Train a pre-trained model based on CT images with MRI images


I want to use and train Nvidia’s pre-trained model “segmentation_ct_lung_tumor”.

Then, I have one question.

I think the “segmentation_ct_lung_tumor” was trained on CT images.
So, do I have to train “segmentation_ct_lung_tumor” on CT images (.nii or .dcm) ?
Is it possible to train on MRI images (.nii or .dcm), instead of CT images ?
This is because I have lung tumor MRI images but don’t have CT images.

Thank you.