The spatial reconstruction of optical properties in Diffuse Optical Tomography (DOT) can be considered as a large-scale optimization procedure. The forward model is based on the Radiative Transfer Equation (RTE) in the frequency domain. In this work, a Gradient-based algorithm which employs the gradient of the objective function has been used. Moreover, the gradient of this function was calculated efficiently through an easy and fast way based on the adjoint method. The results prove that the algorithm is robust to reconstruct the spatial distribution of the optical properties in 2D and 3D geometries.