Background: Although fine needle aspiration (FNA) is widely accepted for the diagnosis of breast cancer, some cases not classifiable as benign or malignant are diagnosed as inconclusive. Although these may require surgical biopsy, many of these lesions are benign. Methods: We used an image analysis and an automated learning system (Xcyt) to categorize 57 (38 benign, 19 malignant) breast FNAs diagnosed as "indeterminate" and compared the computer diagnosis to the surgical biopsy. For each case an operator chose a single group of cells on the FNA slide and digitized this image using a video camera. The outline of each nucleus was manually outlined, and the exact border was delineated by the computer. Based on the analysis of three nuclear features (area, texture and smoothness) the Xcyt system computed a benign or malignant diagnosis and a corresponding probability of malignancy for each case. Results:The system was able to provide a definitive diagnosis on 46 of the 57 cases. The other eleven were assigned an inconclusive diagnosis by the computer and, therefore, were excluded from further statistical analysis. When compared to the surgical biopsy, 40 (87.0%) of the 46 cases were correctly classified with a sensitivity and specificity of 68.8% and 96.7% respectively. The predictive value of a positive and negative test were 91.7% and 85.3% respectively. Conclusion: When faced with inconclusive diagnoses on FNAs of breast masses, we believe that image analysis can be used as an aid in the further classification of such lesions, thereby, providing a more appropriate triage for surgical biopsy.