In Cognitive Radio Ad-Hoc Networks the design of suitable routing solutions is a focal issue to fully unleash the potentials of this networking paradigm. The main challenge is exploiting spectrum holes to build up network paths that remain stable and that achieve specific network performance in terms of delay and percentage of delivered data, even if an opportunistic spectrum access is implemented. In this paper we propose a utility function based on the path connectivity, re-elaborated in a cognitive radio scenario, and we compare it with other utility functions that can be used for routing data in cognitive radio. We show that by using our utility function we select paths for the secondary users transmissions leading to better performance when compared with a utility function that selects paths with the minimum activities of the primary users and an utility function that minimizes the number of hops. These results are derived in great number of topologies and with different primary users activities.