ARM variable for Screening failure patients,

We have a patient who did not meet inclusion criteria but was randomized mistakenly, and we will treat the patient as "Screening Failure".

I understood that ARM variable in SDTM should consist of planned arm (assigned to through randomization).

For this case of SF patients who proceeded randomization mistakenly, which data (treatment group randomized vs. 'Screen Failure' vs. blank) should be filled in ARM variable?

Please consider that this case is not same with the case of patients who are randomized in treatment group but not treated.