Techniques for and uncertainties of MRI-based reconstruction of titanium tandem and ring brachytherapy applicators

Purpose Eliminating patient computed tomography (CT) scans for tandem and ring (T&R) brachytherapy can reduce overall procedure time and eliminates imaging dose. However, reconstructing titanium applicators in magnetic resonance imaging (MRI) is challenging. We evaluated the uncertainty of different applicator reconstruction workflows in MR-guided brachytherapy, and assessed the clinical impact of reconstruction uncertainties. Methods and materials Titanium MRI-compatible T&Rs with aqueous gel in the buildup cap were reconstructed on CTs and MRIs to assess the uncertainties of four different workflows. Reconstruction was performed using (1) proton density–weighted MRIs with solid applicator from a library, (2) applicator-only reference CT fused with MRIs, (3) T2-weighted (T2W) MRIs following GEC-ESTRO guidelines, and (4) patient CTs fused with patient MRIs with in situ applicators. We evaluated dwell positions and plan quality differences using high-risk clinical target volume coverage, and EQD2 D2cc of rectum, sigmoid, bladder, and small bowel. Results The 2σ uncertainty for dwell positions for each workflow were (1) 2.7 mm for both ring and tandem, (2) 1.4 mm ring and 0.8 mm tandem, (3) 0.2 mm ring and 0.8 mm tandem, and (4) 1.9 mm ring and 0.4 mm tandem. Reconstruction uncertainties resulted in dose variations within acceptable levels (below 10%) except for (1) which resulted in larger dose to the rectum (20%). Dose uncertainties were similar between reference CT and patient CT. Conclusions Reconstruction with a reference CT results in similar uncertainty to a patient CT. T2W MRI plans have acceptable uncertainty levels for the applicator reconstruction and resulting dose distributions.