Science Advances: Peking University Yin Yuxin's team develops a new method of AI-assisted metabolomics diagnosis of pancreatic cancer

Pancreatic Cancer is a highly malignant gastrointestinal malignant tumor that is difficult to diagnose and treat
.

Pancreatic Cancer is a highly malignant malignant tumor of the digestive tract that is difficult to diagnose and treat

In recent years, the incidence and mortality of pancreatic cancer have increased significantly.
The early diagnosis rate of pancreatic cancer is not high, and it is often at an advanced stage when it is discovered.
At this time, cancer cells have spread and are difficult to treat.
The 5-year survival rate is less than 7%, which is the best prognosis.
Poor malignant tumors, so it is also called "the king of cancer"
.

According to the latest WHO data, pancreatic cancer is the seventh cancer in China in 2020 (an estimated 120,000 new cases in 2020), and the sixth-largest cancer in deaths (an estimated 120,000 deaths in 2020)

In addition to the traditional blood marker CA19-9 and imaging methods, there is no other effective method for the diagnosis of pancreatic cancer at this stage
.

Therefore, the development of effective detection methods to achieve early, accurate, and non-invasive detection of pancreatic cancer will improve the diagnostic efficiency of pancreatic cancer and reduce its lethality

Recently, Professor Ki Bun new team at Peking University School of Basic Medical Science Advances Online published a report entitled: Metabolic and Detection Systems Analyzes of pancreatic ductal adenocarcinoma through Machine Learning, lipidomics, and Multi-OMICS research papers
.

Science Advances Metabolic detection and systems analyses of pancreatic ductal adenocarcinoma through machine learning, lipidomics, and multi-omics

Yin Yuxin’s team applied machine learning combined with lipidomics and multi-omics techniques to comprehensively analyze the metabolic characteristics of pancreatic ductal adenocarcinoma (PDAC, the most important type of pancreatic cancer), and developed a set of artificial intelligence-assisted PDAC serum metabolism detection methods, and Show the related molecular mechanism
.

Yin Yuxin's team and collaborators have developed a non-invasive detection method for pancreatic cancer using machine learning-assisted metabolomics
.

Using support vector machine-greedy algorithm and high resolution mass spectrometry to analyze non-targeted metabolomics data, 17 serum metabolic markers were screened, and a multi-reaction detection mode targeted metabolism detection method based on liquid chromatography-mass spectrometry was established.

The method tested a total of more than 1800 samples from 4 cohorts, including 1033 patients with pancreatic cancer at different stages
.

In a large external validation cohort of more than 1,000 cases and a prospective clinical cohort containing benign pancreatic lesions, classification detection accuracy of 86.

The study also combined single-cell transcriptome data, tissue proteomics, metabolomics and mass spectrometry imaging and other multi-omics technologies to reveal the mechanism of lipid metabolism changes in pancreatic cancer tissues, and pioneered the use of machine learning-assisted metabolomics An efficient strategy for early detection of pancreatic cancer
.

In summary, this study established a pancreatic cancer detection and analysis method that combines machine learning and targeted metabolomics
.

It demonstrates the advantages of machine learning-assisted serum metabolomics in detecting pancreatic cancer and its application prospects in cancer diagnosis

Wang Guangxi, postdoctoral fellow at Peking University School of Basic Medicine, Associate Researcher Yao Hantao, Institute of Automation, Chinese Academy of Sciences, Deputy Chief Physician Gong Yan, General Hospital of the People’s Liberation Army, and Deputy Chief Physician Lu Zipeng, Jiangsu Provincial People’s Hospital are the co-first authors of the paper, and Professor Yin Yuxin, Institute of Systems Biomedicine, Peking University , Associate Professor Guo Limei, Department of Pathology, Department of Pathology, Peking University School of Basic Medicine, Third Hospital of Peking University, Professor Zeng Qiang of PLA General Hospital, and Professor Jiang Kuirong of Jiangsu Provincial People's Hospital are the co-corresponding authors
.

This work was also supported by the team of Professor Yang Yinmo from Peking University First Hospital, the team of senior engineers from Peking University Analysis and Testing Center Nie Honggang, Professor Zhao Zhicheng and Dr.

Original source:

Original source:

GUANGXI WANG, et al.

Metabolic detection and systems analyses of pancreatic ductal adenocarcinoma through machine learning, lipidomics, and multi-omics in this message