AI-based software tool could help pathologists identify cancer cells
Researchers at UT Southwestern have developed an artificial intelligence algorithm that is able to classify lung cancer cell types from digital pathology images.
Researchers at UT Southwestern have developed an artificial intelligence algorithm that is able to classify lung cancer cell types from digital pathology images.
“It is time-consuming and difficult for pathologists to locate very small tumor regions in tissue images, so this could greatly reduce the time that pathologists need to spend on each image,” says Guanghua Xiao, professor of population and data sciences at UT Southwestern.
"As there are usually millions of cells in a tissue sample, a pathologist can only analyze so many slides in a day,” adds Xiao. “To make a diagnosis, pathologists usually only examine several ‘representative’ regions in detail, rather than the whole slide. However, some important details could be missed by this approach.”
The AI-based software tool—called ConvPath—automatically identifies each cell in the pathology image as tumor cells, stromal cells (connective tissue cells) and lymphocytes (white blood cells), then converts the image into a spatial map while clusters of tumor cells are further identified as tumor regions.
According to a study published in the journal EBioMedicine, ConvPath’s overall classification accuracy was 92.9 percent and 90.1 percent in training and independent testing datasets, respectively.
As a result, the study’s authors conclude that ConvPath could potentially assist pathologists in clinical practice in a number of ways, including helping them to quickly pinpoint tumor cells.
“This tool could help pathologists and clinicians to predict the patient prognosis, and therefore to tailor the treatment plan of individual patients using readily available tissue images,” contend the authors.
In addition, they make the case that ConvPath “could be used to quantify cell-cell interactions and distributions of different types of cells, especially the spatial distribution of lymphocytes and their interaction with the tumor region, which could potentially provide information for patient response to immunotherapy.”
The software tool for lung adenocarcinoma digital pathological image analysis, which incorporates deep learning, has been made publicly accessible by UT Southwestern to “facilitate users in leveraging this pipeline for their research.”
“It is time-consuming and difficult for pathologists to locate very small tumor regions in tissue images, so this could greatly reduce the time that pathologists need to spend on each image,” says Guanghua Xiao, professor of population and data sciences at UT Southwestern.
"As there are usually millions of cells in a tissue sample, a pathologist can only analyze so many slides in a day,” adds Xiao. “To make a diagnosis, pathologists usually only examine several ‘representative’ regions in detail, rather than the whole slide. However, some important details could be missed by this approach.”
The AI-based software tool—called ConvPath—automatically identifies each cell in the pathology image as tumor cells, stromal cells (connective tissue cells) and lymphocytes (white blood cells), then converts the image into a spatial map while clusters of tumor cells are further identified as tumor regions.
According to a study published in the journal EBioMedicine, ConvPath’s overall classification accuracy was 92.9 percent and 90.1 percent in training and independent testing datasets, respectively.
As a result, the study’s authors conclude that ConvPath could potentially assist pathologists in clinical practice in a number of ways, including helping them to quickly pinpoint tumor cells.
“This tool could help pathologists and clinicians to predict the patient prognosis, and therefore to tailor the treatment plan of individual patients using readily available tissue images,” contend the authors.
In addition, they make the case that ConvPath “could be used to quantify cell-cell interactions and distributions of different types of cells, especially the spatial distribution of lymphocytes and their interaction with the tumor region, which could potentially provide information for patient response to immunotherapy.”
The software tool for lung adenocarcinoma digital pathological image analysis, which incorporates deep learning, has been made publicly accessible by UT Southwestern to “facilitate users in leveraging this pipeline for their research.”
More for you
Loading data for hdm_tax_topic #better-outcomes...