Semi-Automatic Rating Method for Neutrophil Alkaline Phosphatase Activity

Area of Science:

• Hematology
• Medical diagnostics
• Image analysis

Background:

• The neutrophil alkaline phosphatase (NAP) score is crucial for diagnosing myeloproliferative neoplasms.
• Manual rating of NAP scores is subjective and time-consuming.
• There is a need for objective and efficient NAP score assessment.

Purpose of the Study:

• To develop and validate a semi-automatic method for NAP score rating.
• To improve the objectivity and efficiency of NAP score analysis.
• To introduce NAP-PS-IJ as a potential precursor to fully automated scoring.

Main Methods:

• Peripheral blood neutrophils from healthy volunteers were stained using Tomonaga's method.
• Images of neutrophils with varying NAP scores were captured.
• Image-J and Photoshop® were used for image processing, including neutrophil segmentation and binarization (P1, P2).
• Neutrophil alkaline phosphatase-positive area (NAP-PA) and granule count (NAP-PGC) were quantified.

Main Results:

• Binarization with P1 significantly differentiated NAP scores between 0-3+ and 4-5+ groups.
• Binarization with P2 allowed significant differentiation of NAP-PGC across all score groups.
• The semi-automatic NAP-PS-IJ method demonstrated a strong correlation (r=0.92) with human examiner scores.
• The method achieved 60% sensitivity and 92% specificity.

Conclusions:

• The developed NAP-PS-IJ method provides a reliable semi-automatic approach for NAP score evaluation.
• This method shows promise as a foundational technique for future fully automated NAP score analysis.
• NAP-PS-IJ enhances diagnostic objectivity for myeloproliferative neoplasms.