Sonographic correlations with the new molecular classification of invasive breast cancer.

Ultrasound is a useful adjunct to mammography for the characterisation and biopsy of solid breast lesions. Protein expression profiling of breast cancer has identified specific subgroups with potential clinical, biological and therapeutic implications. The aim of this study was to determine the ultrasound correlates of these novel molecular classes of invasive breast cancer. The ultrasound findings in 358 patients with operable breast cancer were correlated with the previously described protein expression classes identified by our group using immunohistochemical (IHC) assessment of a large series of breast cancer cases in which 25 proteins of known relevance in breast cancer were assessed, including hormone receptors, HER2 status, basal and luminal markers, p53 and e-cadherin. The proportion of occult lesions was not significantly different in the two groups. Significant differences were noted between the two groups expressing luminal epithelial markers and hormone receptors (1 and 2), including a greater proportion of ill-defined, irregular and distally attenuating tumours in group 2. Tumours characterised by c-erbB2/MUC1 expression, with weak hormone receptor positivity (group 3) were also more likely to be ill defined. Tumours expressing basal markers (group 5) were less likely to have an echogenic halo. The ultrasound features of breast cancer show areas of significant correlation with molecular classes of invasive breast cancer identified by IHC analysis. The biological reasons for these findings and their implications regarding imaging protocols require further study and may enable improved detection of these lesions.

Is acinic cell carcinoma a variant of secretory carcinoma? A FISH study using ETV6'split apart' probes.

AIMS: Acinic cell carcinomas (ACCs) and secretory carcinomas (SCs) of the breast are rare, low-grade malignancies that preferentially affect young female patients. Owing to the morphological and immunohistochemical similarities between these lesions, they have been proposed to be two morphological variants of the same entity. It has been demonstrated that SCs of the breast consistently harbour the t(12;15)ETV6-NTRK3 translocation. The aim was to determine whether ACCs also harbour ETV6 gene rearrangements and are thus variants of SCs. METHODS AND RESULTS: Using the ETV6 fluorescence in situ hybridization DNA Probe Split Signal (Dako), the presence of ETV6 rearrangements in three SCs and six ACCs was investigated. Cases were considered as harbouring an ETV6 gene rearrangement if >10% of nuclei displayed 'split apart signals' (i.e. red and green signals were separated by a distance greater than the size of two hybridization signals). Whereas the three SCs displayed ETV6 split apart signals in >10% of the neoplastic cells, no ACC showed any definite evidence of ETV6 gene rearrangement. CONCLUSIONS: Based on the lack of ETV6 rearrangements in ACCs, our results strongly support the concept that SCs and ACCs are distinct entities and should be recorded separately in breast cancer taxonomy schemes.

The influence of basal phenotype on the metastatic pattern of breast cancer.

AIMS: To assess whether basal phenotype influences the metastatic pattern and survival in patients with metastatic breast cancer. MATERIALS AND METHODS: The basal phenotype status of a well-characterised series of consecutive primary operable breast cancers (1868 cases) was ascertained using the basal cytokeratin markers CK5/6 and CK14. Follow-up data, including time, site and pattern of distant metastasis and post-metastasis survival, were available for 113 women with basal phenotype cancers and they were compared with 178 matching cases from women in the non-basal phenotype group. RESULTS: Patients with basal phenotype were more likely to present with intrapulmonary (25/48, [52%] vs 15/64, [23%]; P=0.0009) and/or brain metastases (20/113, [18%] vs 3/178, [2%]; P<0.0001) than non-basal phenotype patients. Patients with non-basal phenotype were more likely to present with bone metastases in the absence of visceral disease (48/102, [47%] vs 14/62, [23%]; P=0.0017) than patients with basal phenotype. There was no significant difference in the frequency of pleural or liver metastases between both groups. Basal phenotype was also associated with a shorter median survival with metastatic disease (10.1 months vs 25 months, P<0.001). The multivariate analysis, including other established prognostic variables in breast cancer, showed that basal phenotype is an independent poor prognostic factor. CONCLUSION: Intrapulmonary and brain metastases are seen more frequently at metastatic presentation in basal phenotype breast cancer patients, and the basal phenotype is associated with a poorer survival after metastatic presentation. Assessment of basal cytokeratin expression status may provide valuable prognostic information relevant to breast cancer patients' management.

