A computational streak mode cytometry biosensor for rare cell analysis.

Area of Science:

• Biomedical Engineering
• Analytical Chemistry
• Cell Biology

Background:

• Streak mode imaging flow cytometry captures fluorescent cells as streaks, posing challenges for rare cell detection due to dynamic imaging conditions and low signal-to-noise ratios (SNR).
• Existing methods struggle with enumerating rare cells, particularly when using low-resolution cameras suitable for point-of-care testing (POCT).

Purpose of the Study:

• To develop and validate a novel computational biosensor approach for overcoming imaging challenges in rare cell detection.
• To enhance the accuracy and efficiency of enumerating rare cells in large volumes using streak mode imaging.

Main Methods:

• A computational biosensor was developed, integrating biosensing hardware with algorithms to analyze streak intensity, length, and location in consecutive frames.
• Cell identification involved a three-part process: streak detection, candidate cell identification, and spurious cell filtering.
• Experiments analyzed samples with 1 cell/mL at 10 mL/min flow rate, imaged at 4 frames per second.

Main Results:

• The computational biosensor achieved 98% total cell detection (TD) compared to ground truth (GT) for images with SNR > 4.4 dB.
• Detection accuracy decreased to 66% for low SNR images.
• True positive cells detected (TP/GT) reached 91% for high SNR images.

Conclusions:

• The computational biosensor demonstrates significant analytical capabilities for rare cell enumeration in large volumes, surpassing current technologies.
• This approach offers a promising solution for accurate rare cell detection in POCT settings, addressing limitations of traditional imaging methods.