The increasing penetration of large-scale Renewable Energy Sources (RESs) has raised several challenges for power grid operation. Power management solutions supporting the integration of RESs, such as those based on energy storage technologies, are generally costly. Alternatively, promoting a more proactive role of the Distribution System Operator (DSO) to successfully manage RESs’ uncertainty, and take advantage of their flexible resources for the provision of ancillary services, can avoid installing expensive devices in the network and reduce costs. In this line, improved coordination between Transmission System Operators (TSOs) and DSOs is highly desirable. In this paper, the feasibility of solving different aspects of the integration of RESs through an improved TSO/DSO coordination is evaluated. In particular, a Systematic Literature Review (SLR) is conducted to study the most relevant TSO/DSO coordination approaches, exclusively focused on integrating distributed RESs, currently available in the literature. Their main operational, managerial, economic, and computational challenges, advantages, and disadvantages are discussed in detail to identify the most promising research trends and the most concerning research gaps to pave the way for future research toward developing a solid TSO/DSO coordination mechanism for integrating RESs efficiently. The main results of the SLR show a clear trend in implementing decentralized TSO/DSO coordination models since they provide efficient facilitation of RESs’ services, while reducing computational burden and communication complexity and, consequently, reducing operative costs. In addition, while different aspects of the TSO/DSO coordination implementation, such as reactive power and voltage regulation, operational cost minimization, operational planning, and congestion management, have been thoroughly addressed in the literature, further research is needed regarding data exchange mechanisms and RESs’ uncertainty modeling and prediction. In this line, the development of standardized communication solutions, based on the Common Grid Model Exchange Standard (CGMES) of the International Electrotechnical Commission (IEC), has shown promising interoperability results, whereas the use of learning-based approaches to predict RESs’ uncertain behavior and distribution networks’ responses, using only historical data, which relieves the need for access to commercially sensitive and proprietary network data, has also shown itself to be a promising research direction.