Automating Pitted Red Blood Cell Counts Using Deep Neural Network Analysis: A New Method for Measuring Splenic...

Area of Science:

• Biomedical Engineering
• Hematology
• Medical Diagnostics

Background:

• Spleen function is crucial for bacterial infection defense, particularly in conditions like sickle cell anemia (SCA).
• Current methods for measuring splenic function, such as manual pitted red blood cell (RBC) counts, are laborious and user-dependent.
• There is a need for a more efficient and reliable method to assess splenic function.

Purpose of the Study:

• To develop and validate an automated workflow for counting pitted RBCs using deep neural network analysis.
• To assess the durability of fixed RBCs for pitted RBC counts over extended storage periods.

Main Methods:

• Deep neural network analysis was employed to segment and count pitted RBCs from differential interference contrast (DIC) microscopy images.
• Automated counts were compared with traditional manual counts from samples of children with SCA and healthy controls.
• Fixed RBC samples were stored for up to 24 months under various conditions to evaluate storage durability.

Main Results:

• Automated pitted RBC counts demonstrated comparable results to manual counting methods.
• Storage of fixed RBCs for up to 24 months, under varying temperatures and light, did not significantly alter pitted RBC percentages (%PIT).
• The automated method proved less time-consuming and offered improved consistency across measurements.

Conclusions:

• Automated deep neural network analysis provides a viable and comparable alternative to manual pitted RBC counting for assessing splenic function.
• Fixed RBC samples are stable for long-term storage, facilitating future diagnostic and research applications.
• This automated approach enhances the efficiency and reliability of spleen function assessment in clinical and research settings.