New Precision Diagnostic Platform For Colorectal Cancer Using AI and Machine Learning.
Source: Thailand Medical News Jan 25, 2020 4 years, 8 months, 2 weeks, 4 days, 13 hours, 56 minutes ago
A recently developed technique could dramatically reduce the number of
colorectal cancer patients who unnecessarily undergo major surgery to remove tumors, instead of a minimally invasive procedure.
Leading
biomedical scientist Dr Xin Wang of City University of Hong Kong and colleagues have used
AI (
artificial intelligence) and
machine learning to identify a gene expression pattern, or signature, associated with
colorectal cancer spreading, or metastasizing, to the lymph nodes. The signature can be used to more accurately predict the risk of
cancer spreading and the appropriate surgical approach for tumor removal.
Currently
, colorectal cancer is one of the leading causes of
cancer-related deaths worldwide. Localized tumors that are at low risk for developing lymph node metastases can be removed by a minimally invasive endoscopic procedure. However, current methods tend to grossly overestimate the risk of developing metastases. More patients than necessary are referred for major surgery.
Dr Wang and colleagues in the U.S. and Japan used
AI and
machine learning and statistical tools to sift through the gene expression patterns of
colorectal cancer patients with and without lymph node metastases. They were eventually able to pinpoint the expression pattern of a panel of eight genes that can robustly identify patients at high risk of having lymph node metastases.
The
biomedical researchers analyzed tissues of 136
colorectal cancer patients following major surgery to remove small 'T1' tumors, which are only in the inner layer of the bowel. Only 19 (14 percent) of these patients actually had lymph node metastases, while 117 (86 percent) did not.
The research team then compared the gene signature approach to conventional clinical analysis to assess the risk of lymph node metastasis. Their gene signature approach identified only 23 percent as being high risk, far less than the 84 percent identified by conventional analysis.
The novel approach, says Dr Wang, could reduce unnecessary treatment for early stage
colorectal cancer patients by nearly 60 percent. He and his team are planning further studies to evaluate and validate the performance of their method as a step towards translating it into clinical practice.
Dr Wang told
Thailand Medical News via a phone interview, "We hope one day we can develop more precise
diagnostic tools for
colorectal cancer patients so that they would not have to suffer unnecessary surgeries."
t;
Reference : Raju Kandimalla et al. Gene Expression Signature in Surgical Tissues and Endoscopic Biopsies Identifies High-Risk T1 Colorectal Cancers, Gastroenterology (2019). DOI: 10.1053/j.gastro.2019.02.027