The presence and availability of computer-aided diagnostic tools have increasingly become a reality in recent years, assuming an increasingly intrinsic role in the process. Thus, in the field of mammography, the use of computer-aided diagnostic schemes such as CADe or CADx is being proposed as a means to reduce the inherent difficulties in the evaluation process of such images and as an alternative to the need for another professional to be available for a second opinion, which could enhance the accuracy of the examinations. Therefore, this project aims to conduct a statistical survey regarding the performance of a Driven CADx scheme, which classifies only findings within a previously demarcated region of interest, developed for digitized mammographic images, now applied to two image databases obtained from current DR mammography equipment, one derived from a database of real images and the other obtained through exposure of a breast phantom. Consequently, the results obtained from the evaluation will enable the refinement of this computational scheme for practical and effective use.