In concert with increasing co-ordination of energy markets, transmission planning for location-constrained generation is forcing a shift from regional to wide-area transmission network planning. Planning for location-constrained generators, notably wind and solar resources, is complicated by uncertainty in policies driving their growth. To design transmission systems robust to uncertain future expansion of location-constrained generators, planners must incorporate stochastic generation into models. The size of the transmission expansion problem makes this computationally intractable. This paper presents a screening model to reduce the number of transmission investments which must be considered. The screening model combines sampled generation expansion scenarios with a two-stage relaxed planning model to identify corridors of interest. A St.Clair curve-based filter translates corridors to investment variables. In a 240 bus model of WECC, this method reduced the number of corridors of interest 98% from 19,503 to 303 and reduces the number of investment variables 99% from 58,503 to 432.