The aim of this study is to evaluate the dosimetric impact of gadolinium contrast medium (Gadovist) in a transverse MR-Linac system using Monte Carlo methods. The dose distributions were calculated using two heterogeneous multi-layer phantoms consisting of Gadovist, water, bone, and lung. The photon beam was irradiated with a filed size of 5 × 5 cm2, and a transverse magnetic field of 0-3.0 T was applied perpendicular to the incident photon beam. Next, dose distributions for brain, head and neck (H&N), and lung cancer patients were calculated using a patient voxel-based phantom with and without replacing the patient's GTV with Gadovist. The dose at the water-Gadovist interface increased by 8% without a magnetic field. A similar dose increment was observed at 0.35 T. In contrast, the dose increment at the water-Gadovist interface was small at 1.5 T and a dose decrement of 5% was observed at 3.0 T. The dose variation at the lung-Gadovist interface was larger than that at the water-Gadovist interface. The mass collision stopping power ratio for Gadovist was 7% lower than that for water, whereas, the electron fluence spectra at the water-Gadovist interface increased by 17.5%. In a patient study, Gadovist increased the Dmean for brain, H&N, and lung cancer patients by 0.65-8.9%. The dose variation due to Gadovist grew large in the low-dose region in H&N and lung cancer. The GTV dose variation due to Gadovist in all treatment site was below 2% at 0-3 T if the Gadovist concentration was lower than 0.2 mmol/ml-1.