Checkpoint inhibitors and progressive multifocal leukoencephalopathy: friends of foes?