• Facebook
  • Yahoo
  • Twitter
  • Google
  • Live
  • Facebook
  • Yahoo
  • Twitter
  • Google
  • Live

High Accuracy Classification of Cancer Cells with AI-augmented microscopy

14 Sep | By Sebastian Karpf
High Accuracy Classification of Cancer Cells with AI-augmented microscopy
UCLA and Nantwork Researchers Achieve High Accuracy Classification of Cancer Cells in their Native Form with AI-augmented microscopy
Image source: https://www.nature.com/articles/srep21471

A UCLA Engineering research group in collaboration with scientists from Nantworks has announced a new cell diagnostics method that could open up a new path to data-driven cancer diagnostics and drug development. 

Blood tests typically label cells with chemicals in order to identify them. But these chemicals known as biomarkers cause adverse changes in cells and compromise subsequent analysis such as DNA sequencing. While there are label-free cell identification techniques, they mostly rely only on a single feature of the cell. Just as it would be difficult to identify a person based on a single attribute, say hair color, cell identification based on a single feature has limited accuracy. 

The new system integrates deep learning with a novel microscope enabled by the photonic time stretch technique to achieve record high accuracy in label-free cell classification. The microscope captures fast images of cells then boosts the images and slows them in time - all performed optically - so they can be digitized and analyzed in real-time. Unlike previous implementations, the new system measures multiple biophysical parameter of cells eliminating the need to label the them. It then uses deep learning to distinguish cancer cells from normal white blood cells. 

The technology was developed in the Photonics Lab at UCLA led by Bahram Jalali, a professor and holder of the Northrop-Grumman Chair in Optoelectronics, and doctoral student Claire Chen and postdoctoral fellow Ata Mahjoubfar. Sponsored by Nantworks LLC, the team included also Nantworks senior scientist, Dr. Kayvan Niazi. 

The work was published in Scientific Reports and Springer book titled Artifical Intelligence in Label-free Microscopy. The paper also reports the classification of algae cell based on their lipid content. Since high lipid content cells are better for biofuel production, the technology may find use in development of alternative fuels for clean energy. 


Share on Biophotonics.World:
Share on Social Media:

Comments:

No comments
You need to sign in to comment

Categories

Most Viewed

Most Discussed