Data dissemination is a fundamental task in wireless sensor networks. Because of the radio range limitation and energy consumption constraint, sensor data is commonly disseminated in a multihop fashion (flat networks) through a tree topology. However, to the best of our knowledge none of the current solutions deals with the moment when the dissemination topology needs to be rebuilt. This work addresses such a problem introducing the use of information fusion mechanisms, where the traffic is handled as a signal that is filtered and translated into evidences that indicate the likelihood of critical failures occurrence. These evidences are combined by a Dempster-Shafer engine to detect the need for a topology reconstruction. Our solution, called Topology Rebuilding Algorithm (TRA), is evaluated through a set of simulations. In the experiments, TRA showed to be efficient in avoiding unnecessary topology reconstructions. Compared to the periodic rebuilding, in some cases, TRA could reduce the traffic overhead in nearly 35% of the traffic produced by the periodicu rebuilding.