Astronomical algorithms for automated analysis of tissue protein expression in breast cancer.

BACKGROUND: High-throughput evaluation of tissue biomarkers in oncology has been greatly accelerated by the widespread use of tissue microarrays (TMAs) and immunohistochemistry. Although TMAs have the potential to facilitate protein expression profiling on a scale to rival experiments of tumour transcriptomes, the bottleneck and imprecision of manually scoring TMAs has impeded progress. METHODS: We report image analysis algorithms adapted from astronomy for the precise automated analysis of IHC in all subcellular compartments. The power of this technique is demonstrated using over 2000 breast tumours and comparing quantitative automated scores against manual assessment by pathologists. RESULTS: All continuous automated scores showed good correlation with their corresponding ordinal manual scores. For oestrogen receptor (ER), the correlation was 0.82, P<0.0001, for BCL2 0.72, P<0.0001 and for HER2 0.62, P<0.0001. Automated scores showed excellent concordance with manual scores for the unsupervised assignment of cases to 'positive' or 'negative' categories with agreement rates of up to 96%. CONCLUSION: The adaptation of astronomical algorithms coupled with their application to large annotated study cohorts, constitutes a powerful tool for the realisation of the enormous potential of digital pathology.

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.

Associations between genomic stratification of breast cancer and centrally reviewed tumour pathology in the METABRIC cohort.

The integration of genomic and transcriptomic profiles of 2000 breast tumours from the METABRIC [Molecular Taxonomy of Breast Cancer International Consortium] cohort revealed ten subtypes, termed integrative clusters (IntClust/s), characterised by distinct genomic drivers. Central histopathology (N = 1643) review was undertaken to explore the relationship between these ten molecular subtypes and traditional clinicopathological features. IntClust subtypes were significantly associated with histological type, tumour grade, receptor status, and lymphocytic infiltration (p < 0.0001). Lymph node status and Nottingham Prognostic Index [NPI] categories were also significantly associated with IntClust subtype. IntClust 3 was enriched for tubular and lobular carcinomas, the latter largely accounting for the association with CDH1 mutations in this cluster. Mucinous carcinomas were not present in IntClusts 5 or 10, but did not show an association with any of the remaining IntClusts. In contrast, medullary-like cancers were associated with IntClust 10 (15/26). Hormone receptor-positive tumours were scattered across all IntClusts. IntClust 5 was dominated by HER2 positivity (127/151), including both hormone receptor-positive (60/72) and hormone receptor-negative tumours (67/77). Triple-negative tumours comprised the majority of IntClust 10 (132/159) and around a quarter of IntClust 4 (52/217). Whilst the ten IntClust subtypes of breast cancer show characteristic patterns of association with traditional clinicopathological variables, no IntClust can be adequately identified by these variables alone. Hence, the addition of genomic stratification has the potential to enhance the biological relevance of the current clinical evaluation and facilitate genome-guided therapeutic strategies.

Reduction of E-cadherin expression is associated with non-lobular breast carcinomas of basal-like and triple negative phenotype.

AIM: E-cadherin inactivation in breast cancer has been shown to be strongly associated with lobular breast cancer. However, little is known about the levels of E-cadherin expression according to the breast cancer "molecular" subtypes. The aim of this study was to address the distribution of E-cadherin expression according to the different molecular subtypes of breast cancer. METHODS: E-cadherin expression was immunohistochemically analysed in a tissue microarray containing duplicate cores of 245 invasive breast carcinomas, of which 182 cases were of non-lobular histology, using a semi-quantitative scoring system based on the percentage of cells showing membrane immunopositivity. RESULTS: In non-lobular breast carcinomas, reduced and/or negative E-cadherin expression was significantly associated with lack of oestrogen receptor expression, low levels of CCND1 expression, positivity for cytokeratins 5/6 and 17, epidermal growth factor receptor and caveolins 1 and 2, p53 expression, high MIB-1 proliferation indices, basal-like phenotype and triple negative phenotype. CONCLUSION: This study demonstrates that in the group of non-lobular breast cancers, reduction/lack of E-cadherin expression is preferentially found in basal-like breast carcinomas.

APC mutations in sporadic medulloblastomas.

The cerebellar medulloblastoma (WHO Grade IV) is a highly malignant, invasive embryonal tumor with preferential manifestation in children. Several molecular alterations appear to be involved, including isochromosome 17q and the p53, PTCH, and beta-catenin gene mutations. In this study, 46 sporadic medulloblastomas were screened for the presence of mutations in genes of the Wnt signaling pathway (APC and beta-catenin). Single-strand conformational polymorphism (SSCP) analysis followed by direct DNA sequencing revealed 3 miscoding APC mutations in 2 (4.3%) medulloblastomas. One case contained a GCA-->GTA mutation at codon 1296 (Ala-->Val), and another case had double point mutations at codons 1472 (GTA-->ATA, Val-->Ile) and 1495 (AGT-->GGT, Ser-->Gly). Miscoding beta-catenin mutations were detected in 4 tumors (8.7%). Three of these were located at codon 33 (TCT -->TTT, Ser-->Phe) and another at codon 37 (TCT-->GCT, Ser-->Ala). Adenomatous polyposis coli (APC) gene and beta-catenin mutations were mutually exclusive and occurred in a total of 6 of 46 cases (13%). Although germline APC mutations are a well established cause of familial colon and brain tumors (Turcot syndrome), this study provides the first evidence that APC mutations are also operative in a subset of sporadic medulloblastomas.

Beta-catenin mutations are frequent in human hepatocellular carcinomas associated with hepatitis C virus infection.

Hepatocellular carcinoma (HCC) is one of the most common fatal cancers worldwide. Hepatitis B virus and hepatitis C virus infections, exposure to aflatoxin, and excessive intake of alcohol have been identified as major risk factors. However, the molecular mechanisms underlying their development are still poorly understood. Recently, beta-catenin, one of the key components of the Wnt signaling pathway, has been found to be mutated in about 20% of HCCs, suggesting a role of the Wnt pathway in their development. In this study, we examined beta-catenin and APC mutations in 22 HCCs associated with HCV infection, using single-strand conformation polymorphism (SSCP) followed by direct DNA sequencing. beta-Catenin mutations were found in nine (41%) cases, but no APC mutations were found. beta-Catenin immunohistochemistry revealed nuclear accumulation of beta-catenin protein in all nine tumors with a beta-catenin mutation and two additional tumors without a mutation. These results suggest that activation of the Wnt signaling pathway by beta-catenin mutation contributes significantly to the hepatocellular carcinogenesis associated with HCV infection.