Recently the iStar modeling language has evolved to a new version (iStar 2.0), where this version joins the di erent variations of iStar and standardizes it. We believe that it is natural and necessary the core of iStar and any other language to evolve. Nevertheless, iStar has been extended many times, so it is time to research about iStar extensions. Extending a modeling language is to add new constructs or modify the old ones. According to the way new concepts are proposed, an extension can be developed using a light-weight or heavy-weight strategy. The rst iStar extension was proposed 16 years ago (2001 - 2017). So, this talk reviews the existing iStar extensions by presenting a quantitative analysis and the perspectives of iStar researchers. Some challenges will be discussed such as extension mechanisms and a process to guide iStar extensions.