Google and the US Department of Defense have teamed up to create a new artificial intelligence tool to study and identify cancer among military veterans.
The Pentagon’s Defense Innovation Unit (DIU) has asked Google to produce a new AI model of information that could be used with virtual reality microscopes, Fox News reported exclusively.
The project, called Predictive Health, will help physicians find tumor cells under the microscope to procure information about its genetic profile and structure.
Researchers say the program will help lower the healthcare costs by identifying cancer in its earlier stages and lightening the burden on pathologists so they can focus on treating patients.
Google and the Pentagon’s Defense Innovation Unit are teaming up to create a new AI tool to help military veterans with cancer. Pictured: Google office building in Silicon Valley, California, June 2020
The model, called Predictive Health, will identify tumors and help pathologists learn its genetic profile and structure. Pictured: Staff at Naval Medical Center San Diego’s Radiology department prepare a patient for a brain MRI, August 2020
‘Health care is critical to the military’s force readiness,’ Mike Daniels, Vice President of Global Public Sector at Google Cloud, told Fox News.
‘More important than that, or equally important, is funding for certain cancer-related programs. The [Defense Health Agency] spends $1.7 billion of its annual budget on cancer research.’
Daniels says there are two outcomes Google hopes to achieve with the program.
The first is to increase the odds of good patient outcomes and the second is to help doctors sort through data faster and more efficiently.
‘Medical professionals are already overworked,’ said Nathanael Higgins, the Pentagon’s support contractor managing the program, in a statement.
‘We’re essentially giving them an additional tool that will help them make confident decisions – and know that they made the right decision – so that we’re not facing as many false negatives or false positives.
‘And ultimately we’re able to identify these types of disease states earlier, and that’ll help the long-term prognosis.’
The program will first focus on certain types of cancer including breast, colon and prostate cancer, reported Fox News.
The tool will also look at cervical dysplasia, a precancerous condition in which abnormal cells grow on the cervix lining, and lymph node metastasis.
Researchers will train the AI tool by getting it to recognize a picture of cancer and learn which type it is.
Dr Niels Olson, the DIU chief medical officer and originator of the Predictive Health project, says this is similar to a parent teaching a child how to identify objects.
‘The kid asks: “Mom, is that a tree?” And Mom says: “No, that’s a dog,”‘ Olson said in a press release.
‘The kids learn by getting it wrong. You make a guess. We formally call that an inference, a guess is an inference. And if the machine gets it wrong, we tell it that it got it wrong.’
Higgins adds that this project will help drive down costs by catching cancer earlier and requiring patients to undergo less invasive treatments.
Researchers say this will help drive down healthcare costs and lessen the workload for clinicians. Pictured: Deanna Eichenlaud, a patient at Naval Medical Center Portsmouth, undergoes a PET scan at the medical center’s Nuclear Medicine Clinic, August 2013
The program will first focus on certain types of cancer including breast, colon and prostate cancer but could expand in the future. Pictured: Cmdr Christopher Kuzniewski, lead diagnostic cardiopulmonary radiologist at Naval Medical Center Portsmouth, reads a CT scan, February 2013
It will also lessen the workload on healthcare workers.
Aashima Gupta, the director of global health care solutions for Google Cloud, told Fox News that pathologists looking through microscopes is a ‘laborious’ task.
‘When they’re looking and analyzing tissue, small cells need to be noted and this can be used to determine the stage of the cancer,’ she said.
‘Our hope is to limit the unidentified data by bringing this technology into the clinical workflow so that while the pathologist is doing that examination, the model overlays and identifies the areas where it may need more focus.
‘Doctors can tell if there are certain areas where they can go deeper, or if certain segments need more attention than others.’
The Pentagon says the project is expected to end within two years and could expand different types of cancer as well as other treatments such blood transfusions or vaccines.
Google did not reply to DailyMail.com’s request for comment in time for publication.