2008 Summer Meeting - Leeds - The Pathological Society of Great ...

O253 Dimensional Pathology with Virtual SlidesDTreanor 1 , K Brabazon 2 , D Magee 2 ,PQuirke 11 Pathology and Tumour Biology, Leeds Institute of Molecular Medicine,University of Leeds, 2 School of Computing, University of LeedsVisualisation of tissue in 3 dimensions (3D) has many potential uses in tissuebased research. Previously, obtaining 3D models of tissue has been a laboriousprocess requiring manual photomicrography of standard pathology slides or theuse of alternative acquisition methods which result in a loss of cellularinformation.We describe a method for 3D reconstruction of tissue using virtual slides whichallows reconstruction of structures at high resolution (that of tissue microarchitecture)using standard pathology slides. The method uses novel customdevelopedsoftware to align (register) virtual slides and extract 3D informationfrom them.Our methods differ from previous research in this area in that they allow a highthroughputprocess minimising the need for manual intervention. Virtual slidescan be scanned automatically under standard scanning conditions, easilyproducing sequential images to use in 3D construction. Once low resolutionimages have been registered, a further novel aspect of our approach is that smalltissue micro-architectural details can be reconstructed in 3D from any part ofthe registered 3D volume without the need to re-scan the area at higherresolution.Using this method we have produced demonstrator 3D reconstructions of amouse embryo and cirrhotic human liver with 50-200 sections, and an entireglomerulus with over 200 semi-thin sections.This technique can be used to study disease processes in 3 dimensions usingstandard histochemical and immunohistochemical techniques.O27Immunohistochemical Expression of CD10 in Cutaneous Basaland Squamous Cell CarcinomaHAiad 1 , H Hanout 11 Minofyia University, Faculty of Medicine, Pathology DepartmentCD10 is a zinc-dependent metallopeptidase. CD10 expression has beeninvestigated in some cutaneous tumours, however, data regarding its expressionin cutanous epithelial neoplasms are very limited. We aimed to examine theimmunohistochemical expression of CD10 in basal cell carcinoma (BCC)(21cases) and squamous cell carcinoma (SCC) (16 cases). Positive CD10staining was identified as brown cytoplasmic with or without cell membranestaining. In all the 16 SCC cases, tumour cells failed to stain to CD10 incontrast to the stromal cells that showed CD10 expression in 13 cases (81%). InBCC cases, the expression of CD10 was noted in tumour cells in 47.6 % andstromal cells in 95.24%. Most of CD10+ 7/10 BCC showed well-circumscribeddeep margin, however, most of CD10- cases 9/11 showed infiltrating base(P=0.030). BCCs with infiltrating deep margins (12 cases) tended to showCD10 negative basaloid cells (9/12) and CD10 positive stromal cells (12/12)(P=0.0003). There was no significant relation between CD10 and otherclinicopathological factors in SCC or BCC. From our results we suggest thatCD10 might be a useful immunohistochemical marker to differentiate betweenBCC and SCC. At least, if tumour cells were CD10 positive, this would favorBCC over SCC. Absence of CD10 in all the SCC and in infiltrating BCCtogether with its overexpression in the surrounding stromal cells might conferinvasive properties to such tumours. However, its relation to other poorprognostic factors needs larger studies to be confirmed.O26The Immunohistochemical Differentiation Between MetastaticBreast Cancer and Malignant Pleural MesotheliomaGR Powell 1 , WR Roche 11 School of Medicine, University of SouthamptonBackground. The mesothelioma specific markers, calretinin and cytokeratin5/6, and the breast cancer marker oestrogen receptor are frequently used todifferentiate metastatic breast cancer in the pleura from mesothelioma.However, some metastatic breast cancers may mimic the immunohistochemicalprofile of mesothelioma. The objectives of this study were to firstly examinethe expression of these markers in invasive breast cancer, and secondly, to auditthe immunostaining of archived pleural malignancies.Methods. 64 cases of primary invasive breast cancer were immunostained forcalretinin. Tumours that expressed calretinin were additionally immunostainedfor cytokeratin 5/6. Pleural malignancies diagnosed in biopsies from patientswith a history of breast cancer but without asbestos exposure were retrieved andre-examined.Results. 5 cases of breast cancer (8%) expressed calretinin and 4 of these 5tumours also expressed cytokeratin 5/6. The pattern of expression of bothmarkers was indistinguishable from that seen in epithelioid mesothelioma. Allof the 5 tumours were Blooms grade 3 invasive ductal carcinomas and noneexpressed oestrogen receptor. The audit found 2 patients had been diagnosedwith mesothelioma despite a past history of breast cancer and no history ofasbestos exposure. These tumours expressed calretinin but not oestrogenreceptor.Conclusions. Concurrent expression of calretinin and cytokeratin 5/6 andnegative expression of oestrogen receptor in high-grade breast tumours is apotential diagnostic problem in pleural biopsies. This may lead to patients withmetastatic breast cancer being misdiagnosed with malignant mesothelioma.O28Genomic Analysis of Follicular Lymphoma SubgroupsD Marjenberg 1 , PJ Batstone 2 , KE Robertson 1 ,DA Levison 1 ,JR Goodlad 1,31 Division of Pathology & Neuroscience, University of Dundee, Dundee,UK, 2 Department of Pathology, Raigmore Hospital, Inverness, UK,3 Department of Pathology, Western General Hospital, Edinburgh, UKBackground: The characteristic genetic feature of follicular lymphoma (FL) is at(14;18)(q32;q21) present in 85% of cases, leading to inhibition of apoptosisand an indolent phenotype. However, recent studies suggest that there may be at(14;18)(q32;q21)-negative subgroup of FL that is more closely related to, andbetter treated as, diffuse large B-cell lymphoma (DLBCL). We sought todetermine if there were genetic abnormalities to support such a hypothesis, bystudying the genomic profile of a large cohort of t(14;18)(q32;q21)-positive andnegative lymphomas of follicle centre origin.Methods: A combination of morphology, immunohistochemistry andfluorescence in situ hybridization was used to identify groups of translocationpositive and negative FL, and DLBCL of germinal centre type. Karyotypicanalysis (N=100) and microarray-based comparative genomic hybridisation(aCGH) (N=40) were then used to determine the chromosomal aberrations ineach group.Results: t(14;18)(q32;q21)-positive and negative FL were found to havealmost identical genetic profiles, with the exception of a loss of 1p36 found in70% of t(14;18)(q32;q21)-positive FL but only 10% of t(14;18)(q32;q21)-negative FL. A greater difference in chromosomal profiles was witnessed whent(14;18)(q32;q21) negative FL was compared with the cohort of DLBCL.Conclusion: The results indicate that t(14;18)(q32;q21)-negative FL isgenetically more closely related to t(14;18)(q32;q21)-positive FL than toDLBCL. Further investigations are required to determine whether there aremore subtle genetic abnormalities or epigenetic phenomena that might explainthe pathogenesis of t(14;18)(q32;q21)-negative FL and account for theaggressive behaviour of a subset of cases.84 Summer Meeting (194 th ) 1–4 July 2008 Scientific Programme