Science & Enterprise subscription

Follow us on Twitter

  • A company developing health metrics with artificial intelligence received a grant to better understand polycystic o… https://t.co/ElPrTv6NfS
    about 11 hours ago
  • New post on Science and Enterprise: Small Biz Grant Funds AI Repro Health Technology https://t.co/kf8kmvP8Xa #Science #Business
    about 11 hours ago
  • The Food and Drug Administration began a pilot program for assessing alternative technologies to test drugs, biolog… https://t.co/r5mRCGGL4B
    about 15 hours ago
  • New post on Science and Enterprise: FDA Begins Qualifying Non-Animal Testing Options https://t.co/pf0GN32FfV #Science #Business
    about 15 hours ago
  • A new enterprise is developing personalized digital therapeutics for people with Alzheimer's disease and other neur… https://t.co/aFxlnK5VHH
    about 1 day ago

Please share Science & Enterprise

Liquid Biopsy, A.I. Detect Four Cancers

Blood sample tubes

(CDC.gov)

29 Jan. 2020. A single blood test analyzed by machine learning is shown in tests with cancer patients to detect the presence of four types of solid tumor cancers. A team from biotechnology company Bluestar Genomics presented its findings at last week’s Precision Medicine World Conference in Santa Clara, California, and a report appears this week in the pre-print online journal MedRxiv.

Bluestar Genomics, in San Francisco, develops liquid biopsies, simple blood samples to substitute for invasive surgical tumor tissue biopsies to detect and characterize cancer. The company’s technology is spun-off from the lab of Stanford University biotechnology professor Stephen Quake that applies epigenetics, the study of inherited changes in gene expression from outside the basic DNA sequence, to diagnosing disease. The company also employs high-throughput bioinformatics and artificial intelligence to provide actionable findings from epigenetic data for clinicians.

In the paper, researchers led by company scientist Anna Bergamaschi, analyzed blood samples from 188 patients with breast, lung, prostate, or pancreatic cancer. The team also took blood samples from 180 healthy individuals matching the age, sex, and smoking behaviors of the cancer patients for comparison.

The liquid biopsy in the study tests blood plasma for 5-Hydroxymethylcytosine, or 5hmC, a base chemical derived from cytosine, in cell-free DNA, the DNA floating in blood not contained in other cells. Cytosine is one of the four bases making up DNA. Cell-free DNA can result from overproduction of DNA in tumors, which enters the blood stream, and patterns made by 5hmC deposits on the DNA offer signatures for different types of tumors from where they originate. The company analyzes the 5hmC-DNA patterns with sequencing and a classification technique based on machine learning.

The findings indicate Bluestar’s blood test results largely predicted and accurately classified the patients’ cancers, with the highest accuracy for pancreatic cancer, followed closely by breast, lung, and prostate cancer. The company cites data showing these cancers account for 41 percent of cancer occurrences in the U.S. In addition, most of the breast and pancreatic cancer cases in the study were in early stages of progression, with some breast tumors smaller than two centimeters, yet still indicating a high cancer probability.

“There are significant limitations in screening for various cancers,” says Kelly Bethel, Bluestar’s chief medical officer in a company statement released through BusinessWire. Bethel adds, “Detection of small early malignancies is challenging by usual imaging methods, and our platform technology also demonstrates the ability to detect the presence of malignant tumors smaller than two centimeters. Overall, these findings suggest a clinical path toward early detection as part of a multi-cancer screening test.”

More from Science & Enterprise:

*     *     *

Comments are closed.