[en] Purpose/Objective: A novel image-guided robotic radiosurgical system, capable of irradiating 102 non-coplanar nodes in 3 π geometry, produces complex dose distributions which are difficult or impractical to measure with conventional dosimetry instrumentation. The recently developed BANG polymer gel dosimetry system provides accurate, high resolution and three dimensional dose distributions data and is ideally suited for the task described above. In this study, the polymer gels were used for imaging the dose distributions produced by this extremely flexible radiosurgical system. Materials and Methods: The dosimeter materials consist of 2-liter BANG polymer gels in spherical, clear glass flasks, closed with ground glass stoppers, with glass rods extending to the center of the gel that serve as a target for the frameless robotic radiosurgery. A compact 6 MV x-band linac (285 lbs) is mounted and maneuvered by a 6 degree-of-freedom robotic arm. The gels were irradiated using a 25 mm circular insert. A total of 10 Gy was delivered at isocenter at a dose rate of 300 cGy/min using all of the available 102 nodes. The gels were then imaged by MRI(GE Signa) at 1.5 T, using a series of Hahn spin echoes of TR = 3s, TE = 20,100,200,400 ms. Transverse relaxation rate (R₂) maps were constructed from those multiple images, using the non-linear least-squares Lavenberg-Marquardt algorithm and a data analysis and display program 'DoseMap' which was written using the scientific computational program MATLAB. R₂ maps were converted to dose maps using an R₂-to-dose calibration curve. Dose maps and isodose curves were then compared with corresponding data from the treatment planning computer software. Results: The dose dependence of the NMR transverse relaxation rate, R₂, is reproducible (less than 2 % variation) and is linear up to about 10 Gy, with a slope of 0.25 s^-1Gy^-1 at 1.5 Tesla. Isodose curves in three orthogonal (axial, sagittal and coronal) planes show excellent agreement between the gel data with those calculated by the treatment planning system. Conclusion: The tissue-equivalent polymer-gel dosimeter provides a unique and reliable approach for visualization and validation of three dimensional dose distributions produced by the novel robotic radiosurgerical system. Furthermore, the polymer-gel dosimeter is stable and can be imaged with delays of several months if desired without a loss of the dose distribution data