ANAlyte: A modular image analysis tool for ANA testing with indirect immunofluorescence.

Area of Science:

• Medical Imaging Analysis
• Computational Pathology
• Autoimmune Disease Diagnostics

Background:

• Indirect immunofluorescence (IIF) assays for Anti-Nuclear Autoantibody (ANA) testing are vital for autoimmune disease diagnosis.
• Manual interpretation of HEp-2 cell IIF images suffers from subjectivity and lack of repeatability.
• Automating ANA image analysis is essential to overcome diagnostic inconsistencies.

Purpose of the Study:

• To develop and validate a modular, automated tool (ANAlyte) for characterizing IIF images of HEp-2 cells.
• To address the challenge of integrating and assessing the effectiveness of individual ANA image analysis workflow steps.
• To provide a robust solution for accurate intensity level and fluorescent pattern classification in ANA testing.

Main Methods:

• Development of ANAlyte, a modular tool for automated IIF image analysis.
• Integration of an Intensity Classifier module using multi-scale contrast assessment.
• Incorporation of a Cell Segmenter module for individual HEp-2 cell isolation.
• Inclusion of a Pattern Classifier module to determine fluorescent patterns based on cell-level analysis.

Main Results:

• Experimental validation of ANAlyte on two public benchmarks using a leave-one-out cross-validation strategy.
• Achieved overall accuracy of approximately 85% for fluorescent intensity classification.
• Demonstrated above 90% accuracy for fluorescent pattern classification.
• Results compared favorably with state-of-the-art techniques.

Conclusions:

• ANAlyte automates all major steps in ANA image analysis, unlike most existing methods.
• The tool accurately characterizes HEp-2 slides for intensity and fluorescent patterns.
• ANAlyte's performance is comparable or superior to state-of-the-art methods, even with manually segmented cells.
• ANAlyte offers a valid solution for the automatization of ANA testing.