Mechanical Evolution of Metastatic Cancer Cells in 3D Microenvironment

Area of Science:

• Biophysics
• Cancer Biology
• Medical Imaging

Background:

• Cellular biomechanics is crucial for cancer metastasis and tumor progression.
• Current research often uses 2D models, which may not accurately reflect 3D physiological environments.
• Probing cellular elasticity in 3D cancer models remains a significant technological challenge.

Purpose of the Study:

• To longitudinally image and analyze the mechanical properties of cancerous spheroids in 3D using optical Brillouin microscopy.
• To investigate the potential of biomechanical features for improving cancer cell classification.
• To develop a deep learning model for differentiating cancerous from normal spheroids based on mechanical data.

Main Methods:

• Utilized optical Brillouin microscopy to acquire longitudinal mechanical images of growing cancerous spheroids over 8 days.
• Extracted spatially resolved and temporally evolving mechanical features from the acquired images.
• Employed machine learning algorithms to assess the impact of mechanical features on cancer cell classification accuracy.

Main Results:

• Dense mechanical mapping provided previously inaccessible insights into spheroid mechanics.
• Incorporating extracted mechanical features improved cancer cell classification accuracy from 74% to 95%.
• A deep learning pipeline successfully differentiated cancerous from normal spheroids using only Brillouin images.

Conclusions:

• Spatially and temporally resolved mechanical features from Brillouin microscopy are valuable for cancer research.
• Cancer cell mechanical properties can serve as a novel biomarker for improved cancer classification and detection.
• Optical Brillouin microscopy offers a promising non-invasive tool for analyzing 3D cancer models.