Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Poster display session: Basic science, Endocrine tumours, Gastrointestinal tumours - colorectal & non-colorectal, Head and neck cancer (excluding thyroid), Melanoma and other skin tumours, Neuroendocrine tumours, Thyroid cancer, Tumour biology & pathology

1634 - Prediction and Intervention of Colorectal Cancer Risk with Artificial Intelligent System

Date

21 Oct 2018

Session

Poster display session: Basic science, Endocrine tumours, Gastrointestinal tumours - colorectal & non-colorectal, Head and neck cancer (excluding thyroid), Melanoma and other skin tumours, Neuroendocrine tumours, Thyroid cancer, Tumour biology & pathology

Presenters

Yang Ge

Citation

Annals of Oncology (2018) 29 (suppl_8): viii1-viii13. 10.1093/annonc/mdy268

Authors

Y. Ge1, C.W. Ma2, L. Tao1, M.F. Han1, L. Ma3, G. Li4, L.M. Ma3, Q. Lu2, Z. Wu3, P.F. Yu3, Y.F. Yang3, Y.Y. Ma2, B.B. Song3, W. Gao3

Author affiliations

  • 1 Research Center, Fuyang City No.2 Hospital, 236015 - Fuyang/CN
  • 2 Scientific Research, American Intelligent Health, 98006 - Bellevue/US
  • 3 Scientific Research, Beijing Yiwang Data Technology, 100085 - Beijing/CN
  • 4 Research Support, Fu Wai Hospital, 100037 - Beijing/CN
More

Resources

Abstract 1634

Background

Colorectal cancer is the third most common cancer worldwide, with nearly 1.4 million new cases diagnosed in 2012. It is also one of the most common cancers in China, with nearly 200,000 people dying annually as a result of colorectal cancer. Prediction and intervention of colorectal cancer risk would save millions of lives worldwide. The purpose of this study is to provide a simple, effective and economic method to help people predict and intervene in colorectal cancer risk.

Methods

A total of 720182 subjects including 18406 colorectal cancer patients and 701776 normal people (see the table for details) were involved in the study. The data were used in the study including demographic, CBC, CMP, lipids and urinalysis data. Analysis of covariance, logistic analysis and discriminant analysis were used to identify the significant factors and to build the colorectal cancer risk prediction model and the significant level was set at p < 0.05. SAS was used as the primary statistical analysis tool.Table: 12P

Subject by gender

Subject2012-20152016
MaleFemaleTotal% of MalesMaleFemaleTotal% of Males
Colorectal cancer patients8,0025,22413,22660.5%3,1152,0655,18060.1%
Healthy individuals335,021205,678540,66962.0%99,34961,728161,07761.7%
Total343,023210,902553,92561.9%102,46463,793166,25761.6%

Results

The analysis showed that CBC, CMP, lipids and urinalysis data can significantly distinguish healthy individuals from colorectal cancer patients and those data can be used to build colorectal cancer risk prediction models. The predicting accuracy was 96.1% and the clinical verification rate was 95.7%. Top parameters were selected through the discriminant analysis and logistic analysis. Some parameters, such as red cell distribution width, red cell count, leukocyte percentage and age are positively correlated with colorectal cancer risk, and others, such as albumin, platelet count, hematocrit and platelet distribution width are negatively correlated with colorectal cancer risk.

Conclusions

This research shows that the routine blood and urine test results can be used to predict colorectal cancer risks and the accuracy of the prediction is over 95%. The research would provide an effective, convenient and economical method to help people predict and intervene in colorectal cancer risk.

Clinical trial identification

Legal entity responsible for the study

Fuyang No2. Hospital and Beijing Yiwang Data Technology.

Funding

Has not received any funding.

Editorial Acknowledgement

Disclosure

All authors have declared no conflicts of interest.

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings