Improved semiautomatic method for morphometry of angiogenesis and lymphangiogenesis in corneal flatmounts F. Bock a, * , J. Onderka a , D. Hos a , F. Horn a , P. Martus b , C. Cursiefen a a Department of Ophthalmology, University of Erlangen-Nu ¨rnberg, Germany b Institute for Biostatistics and Clinical Epidemiology, Charite ´ University Medicine Berlin, Germany article info Article history: Received 21 November 2007 Accepted in revised form 7 August 2008 Available online 26 August 2008 Keywords: lymphatic vessels blood vessels cornea image analysis grey value threshold angiogenesis lymphangiogenesis morphometry semiautomatic abstract Purpose of the study was to describe a novel semiautomatic, quantitative image analysis method based on threshold analysis for morphometry of corneal (lymph)angiogenesis and to test its validity, reliability and objectivity. Murine corneas were vascularized by using a suture-induced neovascularization assay. For immunohistochemistry, flatmounts of the vascularized corneas were stained with LYVE-1 as a specific lymphatic vascular endothelial marker and with CD31 as panendothelial marker. Morphometry of corneal hem and lymphangiogenesis was performed semi-automatically on digital images using image analysis software. Data were analyzed by a paired t-test, intraclass-correlation and systemic difference analysis compared to a manual method. The semiautomatic method based on threshold analysis was more valid in measuring the area covered by blood or lymphatic vessels. Both methods had a good reproducibility with respect to both vessel types (blood vessels: manual: 0.969, semiautomatic: 0.982; lymphatic vessels: manual: 0.951, semiautomatic: 0.966), whereas the systemic difference was significant for both groups measuring lymphatic vessels (manual: p < 0.003; semiautomatic: p < 0.035) and for the manual method measuring blood vessels (manual: p < 0.0001; semiautomatic: p < 0.419). The new semiautomatic morphometry method based on threshold analysis provides higher accuracy, is more valid than and at least as reproducible and objective as the manual outlining method. Therefore the semiautomatic method can be used to detect even small effects on hem and lymphangiogenesis in murine corneal flatmounts with greater precision. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction Angiogenesis and lymphangiogenesis become more and more important for pathological processes such as tumor growth (Cav- allaro and Christofori, 2000) and graft rejection (Dana and Streilein, 1996; Cursiefen et al., 2004a,b). To analyze the molecular mecha- nism of these processes and to test novel drugs several animal models have been developed (for review see Auerbach et al., 2003). Some of these models use the cornea for two reasons: first, a normal healthy cornea is devoid of both blood and lymphatic vessels, but can be invaded by both vessel types. This process can therefore be stimulated and inhibited pharmacologically. Second, the transparency of the cornea and its normal avascularity (‘‘angiogenic privilege’’) permit excellent visualization of sprouting vessels with low background staining. Since lymphatic vessels are not visible using normal slit lamp magnification, blood and lymphatic vessels together can only be studied histologically. By specific staining for blood vessels (e.g., with CD31/von Willebrand factor) and lymphatic vessels (e.g., with LYVE-1, Podoplanin), hem and lymphangiogenesis can be quantified. Recent studies used different methods to measure the vascularized areas in animal corneas: (i) semiquantitative method: the cornea was divided into several quadrants and each sector was assessed with scores dependent on the number of vessels and the density of the vessel network (Cursiefen et al., 2001a,b; Lai et al., 2007) and (ii) manual quantitative method by outlining of vascularized areas: digital pictures of the cornea along with the limbus were taken and analyzed outlining the total corneal area using the innermost vessel of the limbal arcade as the border. In the latter instance, the neo- vascularized areas were quantified by manually circumscribing the blood vessel network or each single lymphatic vessel with the cursor. The total area of neovascularization was then normalized to the total corneal area, and the percentage of the cornea covered by vessels was calculated (Wu et al., 2003; Cursiefen et al., 2004a,b, 2006; Yoon et al., 2006). (iii) Here we describe a novel semiautomatic, quantitative method based on threshold analyses: digital grey value threshold measurements are used to quantify the areas of neo- vascularization (Conrad et al., 1994; Bock et al., 2007a,b; Dietrich et al., 2007; Usui et al., 2007). In contrast to the quantitative manual outlining method, this new grey value based analysis method has * Corresponding author. Tel.: þ49 9131 85 34141; fax: þ49 9131 85 34636. E-mail address: felix.bock@uk-erlangen.de (F. Bock). Contents lists available at ScienceDirect Experimental Eye Research journal homepage: www.elsevier.com/locate/yexer 0014-4835/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.exer.2008.08.007 Experimental Eye Research 87 (2008) 462–470