Faculty Bibliography

We present the case of a 73-year-old woman with pure squamous cell carcinoma in situ of the left breast. This non-invasive malignancy exhibited pure squamous type of architectural and cytologic features without any evidence of glandular differentiation either in the initial needle core biopsy or in the subsequently performed excisional biopsy and simple mastectomy. The tumor spanned 1.6 cm, involved numerous ducts and terminal ducts and extended into lobules, and was characterized by keratinizing squamous cells with intermediate-grade nuclei. Intercellular bridges extended between the malignant squamous cells. Keratinous debris with "pearl" formation was evident in most involved glands. No invasive carcinoma was identified. There was no evidence of metastatic disease in the ipsilateral sentinel lymph nodes. Thus far, only three cases of squamous cell carcinoma in situ of the breast have been reported in one series-none of which showed any evidence of recurrent or metastatic disease at last follow-up. In our case, treated exclusively by surgery, there was no evidence of disease 11 years after diagnosis.

CONTEXT/BACKGROUND:-Small glandular proliferations of the breast encompass a variety of benign, atypical, and malignant lesions that show some overlapping morphologic features. Myoepithelial stains are frequently used in the workup of these lesions in order to rule out or establish a diagnosis of invasive carcinoma. Some benign lesions show absent or diminished myoepithelial staining, and may represent an interpretative pitfall, particularly in small core biopsy samples. OBJECTIVE:-To review small glandular proliferations of the breast that show absent or diminished staining with myoepithelial immunohistochemical markers. DATA SOURCES/METHODS:-The study comprised a review of published literature and clinical case material. CONCLUSIONS:-The interpretation of myoepithelial stains in small glandular proliferations of the breast can, on some occasions, represent a challenge in diagnosing these lesions. Recognition of the key histopathologic features and immunohistochemical staining patterns of the entities in the differential diagnosis is crucial in their workup.

Classification of mammary fibroepithelial tumors (FETs) relies on assessment of mitotic activity, among other histopathologic parameters. Routine hematoxylin and eosin (H&E) mitotic counts can be subjective and time consuming. Difficulty may arise in identifying "true" mitoses for a variety of reasons. Phosphorylation of histone H3 protein (PHH3) is correlated with mitotic chromatin condensation. The utility of PHH3 immunohistochemical staining to identify mitoses has been demonstrated in multiple organ systems. In this study, we examined the utility of PHH3 in assessing mitotic activity in FETs and compared PHH3- with H&E-determined mitotic counts. PHH3-stained mitoses were readily identifiable at ×10 magnification and allowed for rapid identification of mitotic "hot spots." Median mitotic counts/10 high-power fields for fibroadenoma, benign phyllodes tumor, borderline phyllodes tumor (BlnPT), and malignant phyllodes tumor (MPT) were 0, 0.5, 4.25, and 9, respectively on H&E, and 0, 0.75, 4.5, and 8, respectively for PHH3. Among all FETs, there was a strong positive correlation between H&E- and PHH3-determined mitotic counts (r=0.91, P<.001). Using PHH3, 2 cases would be reclassified, both from BlnPT to MPT. PHH3-determined counts correlated with H&E-determined counts in FETs. Using PHH3, a small number of cases were reclassified from BlnPT to MPT, for which treatment is similar. Although H&E-determined counts remain the criterion standard for assessing mitotic activity in FETs, PHH3 may be a useful adjunctive tool in some cases and is helpful in identifying mitotic hot spots.

OBJECTIVES/OBJECTIVE:The Minimal Carcinoma (MC) Triple Stain is a tri-chromogen multiplex immunostain (CK7, p63, and E-cadherin) helpful in classifying morphologically ambiguous and/or small carcinomas as either ductal or lobular and/or in situ or invasive. We compared the utility of this stain with two commercially available duplex/multiplex immunostains: Breast Triple Stain (BTS) (Clarient, Aliso Viejo, CA; CK5, p63, and CK8/18) and LC/DC Breast Cocktail (LCDC) (Biocare, Concord, CA; E-cadherin and p120). METHODS:Ninety-seven mammary carcinomas stained with the MC Triple Stain, BTS, and LCDC were compared. RESULTS:The MC Triple Stain, LCDC, and BTS were diagnostic in 90 (93%) of 97, 82 (85%) of 97, and 85 (88%) of 97 of cases, respectively. All stains showed decreased diagnostic utility due to variability in tissue integrity, quality of the staining, and/or ease of interpretation. In cases where all immunostains were interpretable, the MC Triple Stain yielded the most information. CONCLUSIONS:When technically sufficient, all three immunostains demonstrated relative strengths and weaknesses in their ability to provide diagnostic information with the highest consistency and ease of use. Many cases stained with LCDC were technically insufficient due to a suboptimal staining protocol provided by the company. Overall, the MC Triple Stain outperformed BTS and LCDC by more consistently providing more diagnostic information. The MC Triple Stain is a viable alternative to other multiplex immunostains in evaluating small foci of carcinoma, particularly when both the histologic type and extent of disease (in situ vs invasive) require clarification.

