CAD (computer-aided diagnosis) schemes, developed since the 90’s to aid mammography image analysis, are currently being used as CADe systems, designed only to detect signals of clinical interest, working as second opinion to radiologists. However, expressive number of physicians is not satisfied with them due to high false positive rates as well as to the lack of useful information regarding the case interpretation. This points toward the CADx schemes (designed to provide also information on the classification of found signals) as a most appropriate tool to the mammography image evaluation. Preliminary evaluation of a CADx scheme developed by our research group - considering a set of 147 cases (78 malignant, 69 benign) selected from the DDSM database, overviewed by experienced radiologist in appropriate LCD displays before and after knowing information from CADx processing - shows that the CADx scheme was capable of positively influencing the radiologist opinion, since the rates of diagnoses after knowing the CADx information were always better than those from radiologist's first impression. Therefore, the high sensitivity for diagnosing malignant cases was remarkable as well as the decrease in false positive rates with the CADx aid, pointing out a useful support of such computational tool in mammography image interpretation.