Response to "On patient-level splitting, contrast-free claims and unsupported comparators in Machine learning-based classification of multiple sclerosis lesion activity using multi-sequence MRI radiomics".