Premium
Different assessments of immunohistochemically stained Ki‐67 ocular malignancies
Author(s) -
PAULAVICIENE R,
PETROSKA D,
ASOKLIS R
Publication year - 2013
Publication title -
acta ophthalmologica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.534
H-Index - 87
eISSN - 1755-3768
pISSN - 1755-375X
DOI - 10.1111/j.1755-3768.2013.s090.x
Subject(s) - ki 67 , medicine , stereology , pathology , digital image analysis , immunohistochemistry , computer science , computer vision
Purpose To investigate a tumor proliferation marker Ki‐67 in conjunctival malignancies and to compare the results of different quantitative assessments. Methods This study was designed as a pilot study to assess nuclear proliferation protein (Ki 67) in conjunctival malignancies. Immunohistochemical staining of Ki‐67 was evaluated in 8 surgical samples of ocular surface neoplasias: 6 cases of invasive conjunctival squamous carcinoma and 2 of conjuntival melanoma. Ki‐67 counting and evaluation was assessed on selected areas using different methods: manual counting (MC) of >1500 cells;digital image analysis with Nuclear V9 algorithm and digital image analysis using stereology module for ImageScope of Aperio. Ki‐67 values from manual, stereoscopic and automated analyses were compared. Results The MC, the nuclear V9 algorithm analysis and stereologic method revealed positivity of Ki‐67 in invasive conjunctival squamous carcinoma from 22 to 60%, 21.5 to 43.5 % and from 30.1 to 51.5% respectively; in conjunctival melanomas from 20 to 30 %; from 17.8 to 18%, and from 7.17 to 29.1% respectively. The strongest correlation was observed between Aperio algorithm and stereologic method in conjunctival squamous carcinoma (0.87). Weaker correlation was found between Aperio/manual (0.6) and manual/stereological (0.83) methods. Conclusion Our data demonstrate that different methods of Ki‐67 evaluation may give dissimilar results, but results of this computer assisted nuclear V9 algorithm correlates strongly with stereoscopic analysis and can be helpful in accurate quantitative evaluation of imunochistochemically stained biomarkers.