Real-time tracking of lung tumors using a 1.5T Elekta Unity MR-Linac: first clinical experiences

Abstract

There is a growing interest in the application of MRI-guided adaptive radiotherapy (MRIgART) to improve the treatment of lung tumors. Motion management plays a central role to better localize these tumors and minimize toxicities to surrounding healthy tissue. Elekta recently released their comprehensive motion management system for the Unity 1.5T MR-Linac, enabling real-time tracking and predictive gating using MR cine imaging during treatment. We report our first clinical experiences using this gating technology to treat patients with central lung tumors adjacent to critical structures on the Unity MR-Linac using both conventional and SBRT fractionations. A surrogate structure was tracked if the primary target could not be visualized. Beam gating was automatically performed if the system detected the target moving outside of the defined gating envelope, which was set as the planning target volume. Integrating this technology has profoundly impacted our MRIgART program; we have been able to treat challenging lung tumor patients with greater precision and have begun to reduce our PTV margins, leading to improved dose sparing of nearby critical structures.

Keywords: MR-Linac; adaptive; gating; lung; motion management.

Figure 1

Subset of lung sites treated using MR cine gating on the Elekta Unity MRLinac. The reference CT is shown in columns 1 and 2 with the coronal and sagittal views of the target, green, gating envelope, red, and surrogate (peach, if applicable). Columns 3 and 4 are the MR cine analogs of the reference plan showing the tracking structure, surrogate (if used), and gating envelope.

Figure 2

(A) Deviations from the reference planning DVH objective constraints for online ATS-adapted and retrospective ATP-planned SBRT treatments for lung tumors. Error bars represent the k = 1 standard error of the mean. (B) Distribution of tumor motion with abdominal compression. Compression constancy was tracked for each fraction by comparing the diaphragm motion at the time of simulation against the diaphragm motion measured using a 2D cine prior to treatment.