Segmentation of elastic fibres in images of vessel wall sections stained with Weigert's resorcin–fuchsin. (April 2017)

• Record Type: Journal Article
• Authors: Hernández-Morera, Pablo; Travieso-González, Carlos M.; Castaño-González, Irene; Mompeó-Corredera, Blanca; Ortega-Santana, Francisco
• Is Part Of: Computer methods and programs in biomedicine. Volume 142(2017)
• Page Start: 43
• Page End: 54
• Publication Date: 2017-04
• Subjects: Elastic fibres -- Histopathology -- Image analysis -- Segmentation -- Local threshold -- Thickness calculation; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2017.02.018
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: A new method for segmenting elastic fibres in histological images is presented. Shape features are analysed to improve the identification of elastic fibres. Quantification of the vessel wall area and the elastic fibres area can be computed easily. The results are satisfactory based on sensitivity and specificity values, and according to the subjective evaluation by a team of medical specialists. The method can be executed on portable personal computers. Abstract: Background and objective: The elastic fibres are an essential component of the extracellular matrix in blood vessel walls that allows a long-range of deformability and passive recoil without energy input. The quantitative determination of elastic fibres will provide information on the state of the vascular wall and to determine the role and behaviour of this key structural element in different physiological and pathological vascular processes. Methods: We present a segmentation method to identify and quantify elastic fibres based on a local threshold technique and some morphological characteristics measured on the segmented objects that facilitate the discrimination between elastic fibres and other image components. The morphological characteristics analysed are the thickness and the length of an object. Results: The segmentation method was evaluated using an image database of vein sections stained with Weigert's resorcin–fuchsin. The performance results are based on a ground truth generated manually resulting in values of sensitivity greater than 80% with the exception in two samples, and specificity values above 90% for all samples. Medical specialists carried out a visual evaluation where the observations indicate a general agreement on the segmentation results' visual quality, and the consistency between the methodology proposed and the subjective observation of the doctors for the evaluation of pathological changes in vessel wall. Conclusions: The proposed methodology provides more objective measurements than the qualitative methods traditionally used in the histological analysis, with a significant potential for this method to be used as a diagnostic aid for many other vascular pathological conditions and in similar tissues such as skin and mucous membranes.