Automated selection of DAB-labeled tissue for immunohistochemical quantification

Area of Science:

• Biomedical Imaging
• Computational Pathology
• Histopathology Analysis

Background:

• Immunohistochemistry (IHC) is increasingly used for quantitative protein analysis in clinical and research settings.
• Standardized and automated IHC analysis is crucial for accurate patient diagnosis and prognosis.
• Existing automated imaging techniques face challenges with background interference and consistency.

Purpose of the Study:

• To develop a novel automated imaging technique for quantitative analysis of diaminobenzidene (DAB)-labeled antigens in IHC.
• To compare the accuracy and consistency of the proposed method against seven existing imaging techniques.
• To assess the impact of misclassification on staining quantification using Bland-Altman analysis.

Main Methods:

• Development of a new imaging technique converting brightfield DAB images to normalized blue images.
• Automated identification of positively stained tissue using the novel technique.
• Statistical comparison with seven prior methods using manual analysis by two observers on 18 DAB-stained images.

Main Results:

• The proposed method demonstrated statistically superior accuracy and consistency compared to seven other techniques.
• Manual analysis by observers served as the ground truth for accuracy assessment.
• Bland-Altman analysis confirmed the robustness of the new method in staining quantification.

Conclusions:

• The novel imaging technique provides a more accurate and consistent automated analysis of DAB-stained IHC images.
• This method has the potential to standardize IHC analysis, improving its utility in clinical diagnosis and prognosis.
• The technique effectively overcomes background interference issues present in other methods.