top of page

AI Case Study

Researchers at Osaka University differentiate between different types of cancer cells using a convolutional neural network

Researchers at Osaka University were able to train a convolutional neural network to differentiate between mouse and human cell lines and their radio-resistant clones. Being aware of the existence of radioresistant cells in the bodies of cancer patients is important for determining if radiotherapy would be effective. The system was trained on 8,000 images of cells obtained from a phase-contrast microscope. It successfully distinguished between the two type of cells, and therefore now the team's aim is to train to to differentiate more types of cancer cells.

Industry

Healthcare

Healthcare Providers And Services

Project Overview

"Researchers from Osaka University have recently used artificial intelligence (AI) to develop a system that can differentiate between different types of cancer cells. Profound variation in types of cancer cells observed from patient to patient make it particularly difficult for a human to identify cell types, and this AI-system could greatly alleviate this task.

The research team created this system based on a convolutional neural network, a specific form of AI that models the human visual system. Published in Cancer Research, their study used this system to distinguish cancer cells from mice and humans, with equivalent cells with radiation resistance being evaluated as well.

“We first trained our system on 8,000 images of cells obtained from a phase-contrast microscope,” said corresponding author Hideshi Ishii. “We then tested its accuracy on another 2,000 images, to see whether it had learned the features that distinguish mouse cancer cells from human ones, and radioresistant cancer cells from radiosensitive ones.”

Evaluating the system’s findings with a two-dimensional plot, the results for each type of cell were found to be clustered together. These clusters of like-cells were distinctly separated from the other cells on this plot, showing that the trained AI-system was capable of correctly identifying cells based solely on microscopic images.

Going forward, the Osaka research team plans to further train their system to differentiate more types of cancer cells. Ultimately, they hope to develop a universal system that can automatically identify and distinguish all cancer cells."

Reported Results

"Being able to simply scan a microscopic image with this AI-system to detect presence of radiation-resistant cancer cells is a tool that could greatly facilitate the oncologist’s patient management."

Technology

Function

R And D

Core Research And Development

Background

"“The automation and high accuracy with which this system can identify cells should be very useful for determining exactly which cells are present in a tumor or circulating in the body of cancer patients,” said lead author Masayasu Toratani. “For example, knowing whether or not radioresistant cells are present is vital when deciding whether radiotherapy would be effective, and the same approach can then be applied after treatment to see whether it has had the desired effect”."

Benefits

Data

"“We first trained our system on 8,000 images of cells obtained from a phase-contrast microscope,” said corresponding author Hideshi Ishii. “We then tested its accuracy on another 2,000 images, to see whether it had learned the features that distinguish mouse cancer cells from human ones, and radioresistant cancer cells from radiosensitive ones”."

bottom of page