High grade endometrial stromal sarcoma (HGESS) is an uncommon malignancy recently re-defined in the new WHO classification of endometrial stromal tumors. In this article, we discuss the differential diagnoses of metastatic HGESS in a fine needle aspiration (FNA) of a lymph node and compare the cytomorphology of HGESS in ThinPrep [(TP), Hologic Inc., Boxborough, MA] to conventional smears (CS). The patient had a history of stage I HGESS, status-post supracervical hysterectomy without regional lymph node metastases. Her post-operative course was complicated by pelvic fluid collections and enlarging para-aortic lymph nodes. Diff-Quik (DQ)-stained and Papanicolaou (Pap)-stained smears from a para-aortic lymph node FNA demonstrated a cellular specimen with monomorphic population of plump to oval cells with scant, wispy cytoplasm in aggregates and as single cells. The nuclei showed fine chromatin and small inconspicuous nucleoli. Compared to the CS, HGESS cells in the TP showed similar cytological features, with the exception that the nuclei were slightly smaller, hyperchromatic, and the chromatin pattern was attenuated. In the absence of prior clinical history, the cytomorphology of metastatic HGESS in a lymph node can be difficult to differentiate from a lymphoma, a variety of metastatic spindle cell tumors or metastatic carcinoma. Immunohistochemical analysis and comparison with the primary tumor can be useful in proving the nature of the malignant cells. The cytomorphology of HGESS on TP correlated well in both single cells and aggregates when compared to CS. The differences noted were decreased nuclear size, nuclear hyperchromasia, and slightly attenuated nuclear detail on TP.

OBJECTIVES: The distinction between metastatic breast carcinomas (BCs) and primary lung carcinomas (PLCs) can be difficult. This study tested the utility of trefoil factor 1 (TFF1) for this purpose and compared it with mammaglobin and GATA protein binding 3 (GATA-3). METHODS: Tissue microarrays containing 365 BCs and 338 PLCs were stained with TFF1, mammaglobin, and GATA-3, and an H-score was calculated. Sensitivity, specificity, and accuracy were calculated, and logistical regression analysis was performed. RESULTS: Accuracy of correctly classifying the tumor type was 81.9%, 71.3%, and 64.0% for GATA-3, mammaglobin, and TFF1, respectively. Odds ratios for selecting BCs were 25.69, 93.15, and 4.17, respectively, with P values less than .001. With a single exception, the best immunopanel included GATA-3 and mammaglobin in all comparisons. CONCLUSIONS: TFF1 demonstrated breast specificity but was inferior to mammaglobin and GATA-3. Therefore, its routine clinical use may not be justified. TFF1 showed little benefit when added to an immunopanel.

Inflammatory and reactive lesions of the breast are relatively uncommon among benign breast lesions and can be the source of an abnormality on imaging. Such lesions can simulate a malignant process, based on both clinical and radiographic findings, and core biopsy is often performed to rule out malignancy. Furthermore, some inflammatory processes can mimic carcinoma or other malignancy microscopically, and vice versa. Diagnostic difficulty may arise due to the small and fragmented sample of a core biopsy. This review will focus on the pertinent clinical, radiographic, and histopathologic features of the more commonly encountered inflammatory lesions of the breast that can be characterized in a core biopsy sample. These include fat necrosis, mammary duct ectasia, granulomatous lobular mastitis, diabetic mastopathy, and abscess. The microscopic differential diagnoses for these lesions when seen in a core biopsy sample will be discussed.

A case of a 53-year-old woman with the epithelioid variant of mammary myofibroblastoma, which was initially misinterpreted as invasive lobular carcinoma, is presented. A needle core biopsy of the 1.6 cm mass showed interlacing bundles of epithelioid myofibroblasts amid dense fibrous tissue associated with lobular carcinoma in situ of the classical type. Most epithelioid cells showed nuclear atypia, and a few exhibited signet-ring cytology. Immunoreactivity for estrogen and progesterone receptors further compounded the deception, and the neoplasm was misinterpreted as invasive lobular carcinoma. Excisional biopsy showed a circumscribed stromal tumor with foci suspicious for invasive lobular carcinoma. The latter was excluded by cytokeratin negativity throughout the tumor. The overall histopathological appearance and immunostaining pattern was confirmatory of myofibroblastoma. This case report emphasizes the potential for mistaking epithelioid myofibroblastoma for invasive lobular carcinoma--particularly in the setting of limited sampling, hormone-receptor immunoreactivity of the lesional cells, and synchronous lobular carcinoma in situ.