Basal phenotype: a powerful prognostic factor in small screen-detected invasive breast cancer with long-term follow-up.

OBJECTIVE: To assess the frequency and prognostic significance of a basal phenotype in a group of women with screen-detected invasive breast cancers with long-term follow-up and to focus particularly on women with small ( < 15 mm) breast cancers. METHODS: The study group was derived by finding women common to a consecutive series of 1944 invasive breast cancers diagnosed in Nottingham between 1986 and 1998 with a known basal phenotype status and a prospectively collected database of all screen-detected breast cancers. In total, 356 women constituted the study group. Pathological and radiological features were recorded. Basal cell markers used were CK5/6 (cloneD5/16134) and CK14 (clone LL002). Tumours were classified as of basal phenotype if > or = 10% staining was seen with either marker. RESULTS: Of all screen-detected lesions, 43 (12%) had a basal immunophenotype and 313 (88%) were non-basal. There were 15 (35%) and 40 (13%) breast cancer deaths in the basal group and nonbasal groups, respectively ( P = 0.0006). On univariate analysis, nodal stage, histological grade, lympho-vascular invasion (LVI) status, invasive size and basal phenotype had prognostic significance. On multivariate analysis, basal phenotype, LVI and nodal stage maintain prognostic significance. Of the 189 women with < 15 mm lesions, eight of 20 (40%) of the basal group and eight of 169 (5%) of the non-basal group died of breast cancer ( P < 0.0001). On multivariate analysis, basal phenotype was the only factor to maintain independent prognostic significance. CONCLUSIONS: Basal phenotype is a powerful prognostic factor for women with small screen-detected invasive breast cancer.

Morphological and immunophenotypic analysis of breast carcinomas with basal and myoepithelial differentiation.

The aim of this study was to assess the morphological characteristics and immunohistochemical profile of breast carcinomas with basal and myoepithelial phenotypes to obtain a better understanding of their biological behaviour and nature. One thousand nine hundred and forty-four invasive breast carcinomas were examined, using tissue microarray (TMA) technology and immunohistochemistry, to identify those tumours that showed basal and myoepithelial phenotypes, and their immunophenotype profile was characterized using a variety of markers. In addition, haematoxylin and eosin-stained sections of these tumours were studied for several morphological parameters. The findings were correlated with patient and tumour characteristics and outcome data. Tumours were classified into two groups: (1) tumours with basal phenotype [expressing one or both basal markers (CK5/6 and/or CK14)] and (2) tumours with myoepithelial phenotype (expressing SMA and/or p63). Group 1 was further subdivided into two subgroups: (A) dominant basal pattern (more than 50% of cells positive) and (B) basal characteristics (10-50% of cells positive). Group 1 tumours constituted 18.6% (8.6% and 10% for groups 1A and 1B, respectively) and group 2 constituted 13.7% of the cases. In both groups, the most common histological types were ductal/no specific type, tubular mixed and medullary-like carcinomas; the majority of these tumours were grade 3. There were positive associations with adenoid cystic growth pattern, loss of tubule formation, marked cellular pleomorphism, poorer Nottingham prognostic index, and development of distant metastasis. In addition, associations were found with loss of expression of steroid hormone receptors and FHIT proteins and positive expression of p53 and EGFR. The most common characteristics in group 1 were larger size, high-grade comedo-type necrosis, development of tumour recurrence, and absence of lymph node disease. Group 2 tumours were more common in younger patients and were associated with central acellular zones, basaloid change, and positive E-cadherin protein expression. Group 1 characteristics were associated with both reduced overall survival (OS) [log rank (LR) = 22.5, p < 0.001] and reduced disease-free interval (DFI) (LR = 30.1, p < 0.001), while group 2 characteristics showed an association with OS (LR = 5, p = 0.02) but not with DFI. Multivariate analysis showed that basal, but not myoepithelial, phenotype has an independent value in predicting outcome. Breast cancers with basal and myoepithelial phenotypes are distinct groups of tumours that share some common morphological features and an association with poor prognosis. The basal rather than the myoepithelial phenotype has the strongest relationship with patient outcome.

