Abstract P1-07-17: V Array: A novel tool for constructing virtual tissue microarrays (TMAs), an evaluation of its use in optimizing TMA construction for Ductal Carcinoma in Situ (DCIS).

Abstract Introduction: The use of TMAs has become invaluable in the assessment of large patient cohorts in clinical practice. TMAs facilitate high throughput analysis and increase biomarker standardization. However, there is little evidence in the literature validating the number of cores required to be representative of the whole tumor. With increasing evidence indicating the heterogeneous nature of many tumors such evidence is critical. DCIS is becoming an increasingly common diagnosis with 5000 new cases p.a. in Canada; with women at risk of recurrence and invasion. It is challenging to create TMAs for DCIS in view of the scattered distribution of the involved ducts. Furthermore, ducts affected with DCIS often vary in architecture, nuclear grade and presence of comedo necrosis even within individual patients. This study aims to determine the number of cores required to construct representative TMAs for different biomarkers in the setting of DCIS. Materials and Methods: Tumor blocks from 102 patients presenting with DCIS alone were retrieved from the archives of Sunnybrook Hospital. Sequential tissue sections were stained with H&E, ER, PgR, HER2 and Ki67. All slides were manually evaluated and Histo-scores determined for ER, and PR, % positive cells for Ki67. and HER2 was classified in accordance with the 2007 ASCO/CAP guidelines. Slides were then scanned at x1.25 magnification on the Ariol SL50 Image Analysis system (Leica Microsystems). A map representing a 5 × 2 TMA, with 0.6mm2 cores was placed on the scanned image of the H&E stained slides and 10 regions of interest (ROI) identified (where possible). The H&E and IHC were then slide linked to then allow identification of the same ROI. The slides were then rescanned on x20, this time only the mapped areas were scanned creating virtual “TMA cores”. Using the V Array (virtual array) function within the Ariol software the virtual cores were placed in a V Array. Previously validated algorithms for ER, PR, HER2 and Ki67 were used to directly analyze each core and the results exported to Excel for analysis. The continuous mean was assessed for increasing numbers of cores and used to determine the optimal number of cores required to be representative of the whole tumor. Results: Virtual TMAs were successfully constructed on all cases. The Histo score of increasing numbers of cores was determined and compared to the overall Histo score for the tumor. The mean numbers of cores required to be representative of the whole tumor was three. Discussion: V array proved an excellent tool for the creation of virtual TMAs and helped to identify the minimum number of cores required to be representative. This technology also has wider applications and may prove very useful in the evaluation of samples with insufficient tumor to allow physical cores to be taken, or where tumors are rare. With the increase in digital pathology and access to scanned images V array will be a valuable addition as a research tool. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P1-07-17.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.