Recently, wireless sensor networks are gaining a lot of attention due to the availability of various low cost sensor devices. For example, Wireless Multimedia Sensor Networks (WMSNs) have lot of new potential applications in different domains including, environmental monitoring, home automation, tracking, health care, just to name a few. Another example is the Body Area Networks (BANs), where different sensors are attached with the patient's body. However, these sensor networks also bring several challenges, such as they are resource constraint, requires application specific QoS requirements, high bandwidth demands, and resistance to interference. In this paper, we argue that wireless sensor nodes with cognitive radio capability, together with channel bonding, can help to address these challenges. In this perspective, we discuss cognitive radio wireless sensor network for hospital environment as a case study. We then discuss the way of spectrum characterization and criteria for channel bonding in such networks. Finally, we highlight issues and challenges, which provides the basis to develop algorithms and protocols for the future cognitive radio wireless sensor network deployments in different application scenarios.