purpose: Magnetic Resonance Imaging (MRI)-only workflow eliminates the MRI-computed tomography (CT) registration inaccuracy, which degrades radiotherapy (RT) treatment accuracy. For an MRI-only workflow MRI sequences need to be converted to synthetic-CT (sCT). The purpose of this study was to evaluate a commercially available artificial intelligence (AI)-based sCT generation for dose calculation and 2D/2D kV-image daily positioning for brain RT workflow.

Materials and methods: T1-VIBE DIXON was acquired at the 1.5 T MRI for 26 patients in RT setup for sCTs generation. For each patient, a volumetric modulated arc therapy (VMAT) plan was optimized on the CT, then recalculated on the sCT; and vice versa. sCT-based digitally reconstructed radiographs (DRRs) were fused with stereoscopic X-ray images recorded as image guidance for clinical treatments. Dosimetric differences between planned/recalculated doses and the differences between the calculated and recorded clinical couch shift/rotation were evaluated.

Results: Mean ΔD₅₀ between planned/recalculated doses for target volumes ranged between -0.2 % and 0.2 %; mean ΔD₅₀ and ΔD_0.01ccm were -0.6 % and 1.6 % and -1.4 % and 1.0 % for organ-at-risks, respectively. Differences were tested for clinical equivalence using intervals ±2 % (dose), ±1mm (translation), and ±1° (rotation). Dose equivalence was found using ±2 % interval (p < 0.001). The median differences between lat./long./vert. couch shift between CT-based/sCT-based DRRs were 0.3 mm/0.2 mm/0.3 mm (p < 0.05); median differences between lat./long./vert. couch rotation were -1.5°/0.1°/0.1° (after improvement of RT setup: -0.4°/-0.1°/-0.4°, p < 0.05).

Conclusions: This in-silico study showed that the AI-based sCT provided equivalent results to the CT for dose calculation and daily stereoscopic X-ray positioning when using an optimal RT setup during MRI acquisition.

Keywords: Artificial intelligence; Brain radiotherapy; MRI-only workflow; Synthetic CT; kV-image-based positioning.

Fig. 1

The colorwash of the planned dose distribution on image datasets. (a) Planned dose distribution on the planning CT and (b) on the reconstructed AI-based synthetic CT of the brain (bone windowing, W = 1600, L = 450).

Fig. 2

Dosimetric evaluation of sCT-based plan compared to CT-based plan. (a) Percentage deviation of D₅₀ (ΔD₅₀[%]) of target volumes and organ-at-risks (OAR); and (b) percentage deviation of D_0.01ccm (ΔD_0.01[%]) of OARs. The color bar represents the dose deviation. Values for TP_CT→sCT are indicated with circular (filled) markers, whereas TP_sCT→CT are indicated with rectangular (unfilled) markers.

Fig. 3

Evaluation of PTV coverage of sCT-based plan compared to CT-based plan. (a) The PTV coverage (conformity index (CI)) of TP_CT→sCT and (b) of TP_sCT→CT; (b) the PTV coverage differences between planned and recalculated dose using the two evaluation methods.

Fig. 4

Comparison of sCT-based DRR and CT-based DRR on the 2D/2D kV-X-ray-based patient setup verification system. (a) DRR-X-ray registration at ExacTrac system using CT (top) and sCT (bottom). (b) (Left) couch shift difference (T: translation) and (right) couch rotation difference (R: rotation) for each patient in lateral, longitudinal, and vertical direction. Blue dashed line represents the 0 value (no deviation). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)