Digital mammography: Observer performance study of the effects of pixel size on the characterization of malignant and benign microcalcifications

Rationale and Objectives, The authors performed this study to evaluate the effects of pixel size on the characterization of mammographic microcalcific ations by radiologists. Materials and Methods. Two-view mammograms of 112 microcalcification cluste rs were digitized with a laser scanner at a pixel size of 35 mum Images wit h pixel sizes of 70, 105, and 140 mum were derived from the 35-mum-pixel si ze images by averaging neighboring pixels. The malignancy or benignity of t he microcalcifications had been determined with findings at biopsy or 2-yea r follow-up. Region-of-interest images containing the microcalcifications w ere printed with a laser imager. Seven radiologists participated in a recei ver operating characteristic (ROC) study to estimate the likelihood of mali gnancy. The classification accuracy was quantified with the area under the ROC curve (A(z)). The statistical significance of the differences in the A( z). values for different pixel sizes was estimated with the Dorfman-Berbaum -Metz method and the Student paired t test. The variance components were an alyzed with a bootstrap method. Results. The higher-resolution images did not result in better classificati on; the average A(z) with a pixel size of 35 mum was lower than that with p ixel sizes of 70 and 105 mum The differences in A(z) between different pixe l sizes did not achieve statistical significance. Conclusion. Pixel sizes in the range studied do not have a strong effect on radiologists' accuracy in the characterization of microcalcifications. The low specificity of the image features of microcalcifications and the large interobserver and intraobserver variabilities may have prevented small adv antages in image resolution from being observed.