(주) 뷰노

Press Releases

VUNO Presented an Abstract of Deep Learning-based Analysis of Histopathology Images of Colorectal Cancer at the AACR 2020

  • 06. 25. 2020

VUNO Presented an Abstract of Deep Learning-based Analysis of Histopathology Images of Colorectal Cancer at the AACR 2020

 

VUNO Inc., South Korean artificial intelligence (AI) developer announced that it presented an abstract of its deep learning-based analysis of tissue segmentation in histopathology images of colorectal cancer at the American Association for Cancer Research (AACR) 2020, successfully held for three days from June 22-24, 2020.

 

VUNO was accepted for poster presentation at AACR 2020, to which it submitted its first research abstract for the year. AACR was established in 1907 and currently has over 46 thousand members from more than 120 countries all around the world. It annually holds the largest cancer research conference in the world, and this year’s conference which was held online had 120 presentation sessions encompassing various topics including cancer health disparities, and the effects of COVID-19 on clinical trials as well as over 4,000 abstract presentations. With the AACR presentation under its belts, VUNO plans to submit various research findings discovered by its pathology team to high impact journals.

 

The objective of the research was to use deep learning- based algorithms to analyze the spatial distribution and mapping of stroma and lesions composing the complex TME (Tumor Microenvironment) in association with mRNA levels via high throughput deep learning algorithm that is largely responsible for metastasis.

 

The research team used VUNO Med®-PathLab™, VUNO’s AI-based pathological image quantification and analysis platform to analyze cancer tissue slides of colorectal cancer patients from the Cancer Genome Atlas (TCGA) and found that it was able to produce similar results as the highly costly consensus molecular subtypes (CMS) classification. This means that the analysis of histopathology images via the deep learning-based algorithm can extract information for treating colorectal cancer as much as RNA based gene expression analysis, without the need for additional analysis using genetic information and carrying out laboratory research.

 

The team also found that the results of the stroma tissue quantification analysis is correlated to the gene that can cause or suppress the formation of tumors. The research shows when the group has a higher stroma-to-tumor ratio (STR), it has significantly decreased expression of immune related gene expression such as antigen presentation (APC), natural killer cells (NK cells), and plasmacytoid dendritic cells (pDCs), as well as significantly elevated expression of genes related to blood vessel development such as epithelial-mesenchymal transition, desmoplastic reaction, fibroblastic cytokines, and angiogenesis

 

One of the co-researchers, Professor Sunyoung S. Lee of the University of Texas MD Anderson Cancer Center stated that, “This research is meaningful in that we were able to ascertain that the deep learning-assisted algorithm enables automatic segmentation of stroma and malignant lesions and analyzes genomic information.” She also added that, “It can be used as a cost-efficient alternative to the molecular gene tests for colorectal cancer diagnosis and can be used as a biomarker for colorectal cancer treatment with further follow-up research and verification.” She also added that, “With follow-up research, we expect that this solution can be used for predicting the development and treatment of not only colorectal cancer but other types of cancers as well.”

 

Dr. Kyung-doc Kim, the lead of VUNO’s pathology team informed that, “This research was a critical touchstone for demonstrating the potential for VUNO’s digital pathological solutions,” and maintained that, “Our pathology research team will focus on research so that we can not only provide diagnostic tools, but also come up with an AI pathological solution that can comprehensively aid clinical decision-making including patients’ prognosis prediction, and selecting a method of treatment.”

 

Article URL (Kor.): https://www.aitimes.kr/news/articleView.html?idxno=16850

 

Contact Information

Yerim Kim / PR Manager, VUNO lnc.   

Email: rim@vuno.co

 

 

Tag in

#Press_Releases

#VUNO Med® Solution