Among rapidly development in immuno-oncology, one critical question is whether drugs targeting the receptor (e.g. PD-1) have a different safety and efficacy profile from the drugs targeting the ligand (e.g. PD-L1) on the same axis of the immune system. Direct comparison of these molecules would be ideal but highly unlikely. Existing methods of indirect comparison relying on a network of relative effect are not applicable, either, due to the lack of randomized studies with a common control arm. We proposed a meta-analytic procedure based on a novel Bayesian hierarchical model for indirect comparison without requirement of a network structure, while the performance improved when such a network exists. Simulation studies showed synchronizing effect under the assumption of exchangeability across indications. An example based on real data showed that the proposed model successfully synchronized information across multiple indications and concluded a strong class effect of the drugs targeting either the receptor or the ligand on the same axis, demonstrating the important impact of the proposed method in addressing the critical medical question in real world application.