Expression and co-expression of the members of the epidermal growth factor receptor (EGFR) family in invasive breast carcinoma.

The epidermal growth factor receptor (EGFR) family plays an important role in breast carcinogenesis. Much interest has been focused recently on its members because of their potential role as prognostic indicators in breast cancer and their involvement in cancer therapy. We have evaluated more than 1500 cases of invasive breast carcinoma immunohistochemically using tissue microarray technology to examine the expression of EGFR family receptor proteins. We have found that 20.1 and 31.8% of cases were positive for EGFR and c-erbB-2, respectively, and 45 and 45.1% of tumours overexpressed for c-erbB-3 and c-erbB-4, respectively. The expression of either EGFR or c-erbB-2 was associated with other bad prognostic features and with poor outcome. Neither c-erbB-3 nor c-erbB-4 had any association with survival. c-erbB-2 had an independent prognostic effect on overall and disease-free survival (DFS) in all cases, as well as in the subset of breast carcinoma patients with nodal metastases. Several hetero- and homodimeric combinations have been reported between the EGFR members. Those dimers can evoke diverse signal transduction pathways with variable cellular responses. We stratified cases according to their co-expression of receptors into distinct groups with different receptor-positive combinations. Patients whose tumours co-expressed c-erbB-2 and c-erbB-3, as well as those whose tumours co-expressed EGFR, c-erbB-2 and c-erbB-4 showed an unfavourable outcome compared with other groups, while combined c-erbB-3 and c-erbB-4 expression was associated with a better outcome. In cases showing expression of one family member only (homodimers), we found a significant association between c-erbB-4 homodimer-expressing tumours and better DFS. In contrast, patients with c-erbB-2 homodimer-expressing tumours had a significant poorer DFS compared with other cases. These data imply that the combined profile expression patterns of the four receptor family members together provide more accurate information on the tumour behaviour than studying the expression of each receptor individually.

Expression of the transcription factor CTCF in invasive breast cancer: a candidate gene located at 16q22.1.

CTCF is a ubiquitous 11-zinc-finger protein that plays a role in gene silencing or activation, chromatin insulation and genomic imprinting. The CTCF gene has been mapped to the chromosome band 16q22.1 that shows frequent loss of heterozygosity in breast cancer. The E-cadherin gene is the known tumour suppressor gene (TSG) at this region in lobular carcinomas; however, the target gene in the more frequent ductal tumours is still unknown. Since CTCF targets include TSGs and oncogenes and it has the ability to inhibit cell growth and proliferation, it has been suggested that it may be the target gene at the 16q22.1 in ductal carcinomas. In the present study, tissue microarray technology was used to study the expression pattern of CTCF immunohistochemically in 344 cases of invasive breast carcinoma and its expression was correlated with clinicopathological variables and patient outcome. Results showed that breast tissues express CTCF in the parenchymal cells of the normal ducts and lobules but with a variable percentage of positive cells. Staining of CTCF was detected in the nuclei and cytoplasm of the malignant cells, but no significant loss or decrease of expression was noticed in association with any specific tumour type. There was a significant correlation between expression of CTCF and histological grades; lower expression was associated with grade 3 tumours. Cytoplasmic expression was associated with increased tumour size and with the presence of vascular invasion. However, no association was found between CTCF expression and tumour type, lymph node stage, oestrogen receptor expression or patient outcome. In conclusion, the current results show that CTCF, although it may play a role in breast carcinogenesis, is unlikely to be the TSG targeted by the 16q22.1 loss in breast cancer and thus another gene or genes at this region remain to be identified.

Loss of CD59 expression in breast tumours correlates with poor survival.

