Machine Learning Could Lead to Earlier Mesothelioma Diagnosis

A machine learning tool being developed by Google could lead to earlier mesothelioma diagnosis and better survival rates.

That word comes from Google artificial intelligence experts who recently presented their research at a developer conference in California.

Machine learning specialist Lily Peng told the gathering that advanced A.I. can detect lung cancers like mesothelioma before doctors can even see them.

Earlier mesothelioma diagnosis would allow for earlier intervention which could mean longer survival for victims of one of the world’s deadliest cancers.

Diagnosis Often Comes Too Late

Mesothelioma is a form of lung cancer caused by asbestos exposure. It can take decades to develop, but It is usually fatal within 18 months of diagnosis.

One reason is that earlier mesothelioma diagnosis is rare. Early pleural mesothelioma is difficult to detect using standard imaging tools. By the time a mesothelioma tumor is large enough to be seen on a CT scan, it may already be very advanced.

Late stage pleural mesothelioma is resistant to most kinds of cancer treatments.

Earlier Mesothelioma Diagnosis with A.I.

Google’s A.I. healthcare team has created a machine learning tool that can recognize very early signs of cancer on imaging scans.

For patients at high risk of mesothelioma or lung cancer, some cancer centers are already using low-dose CT scans to try to spot tumors earlier.

But the Google A.I. takes that system up a notch. The new machine learning tool was “taught” how to spot subtle signs of early lung cancer using scans from the National Cancer Institute and Northwestern University.

Peng offered an example of what the technology can do. She cited the case of a patient who showed no signs of lung cancer on a routine CT scan. A year later, the patient had stage III lung cancer.

Five out of six radiologists could see no signs of cancer on the early scan but the A.I. spotted it. Using the tool for earlier mesothelioma diagnosis has the potential to boost survival rates. Peng estimates it could reduce overall lung cancer deaths by as much as 40 percent.

“Clearly this is a promising but early result,” says Peng. “We are very much looking forward to partnering with the medical community to use technology like this to help improve outcomes for patients.”

Peng and her team will soon publish a paper on the new A.I. in the journal Nature Medicine.

Source:

Murphy, Margi, “Google Says Its A.I. Can Spot Lung Cancer a Year Before Doctors”, May 7, 2019, The Telegraph, https://www.telegraph.co.uk/technology/2019/05/07/google-says-ai-can-spot-lung-cancer-year-doctors/

Google Keynote (Google I/O ‘19), YouTube, https://www.youtube.com/watch?v=TQSaPsKHPqs