Radiological over- and underestimation of DCIS extent in breast conserving surgery specimens: radiological-pathological correlation in slice radiographs and respective histological large-format slides

Abstract

PURPOSE: Disease extent is an independent risk factor for local recurrence in ductal carcinoma in situ (DCIS) of the breast. Surgical strategies mainly depend on area size of microcalcifications in mammograms which, however, may differ considerably from final DCIS size found histopathologically. Our aim was to analyse the extent and causes of discrepancies between radiological and pathological DCIS size in breast conserving surgery (BCS) specimens. METHODS: The slices of the BCS specimens were radiographed in the department of radiology using digital technique without magnification. Microcalcifications area size in 276 slice radiographs of 46 BCS specimens was measured by two radiologists and compared to histological DCIS size in the large format slides of the matching slices. Size differences >5 mm were considered discrepant. RESULTS: 216 BCS specimen slices contained microcalcifications, 228 slices contained DCIS. Positive predictive value of microcalcifications for DCIS was 0.77, negative predictive value was 0.86. Overestimation of DCIS size in the slice radiograph occurred in 58 (35%) and underestimation in 67 (40%) of 168 slices containing both DCIS and microcalcifications. Discrepancy rates of 6-9, 10-15, 16-19 and ≥20 mm between radiological and pathological DCIS extent were 20.8%, 14.9%, 8.9% and 19% for 168 slices containing DCIS and microcalcifications, respectively. CONCLUSION: More than two thirds of DCIS sizes were radiologically either over- or underestimated in the slice radiographs. In 89.1% of the analysed 46 BCS specimens, discrepancies between radiological size of microcalcification area and histopathological DCIS extent did not exceed 15 mm. This should be considered in planning of BCS.