Fluorescence tomography of targets in a turbid medium using non-negative matrix factorization

Area of Science:

• Biomedical Optics
• Medical Imaging
• Fluorescence Tomography

Background:

• Accurate detection and localization of fluorescent targets in turbid media are crucial for biomedical imaging.
• Existing methods face challenges in achieving high spatial resolution and speed for complex biological tissues.

Purpose of the Study:

• To develop and validate a near-infrared optical tomography (NIROT) approach for 3D detection, localization, and cross-section imaging of fluorescent targets.
• To assess the potential of this NIROT method for early-stage breast cancer diagnosis.

Main Methods:

• Utilized multisource probing and multidetector acquisition of diffusely transmitted fluorescence signals.
• Employed non-negative matrix factorization (NMF) for blind source separation to determine 3D target locations.
• Applied a Fourier transform back-projection algorithm for estimating target cross sections.

Main Results:

• Successfully demonstrated the approach in a human breast tissue-simulating phantom with two separated fluorescent targets.
• Achieved accurate 3D localization of targets within approximately 1 mm of their known positions.
• Provided reliable estimates of target cross sections with high spatial resolution and noise tolerance.

Conclusions:

• The developed NIROT approach offers high spatial resolution, fast reconstruction, and noise tolerance.
• Its ability to detect small targets indicates significant potential for early detection and localization of fluorescence contrast-enhanced breast tumors.
• This method could improve treatment outcomes by enabling earlier intervention for breast cancer.