Dosimetric Comparison in Malignant Glioma Patients Clinically Treated on Hybrid Magnetic Resonance Imaging (MRI)-Linac (MRL) vs. Conventional Linac

Abstract

The Magnetic Resonance Imaging (MRI)-Linac (MRL) is a hybrid machine integrating a high field strength 1.5T MRI with a linear accelerator, providing superior soft tissue visualization of tumors and organs-at-risk (OARs) during treatment delivery. A special consideration for MRL radiotherapy is accounting for interactions of secondary electrons generated within the magnetic field, which can alter dose deposition at air-tissue interfaces. We evaluated dosimetric outcomes in clinically treated malignant glioma patients who received at least one fraction of radiotherapy on both the MRL and a conventional Linac.Thirty-seven glioma patients treated on both the MRL and a conventional Linac for adjuvant chemoradiotherapy between July 2019 and February 2021 were analyzed. Planning was completed on treatment planning systems (TPS) using a Monte-Carlo algorithm that accounts for magnetic field effects (Monaco v5.40) for the MRL, and a convolution-superposition algorithm (Pinnacle v9.8) for the conventional Linac. Dosimetric parameters of interest from the target, OARs, and air-tissue interface volumes for each patients' clinical treatment plans were extracted and compared. For 10 representative patients, in vivo skin doses during a single fraction of MRL and conventional Linac treatment were obtained using an Optically Stimulated Luminescent Dosimeter (OSLD) placed in a defined location on the patient's skin near the Planning Target Volume (PTV). Student's t-test and Wilcoxon signed-rank test were used to compare parameters between Monaco and Pinnacle. Spearman's correlation was used to assess the relationship between in vivo OSLD measurements and TPS skin dose. Threshold for statistical significance was P < 0.05.Most patients were treated for high grade glioma (76% Grade III or IV, 24% Grade II), and median PTV was 257.4 cm3 (range, 37.1-570.3 cm3). MRL and conventional Linac had similar V100, V95, D98, and D95 for PTV, and D3cc for optic chiasm, optic nerves, and each cochlea (P = NS). However, clinically delivered Monaco plans had significantly greater doses within air cavities (mean Dmean higher by 1.3 Gy, P < 0.0001) and skin (mean Dmean higher by 1.9 Gy, P < 0.0001; mean D2cc higher by 8.1 Gy, P < 0.0001; mean V20 Gy higher by 7.2 cm3, P < 0.0001), compared to clinically delivered Pinnacle plans. In vivo OSLD skin readings were 14.5% greater for treatments delivered on the MRL (P = 0.0027), and were more accurately predicted by Monaco (r = 0.95, P < 0.0001) vs. Pinnacle (r = 0.80, P = 0.0096).In this prospective study of clinically treated glioma patients on both MRL and conventional Linac, the dosimetric impact of the magnetic field was minimal for the target and standard OARs. However, higher doses to skin and air cavities were observed. In vivo correlation of dose to skin was more accurately predicted with Monaco. Future MRL planning processes are being designed to account for skin dosimetry and treatment delivery.