Please select Print from the file menu or right mouse click and choose print from the menu to print this page.
[144] Measuring Extent of DCIS in Breast Excision Specimens: A Comparison of Four Methods
AN Grin, G Horne, M Ennis, FP O'Malley. Mount Sinai Hospital, Toronto, ON, Canada; University of Calgary, Calgary, AB, Canada; Applied Statistician, Markham, ON, Canada
Background: Measuring the extent of non-palpable DCIS in a breast specimen is challenging. An accurate assessment of the size of DCIS is important as it influences patient management. Several methods for estimating the extent of DCIS have been used, but none have been validated. The aim of this study was to compare the extent of DCIS using various methods of estimation.
Design: 78 primary breast excisions with DCIS and an accompanying sliced specimen radiograph were retrospectively reviewed. All specimens had been sampled using the serial sequential method, which involved mapping the location of each block on the sliced specimen radiograph and calculating extent through 3-D reconstruction. This method was considered the gold standard. The other methods included: 1) calculating the extent based on sampling only areas of calcification (calcification method); 2) recording the number of blocks involved by DCIS and multiplying this by 0.3 cm (blocks method); 3) measuring the largest extent of DCIS on a single slide (single slide method). Data was analysed by characterizing the under- and overestimation relative to the serial sequential method, and by calculating the percentage correctly classified into size categories.
Results: There were 4 grade I (5.1%), 40 grade II (51.3%) and 34 grade III (43.6%) cases. By the serial sequential method, the mean number of blocks submitted was 26.7 (SD 11.0). All three alternative methods tended to underestimate the DCIS, with median underestimates of zero when extent (according to the serial sequential method) was <0.5cm, increasing to 1.2cm, 1.6cm, and 3.7cm for the calcification, blocks and single slide methods when extent was >4cm. Table 1 shows the number of cases in each clinically significant size category by the serial sequential method with comparison to the number (%) of cases classified into the same group by alternate methods.
Table 1 | Serial sequential | Calcification method | Blocks method | Single slide method | | <0.5cm | 4 | 4 (100) | 4 (100) | 4 (100) | | 0.5-2cm | 34 | 28 (82.4) | 28 (82.4) | 28 (82.4) | | >2-4cm | 26 | 19 (73.1) | 14 (53.8) | 1 (3.8) | | >4cm | 14 | 9 (64.3) | 8 (57.1) | 0 (0) | | Total | 78 | 65 (83) | 54 69) | 33 (42) |
Conclusions: When DCIS is <0.5cm, all methods of measuring extent are equivalent. For sizes >0.5cm the alternative methods tended to underestimate the DCIS extent relative to the serial sequential method with the underestimation becoming more pronounced as size increased. These size discrepancies are of clinical relevance.
Category: Breast
Monday, March 3, 2008
Poster # 28, Monday Morning
|