Automated segmentation of tissue images for computerized IHC analysis

Abstract

This paper presents two automated methods for the segmentation of immunohistochemical tissue images that overcome the limitations of the manual approach as well as of the existing computerized techniques. The first independent method, based on unsupervised color clustering, recognizes automatically the target cancerous areas in the specimen and disregards the stroma; the second method, based on colors separation and morphological processing, exploits automated segmentation of the nuclear membranes of the cancerous cells. Extensive experimental results on real tissue images demonstrate the accuracy of our techniques compared to manual segmentations; additional experiments show that our techniques are more effective in immunohistochemical images than popular approaches based on supervised learning or active contours. The proposed procedure can be exploited for any applications that require tissues and cells exploration and to perform reliable and standardized measures of the activity of specific proteins involved in multi-factorial genetic pathologies.

Pioglitazone Treatment Reduces Adipose Tissue Inflammation through Reduction of Mast Cell and Macrophage Number and by Improving Vascularity

“…To quantify the amount of collagen and elastin in the tissue section, the same techniques were used as described above, and explained in details in Methods S1, but color thresholding of the ECM segmentation was calculated based on the automated segmentation algorithm to calculate the percent collagen and elastin [28], [29].…”

Histological and immunohistochemical evaluation of biphasic calcium phosphate and a mineral trioxide aggregate for bone healing in rat calvaria

“…The perimeters of the deposited histological bone matrix and osteoid tissue areas were carefully traced, all areas were computed into the total histological area. At the same time, the percentage of positive BMPR-1B protein was accounted by automation, following the protocol established by Di Cataldo et al 2010 9 . This automated counting allowed counting only the percentage of protein present in the whole defect; it is not possible to distinguish their specific immunostaining in cells or bone matrix.…”

Micro‐structural tissue analysis for automatic histopathological image annotation

“…The main challenge of this task consists in dealing with the colocalization problem, a very common phenomenon produced by the chemical reactions of the target proteins with more than one specific dye, or by the co-existence of structures that also respond to various dyes (Cataldo et al, 2010). The color unmixing process herein implemented was followed by normalization of each of the found stain components, allowing to compare feature descriptors between local patches from different tissues.…”