Comparison of Spot Digital and Conventional Mammography in the Evaluation of Microcalcifications

KKT Pak, JLF Chiu, CS Kwok, SC Wong, SY Yiu, SCH Chan

Hong Kong J Radiol 2003;6:195-200

Objective: To compare the diagnostic accuracy of spot digital and conventional screen-film mammography and the influence of clinical data on the characterisation of microcalcifications.

Patients and Methods: Twenty nine patients with 30 groups of microcalcifications were studied. Two radiologists assessed the biopsy-proven calcifications on conventional screen-film cone-magnification views and digital spot views. Ratings on breast density, visibility, and likelihood of malignancy were estimated on a 4-point confidence scale individually by each radiologist. Both radiologists were asked to give a second alignancy rating for conventional screen-film mammograms with patients’ clinical data. Receiver operating characteristic methodology was applied to evaluate the results. The classification accuracy was quantified by area under the receiver operating characteristic curve. Statistical differences in the area under the receiver operating characteristic curve values for the effects of digital mammography and clinical data were estimated.

Results: The average area under the receiver operating characteristic curve values were higher for evaluation of microcalcifications on conventional screen-film images. Prior knowledge of patient’s clinical data did not improve the diagnostic accuracy on conventional screen-film mammography. But the differences in area under the receiver operating characteristic curve values for both comparisons were not statistically significant (p > 0.05).

Conclusion: The diagnostic accuracy of spot digital is comparable to that of cone magnification views on conventional screen-film mammography. The effects of intra- and inter-reader variability should be considered. Consistency in malignancy rating based on lesion characterisation may be affected by variation in interventional threshold among the readers.