The spectrum sensing strategy plays a vital role in Amplify-Forward (AF)-Cognitive Radio Networks (CRNs). However, AF-CRN cannot obtain maximal throughput, when existing sensing strategies are applied to AF-CRNs. In this paper, we present a superposition approach in AF-CRNs, in which firstly a Secondary User (SU) extends its sensing time until right before the beginning of its reporting time slot, and secondly each SU sends its measurement results containing amplified reports to the Cluster Head (CH), while the CH with soft-fusion report is forwarded to the Fusion Center (FC). With such extended sensing intervals and amplified reporting, a better sensing performance can be obtained than with the conventional rigid strategy. In addition to this, the sum rate of primary and secondary networks is also investigated for the superposition approach and the n-out-of-k rule. Numerical experiments show that the proposed strategy guarantees maximum sum rate compare to the conventional strategy under any condition.