Image-analysis software designed to help astronomers pick out indistinct objects in the night sky has proven adept at flagging biomarkers of aggressive breast cancer, according to scientists at the University of Cambridge.
The algorithms have proven just as accurate as the traditional method of spotting these traits through a microscope--and much faster. A study published in the British Journal of Cancer explains how the technique was used on 2,000 samples from breast cancer patients.
The traditional method requires pathologists to distinguish subtle differences in the staining of tumor cells based on specific proteins they express, an announcement from Cancer Research UK explains. This test focused on gauging the levels of three proteins linked to aggressive cancer.
The technology was able to process the 2,000 samples in a day, while inspecting them manually would have taken a week, Reuters reports. In fact, the software can process up to 4,000 samples a day.
Editor's note: FierceHealthIT is sponsoring an executive breakfast at HIMSS13: Using Predictive Analytics to Improve Care and Efficiencies on Wednesday, March 6. Registration is required--to learn how leaders at premier healthcare organizations, including Muir Medical Group, Intermountain Medical Center and Rush University Medical Center are using data to identify and reduce duplications in service, lower costs, decrease readmissions and avoid adverse events, be sure to sign up to join us.
The researchers now plan to refine the process using more than 20,000 samples. They have placed the samples and algorithms in the public domain to encourage further collaboration, according to Reuters.
"It shows that we don't cross-communicate as much as we ought to," lead researcher Raza Ali, a pathologist from Cancer Research UK's Cambridge Institute, told Reuters.
It's the latest example of applying practices from another field to healthcare. Previously British researchers hit upon the idea of taking a page from aviation to prevent complications from heart surgery. Up to 1,000 sensors aboard aircraft help airlines determine when a plane requires maintenance. This research involved creating an algorithm to predict the likelihood of complications in the 24 hours after surgery based on four metrics: systolic blood pressure, heart rate, respiration rate and peripheral oxygen saturation.
The checklists and other practices common to aviation have been hailed as examples for improving patient safety.
And efforts to improve collaboration among researchers are gaining steam. In one major effort, the European Union recently announced that it will invest 38 million euros (more than $50 million) in global research projects to develop new diagnostic tools and new treatments for people with rare diseases. As part of that, a central disease hub involving 70 institutions will allow a crowdsourcing approach to research data sharing for genomic studies.