The ITER high-frequency (HF) magnetic diagnostic system has to provide essential measurements of MHD instabilities with |deltaBMEAS/BPOL|~10^-4 (~1G) for frequencies up to 2MHz to resolve toroidal mode numbers (n) in the range |n|=10-50. A review of the measurement requirements for HF MHD instabilities in ITER was initiated during the TW4 work-program and led to significant interest for physics and real-time control issues in measuring modes with |deltaBMEAS| as low as ~10^-4G at the position of the sensors, with |n|<30 and poloidal mode numbers |m|<60, for a frequency range extending up to ~500kHz. We have examined the ability of the current ITER design for the individual sensors and the diagnostic system as a whole to meet these needs, and have explored what adjustments to the design (of the individual sensors and/or of the system as a whole) or to the requirements would be needed to meet them when considering different hypothesis for the financial costs and risk management over the ITER life-time. First, we find that the proposed diagnostic layout, with 168 sensors in total, does not meet the more stringent measurement requirements and risk management criteria: these can only be met by a revision of the design requiring 350-500 sensors, depending on different costing and risk management options. Second, we find that the current design for the ITER HF Mirnov-type pick-up coil could be usefully revised.