Detecting infection-related changes in peripheral blood smears with image analysis techniques
Analytical and Quantitative Cytology
|December 1, 1983
Summary
High-resolution image analysis can identify subtle white blood cell changes indicative of sepsis. This automated method accurately distinguishes infected cells, offering a promising tool for disease detection.
Area of Science:
- Medical image analysis
- Hematology
- Computational pathology
Background:
- Subtle changes in white blood cell morphology can indicate disease.
- Detecting these changes manually is challenging.
- Hematologic bacterial infections, like sepsis, alter cell appearance.
Purpose of the Study:
- To assess the feasibility of using high-resolution image analysis to detect sepsis.
- To identify quantifiable cell morphology parameters indicative of infection.
- To evaluate the accuracy of image analysis in differentiating septic from healthy blood cells.
Main Methods:
- Digitizing Wright-Giemsa-stained peripheral blood smears from sepsis patients and controls.
- Extracting geometric, color, texture, and shape parameters from neutrophils and lymphocytes.
- Applying image analysis techniques for quantitative cell assessment.
Main Results:
- Significant differences in color, geometric, texture, and shape parameters were observed between septic and control samples.
- Individual neutrophils and lymphocytes were classified with >84% accuracy.
- A 100% accurate classification was achieved using average cell parameters per specimen.
Conclusions:
- Image analysis techniques can detect subtle, disease-related cell morphology changes.
- The method shows high sensitivity for identifying sepsis-related alterations in white blood cells.
- Further research will explore specificity and broader applications for disease detection.
