In TMA data management a crucial step is the correct recognition of each core section position in the array, i.e. the right assignment to each core of proper row and column numbers for reliable further data analysis. The available commercial systems support manual or semi-automatic problem solutions. In this talk we present a new approach, based on a ad hoc developed mathematical algorithm, which automatically locates each core section following the right order, allowing a robotic microscope1 to acquire a correct ordered sequence of single core images at high magnification. The algorithm inputs are the digital image of the complete slide and some parameters chosen during block array preparation; the outputs are the ordered list of the coordinates of the core sections centres and an index associated to each spot, representing its dimension and local environment anomalies. Together with algorithm description we present the first results of algorithm accuracy tests and assessment of working constrains. The tool described here is part of an integrated TMA management system, consisting in an automatic image acquisition environment and a web-based multi-centre data collection system2,3. Single core section images at medium and high magnification are processed by a tool we are developing to provide a quantitative evaluation of the expression of some biomarkers on breast tissue samples in an automatic way. In this talk we present some results of applying this image processing and analysis procedure to a nuclear and a cytoplasmatic marker. [1] Demichelis F, Barbareschi M et. al `The Virtual Case. A new method to completely digitise cytological and histological slides`, Virchow Arch 441:159-164, 2002 [2] Dell’Anna R, Demichelis F, Barbareschi M, Doglioni C. “Tissue Microarray Data System”, Sept. 2002 Tech.Rep. ITC, #0209-09. [3] https://bioinfo.itc.it/TMA
How to drive robotic microscope on TMA slide to measure biomarkers expression: an ordering algorithm
Dell'Anna, Rossana;Demichelis, Francesca;Sboner, Andrea;
2003-01-01
Abstract
In TMA data management a crucial step is the correct recognition of each core section position in the array, i.e. the right assignment to each core of proper row and column numbers for reliable further data analysis. The available commercial systems support manual or semi-automatic problem solutions. In this talk we present a new approach, based on a ad hoc developed mathematical algorithm, which automatically locates each core section following the right order, allowing a robotic microscope1 to acquire a correct ordered sequence of single core images at high magnification. The algorithm inputs are the digital image of the complete slide and some parameters chosen during block array preparation; the outputs are the ordered list of the coordinates of the core sections centres and an index associated to each spot, representing its dimension and local environment anomalies. Together with algorithm description we present the first results of algorithm accuracy tests and assessment of working constrains. The tool described here is part of an integrated TMA management system, consisting in an automatic image acquisition environment and a web-based multi-centre data collection system2,3. Single core section images at medium and high magnification are processed by a tool we are developing to provide a quantitative evaluation of the expression of some biomarkers on breast tissue samples in an automatic way. In this talk we present some results of applying this image processing and analysis procedure to a nuclear and a cytoplasmatic marker. [1] Demichelis F, Barbareschi M et. al `The Virtual Case. A new method to completely digitise cytological and histological slides`, Virchow Arch 441:159-164, 2002 [2] Dell’Anna R, Demichelis F, Barbareschi M, Doglioni C. “Tissue Microarray Data System”, Sept. 2002 Tech.Rep. ITC, #0209-09. [3] https://bioinfo.itc.it/TMAI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
