Mammaglobin vs GCDFP-15: An Immunohistologic Validation Survey for Sensitivity and Specificity

Rohit Bhargava, MD1; Sushil Beriwal, MD2; David J. Dabbs, MD1

Disclosures

Am J Clin Pathol. 2007;127(1):103-113. 

In This Article

Results

Staining with the mammaglobin antibody is characteristically seen in the cytoplasm. The staining intensity in our series ranged from a weak blush to moderate or strong Figure 1. The amount of cells staining with the antibody was further categorized as focal (<10%), patchy (10%-50%), and diffuse (>50%). For meaningful semiquantitative analysis, focal and/or weak staining was considered equivocal staining, and only patchy or diffuse staining with moderate or strong intensity was considered positive.

Invasive breast carcinomas showing weak (A, ×200), moderate (B, ×400), and strong (C, ×400) staining with mammaglobin antibody cocktail.

A higher percentage of breast carcinomas stained with mammaglobin and GCDFP-15 in the whole tissue section group than in the breast TMA group. Although the breast TMA was constructed with 3-fold redundancy, the often patchy pattern of mammaglobin and GCDFP-15 staining partly accounts for this difference. Moreover, the whole tissue section group contained an equal mix of ductal and lobular carcinomas, whereas the 64 TMA cases were randomly selected and contained only 9 lobular carcinomas (14%). Our study shows that mammaglobin and GCDFP-15 stain a higher percentage of cells in lobular carcinoma than in ductal carcinoma. Within the whole tissue section group, the median percentage of cellular staining with mammaglobin was 25 in ductal carcinoma and 75 in lobular carcinoma. Similarly, the median percentage of cellular staining with GCDFP-15 was 5 in ductal carcinoma and 30 in lobular carcinoma. The aforementioned factors are, therefore, also responsible for the apparent discrepancy within the breast whole tissue section and breast TMA groups. Therefore, the results of the whole tissue section breast carcinomas and breast TMA are reported and analyzed separately.

The details of the staining pattern for each case are shown in Table 1 . In a majority of cases, the intensity of staining and proportion of cells stained was significantly higher with the mammaglobin antibody than with GCDFP-15 Figure 2 and Figure 3. Of the 58 breast carcinomas, 40 (69%) showed a higher proportion of cells staining with mammaglobin than with GCDFP-15. In 9 cases (16%), GCDFP-15 stained more cells and in 9 cases (16%), a similar percentage of cells stained with both mammaglobin and GCDFP-15. The median percentage of cellular staining was 55% (mean, 48.7%; SD, 36.3) with mammaglobin compared with only 20% (mean, 25.9%; SD, 25.5) with GCDFP-15. This difference was statistically significant (P < .0001). For the staining intensity, 41 (71%) of 58 cases showed moderate to strong staining with mammaglobin compared with only 28 (48%) with GCDFP-15. This difference was also statistically significant (P=.02).

Invasive, poorly differentiated breast carcinoma (A, H&E, ×200) showing diffuse strong staining with mammaglobin (B, ×400) but only weak positivity for gross cystic disease fluid protein-15 (C, ×400).

Metastatic breast carcinoma (A, H&E, ×400) showing patchy strong staining with mammaglobin (B, ×400) but negativity for gross cystic disease fluid protein-15 (C, ×400).

Based on the intensity and percentage of cells staining, unequivocal positive staining was seen in 41 (71%) of 58 cases with mammaglobin, whereas only 24 cases (41%) showed significant staining with GCDFP-15. The summary of staining patterns (positive, negative, or equivocal) with both antibodies is shown in Table 2 . If equivocal staining is counted with positive staining, the mammaglobin sensitivity for breast carcinoma increases to 93.1% and that of GCDFP-15 approaches 84.5%.

Of the 29 invasive carcinomas, significant staining (patchy or diffuse with moderate to strong intensity) was seen in 20 primary tumors (69%). These included 9 ductal, 10 lobular, and 1 mixed carcinoma. A high degree of concordance was identified between primary and metastatic carcinomas. Overall 27 (93%) of 29 cases showed concordant staining when analyzed for significant staining vs negative or equivocal staining. The 2 discordant cases included 1 ductal ( Table 1 , case 15) and 1 lobular ( Table 1 , case 26) carcinoma. The primary ductal carcinoma showed patchy, strong staining, whereas the tumor in the lymph node showed only focal, weak immunoreactivity. The primary lobular carcinoma showed focal, weak staining, whereas the metastatic tumor showed diffuse, strong positivity.

Mammaglobin positivity was seen in 4 (57%) of 7 grade 1, 8 (80%) of 10 grade 2, and 8 (67%) of 12 grade 3 tumors. The differences in mammaglobin staining with respect to nuclear grades (1 and 2 vs 3) were not statistically significant (P=1). Because the majority of tumors in the study group were positive for estrogen and progesterone receptors, the effect of receptor status on mammaglobin staining could not be studied. Mammaglobin positivity was seen in 15 (68%) of 22 HER-2/neu– tumors and 5 (71%) of 7 HER-2/neu+ tumors. There was no statistically significant difference in mammaglobin staining with respect to HER-2 status (P=1).

Of 64 cases, 63 could be evaluated on TMA with mammaglobin and GCDFP-15 antibodies. The details of staining pattern for each case are shown in Table 3 . Once again, in the majority of cases, the intensity of staining and proportion of cells stained was higher with the mammaglobin antibody than with GCDFP-15. Of the 63 breast carcinomas, 36 (57%) showed a higher proportion of cell staining with mammaglobin than GCDFP-15. In 7 cases (11%), GCDFP-15 stained more cells, and in 20 cases (32%), a similar percentage of cells stained with mammaglobin and GCDFP-15. This difference in the proportion of cellular staining was statistically significant (P < .0001).

Based on intensity and percentage of cells staining, unequivocal positive staining was seen in 26 (41%) of 63 cases with mammaglobin, whereas only 3 cases (5%) showed significant staining with GCDFP-15. This difference in sensitivity was statistically significant (P=.0000001). The summary of staining patterns (positive, negative, or equivocal) with both antibodies is shown in Table 4 . If equivocal staining is counted with positive staining, the mammaglobin sensitivity for breast carcinoma increases to 58.7% and that of GCDFP-15 approaches 19%, and the difference was again statistically significant (P=.000003).

Mammaglobin positivity was seen in 10 (56%) of 18 grade 1, 8 (32%) of 25 grade 2, and 8 (40%) of 20 grade 3 tumors. The differences in mammaglobin staining with respect to Nottingham grades (1 vs 2 vs 3) were not statistically significant (grade 1 vs 2, P=.20; grade 2 vs 3, P=.75; grade 1 vs 3, P=.51; grade 1 vs others, P=.16). Mammaglobin staining with respect to other pathologic variables is summarized in Table 5 .

All tissue cores (tumor and normal) in the lung, stomach, colon, kidney, and bladder arrays were negative. All tumors (squamous cell carcinomas) in the uterine cervical tissue array were also negative, but 13 of 15 cores with normal endocervical glands showed weak (equivocal) to moderate staining. Positive staining was seen in endometrial carcinoma Figure 4A, sweat gland carcinoma Figure 4B, salivary gland, melanoma, ovarian, and pancreatic tissue arrays.

Mammaglobin staining in nonbreast tumors. A, Endometrial endometrioid adenocarcinoma shows diffuse strong staining (×200). B, Sweat gland carcinoma showing positive mammaglobin staining (×200).

The results for nonbreast tumors are given in Table 6 .

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.

processing....