Clinical utility of the K-T cancellation test in a memory clinic population