The Wide Field Infrared Survey Telescope (WFIRST) Coronagraphic Instrument (CGI) technology demonstration and potential science mission, as well as mission concepts with exoplanet imaging capabilities such as HabEx and LUVOIR, all require the identification of the best targets for exoplanet observations. To date, the focus has been primarily on two classes of targets: those with known exoplanets discovered indirectly that may be observable by these imaging missions, and those targets with no known planets that have high completeness values (probabilities of planet detections) under some assumptions of the instrument performance and the overall population of exoplanets. A third class of target, however, has received much less scrutiny: stars with known exoplanets that could not possibly be directly imaged due to the size of their orbits. These are planets that would be guaranteed to spend all of their time inside the inner working angles or outside the outer working angles of all currently proposed coronagraphs. However, these systems could potentially harbor additional planets, either exterior to or interior to the currently known planets, but not yet detectable by indirect means. Here, we discuss how to identify systems from all three categories that would be good targets for WFIRST and other, future, space-based imagers. We present a method for assessing the utility of these targets based on an exploration of the available dynamical phase space of the systems that would result in long-term stable orbits for both the currently known and the potentially discoverable companions, and show how it augments existing methods for assessing target completeness and utility.