An efficient heuristic for geometric analysis of cell deformations.

Area of Science:

• Medical imaging analysis
• Hematology
• Computational biology

Background:

• Sickle cell disease deforms erythrocytes, impairing oxygen delivery and increasing global health burdens.
• Accurate, automated classification of sickle cells is vital for clinical assessment and crisis management.
• Existing shape-based classification methods use elastic distances in shape space.

Purpose of the Study:

• To refine erythrocyte shape analysis for improved sickle cell classification.
• To develop a computationally efficient method for distinguishing healthy and sickled erythrocytes.

Main Methods:

• Modeled erythrocytes as closed planar curves in shape space.
• Introduced a fixed parameterization based on the cell's major axis for distance computation.
• Aligned cell shapes to templates prior to distance calculation to simplify analysis.

Main Results:

• Achieved a 96.03% accuracy rate in both supervised classification and unsupervised clustering.
• Demonstrated maintained or improved accuracy compared to previous shape space models.
• Significantly reduced computational costs associated with erythrocyte classification.

Conclusions:

• The refined shape analysis method offers efficient and accurate sickle cell classification.
• This approach enhances the potential for automated analysis in clinical settings.
• The method provides a valuable tool for managing sickle cell disease burdens.