Therapeutic treatment plan evaluation is often based on examining the radiotherapy treatment planning (RTP) system dose distributions in the target and surrounding normal structures. To study the effects of tissue inhomogeneities on photon dose distributions, we compared FOCUS RTP system dose distributions from the measurement-based Clarkson and model-based MultiGrid Superposition (MGS) algorithms with those from the BEAM Monte Carlo code system in a set of heterogeneous phantoms. The phantom inhomogeneities mimic relevant clinical treatment sites, which include lung slab, lung-bone slab, bone-lung slab, mediastinum, and tumor geometries. The benchmark comparisons were performed in lung densities of 0.20 and 0.31 g/cm3, and a bone density of 2.40 g/cm3 for 5x5 cm2 and 10x10 cm2, 6- and 15-MV photon beams. Benchmark comparison results show that the MGS model and BEAM doses match better than 3% or 3 mm, and the MGS model is more accurate than the Clarkson model in all phantoms. The MGS model, unlike the Clarkson model, predicts the build-down and build-up of dose near tissue interfaces and penumbra broadening in lung associated with high energy beams. The Clarkson model overestimates the dose in lung by a maximum of 10% compared to BEAM. Dose comparisons suggest turning-off the effective path length inhomogeneity correction in the Clarkson model for lung treatments.