[en] Purpose: Dose distributions around low energy (< 60 keV) brachytherapy sources, such as ¹²⁵I, are known to be very sensitive to changes in tissue composition. Available ¹²⁵I dosimetry data describe the effects of replacing the entire water medium by heterogeneous material. This work extends our knowledge of tissue heterogeneity effects to the domain of bounded tissue heterogeneities, simulating clinical situations. Our goals are three-fold: (a) to experimentally characterized the variation of dose rate as a function of location and dimensions of the heterogeneity, (b) to confirm the accuracy of Monte Carlo dose calculation methods in the presence of bounded tissue heterogeneities, and (c) to use the Monte Carlo method to characterize the dependence of heterogeneity correction factors (HCF) on the irradiation geometry. Methods and Materials: Thermoluminescent dosimeters (TLD) were used to measure the deviations from the homogeneous dose distribution of an ¹²⁵I seed due to cylindrical tissue heterogeneities. A solid water phantom was machined accurately to accommodate the long axis of the heterogeneous cylinder in the transverse plane of a ¹²⁵I source. Profiles were obtained perpendicular to and along the cylinder axis, in the region downstream of the heterogeneity. Measurements were repeated at the corresponding points in homogeneous solid water. The measured heterogeneity correction factor (HCF) was defined as the ratio of the detector reading in the heterogeneous medium to that in the homogeneous medium at that point. The same ratio was simulated by a Monte Carlo photon transport (MCPT) code, using accurate modeling of the source, phantom, and detector geometry. In addition, Monte Carlo-based parametric studies were performed to identify the dependence of HCF on heterogeneity dimensions and distance from the source. Results: Measured and calculated HCFs reveal excellent agreement (≤ 5% average) over a wide range of materials and geometries. HCFs downstream of 20 mm diameter by 10 mm thick hard bone cylinders vary from 0.12 to 0.30 with respect to distance, while for an inner bone cylinder of the same dimension, it varies from 0.72 to 0.83. For 6 mm diameter by 10 mm thick hard bone and inner bone cylinders, HCF varies 0.27-0.58 and 0.77-0.88, respectively. For lucite, fat, and air, the dependence of HCF on the 3D irradiation geometry was much less pronounced. Conclusion: Monte Carlo simulation is a powerful, convenient, and accurate tool for investigating the long neglected area of tissue composition heterogeneity corrections. Simple one dimensional dose calculation models that depend only on the heterogeneity thickness cannot accurately characterize ¹²⁵I dose distributions in the presence of bone-like heterogeneities