CD59 (protectin), a phosphatidylinositol-anchored glycoprotein, is a member of the cell membrane-bound complement regulatory proteins that inhibits the formation of the terminal membrane attack complex (MAC) of complement. In this study, the expression of CD59 was evaluated in 520 breast carcinomas from patients with a mean follow-up of 87 months. This expression was correlated with clinicopathological features and patient survival. Marked variation in the intensity of CD59 expression, which correlated with histological grade and Nottingham prognostic index (NPI), was found, with higher expression of CD59 found more often in well and moderately differentiated tumours and those of good prognosis (NPI < or = 3.4). In contrast, high grade and poor prognosis (NPI > 5.4) carcinomas significantly demonstrated lack of CD59 expression (p < 0.001). Moreover, it was found that the percentage of CD59-positive cells correlated significantly with patient survival, ie patients with a high percentage of positive cells (>50%) had a better overall survival (p = 0.006). A correlation was also found between the percentage of CD59-positive cells and tumour type and also the development of distant metastases. No association was found between either the intensity or the percentage of cells expressing CD59 and vascular invasion, lymph node stage, tumour size, patient age or menopausal status. In multivariate analysis, CD59 percentage positivity was of independent prognostic significance with grade and lymph node stage. These findings indicate that loss of CD59 may offer a selective advantage for breast cancers, resulting in more aggressive tumours and conferring a poor prognosis for patients.

Correlation between immunohistochemistry (HercepTest) and fluorescence in situ hybridization (FISH) for HER-2 in 426 breast carcinomas from 37 centres.

Accurate diagnostic assessment of HER-2 is essential for the appropriate application of the humanized anti-HER-2 monoclonal antibody trastuzumab (Herceptin) to the treatment of patients with metastatic breast cancer. The diagnostic test needs to be applicable to archival, fixed tissue removed at excision, in many cases several years earlier. We compared the assessment of HER-2 by immunohistochemistry (IHC; HercepTest) and fluorescence in situ hybridization (FISH) in 426 breast carcinomas from patients being considered for trastuzumab therapy. The tumours were tested in three reference centres having been sent in from 37 hospitals. Only 2/270 (0.7%) IHC 0/1+ tumours were FISH positive. Six of 102 (5.9%) IHC 3+ tumours were FISH negative. Five of the six had between 1.75 and 2.0 HER-2 gene copies per chromosome 17 and the sixth had multiple copies of chromosome 17. Thirteen per cent of tumours were IHC 2+ and overall 48% of these were FISH positive but this proportion varied markedly between the centres. Sixty IHC-stained slides selected to be enriched with 2+ cases were circulated between the three laboratories and scored. There were 20 cases in which there was some discordance in scoring. Consideration of the FISH score in these cases led to concordance in the designation of positivity/negativity in 19 of these 20 cases. These data support an algorithm in which FISH testing is restricted to IHC 2+ tumours in reference centres. The results may not extrapolate to laboratories with less experience or using different methodologies.

The role of blood tumor marker measurement (using a biochemical index score and c-erbB2) in directing chemotherapy in metastatic breast cancer.

The role of blood tumor markers in monitoring response in advanced breast cancer is established in endocrine therapy and standard chemotherapy. This study examines marker levels in patients receiving new chemotherapy regimens. Thirty patients were recruited into two multicenter trials in which docetaxel-based regimens were used in 15 patients. The other 15 received doxorubicin-based regimens. Biochemical response calculated from a score using CA15.3, CEA and ESR was compared with UICC response. Marker changes at 2, 4 and 5 months correlated with UICC response at 3, 4(1/2) and 6 months, respectively (p < 0.03). Eleven patients achieved both clinical/radiological and biochemical response at the end of treatment; markers had not yet returned to below cutoffs in seven, suggesting a possible advantage to continue chemotherapy. No patient showed a biochemical response whilst judged clinically/radiologically progressive. Nineteen patients had progressed either clinically/radiologically or biochemically at six months; of these, eight showed progression assessed earlier by markers so that a median of four cycles of chemotherapy could have been saved. Measurements of serum c-erbB2 showed a correlation with tissue c-erbB2 staining in the primary tumor (p < 0.003). Among the patients with positive tissue staining, sequential changes in serum c-erbB2 completely paralleled initial response.