AI that analyzes breast cancer tests better than humans.

Image credits: Technion press room.

A deep learning system developed in Israel analyzes breast cancer tests better than a human.

One in nine women in the developed world will be diagnosed with breast cancer at some point in her life.

The prevalence of breast cancer is on the rise, an effect caused in part by modern lifestyles and increased life expectancy.

Fortunately, treatments are becoming more efficient and personalized.

However, what is not increasing, rather decreasing, is the number of pathologists, i.e. doctors whose specialization is to examine the tissues of the body to provide the specific diagnosis necessary for personalized medicine.

Therefore, a team of Israeli researchers decided to turn computers into effective assistants for pathologists, simplifying and improving the work of the human doctor.

Their new study was recently published in Nature communications.

The specific task that Dr Gil Shamai and Amir Livne of Professor Ron Kimmel’s laboratory set out to accomplish falls within the field of immunotherapy.

Immunotherapy has gained prominence in recent years as an effective, sometimes even innovative treatment for various types of cancer.

The basis of this form of therapy is to encourage the body’s immune system to attack the tumor.

However, such therapy must be individualized as the correct drug must be administered to patients who will benefit based on the specific characteristics of the tumour.

Multiple natural mechanisms prevent our immune system from attacking our own body.

These mechanisms are often exploited by cancerous tumors to evade the immune system.

One such mechanism is linked to the PD-L1 protein: some cancers display it and it acts as a sort of password by mistakenly convincing the immune system that cancer shouldn’t be attacked.

PD-L1 specific immunotherapy can persuade the immune system to ignore this particular password, but of course it would only be effective when the tumor expresses PD-L1.

It is the job of a pathologist to determine whether a patient’s tumor expresses PD-L1.

Expensive chemical markers are used to stain a biopsy taken from the tumor to get the answer.

The process is non-trivial, time-consuming, and sometimes inconsistent.

Dr. Shamai and his team have taken a different approach.

In recent years, it has become FDA-approved practice to scan biopsies so they can be used for digital pathology analysis. Amir Livne, Dr. Shamai and Prof. Kimmel set out to see if a neural network could use these scans to make the diagnosis without requiring further processing.

“They told us it couldn’t be done,” the team said, “so obviously we had to prove them wrong.”

Neural networks are trained similar to how children learn: they are presented with multiple labeled examples.

A child is shown many dogs and various other things, and from these examples an idea is formed of what “dog” is.

The neural network developed by Professor Kimmel’s team was presented with digital biopsy images of 3,376 patients labeled as expressing or not expressing PD-L1.

After preliminary validation, we were asked to determine whether biopsy images from additional clinical studies of 275 patients were positive or negative for PD-L1.

It worked better than expected: 70% of patients were able to determine the answer with confidence and accuracy.

For the remaining 30% of patients, the program was unable to find the visual patterns that would allow them to decide one way or another.

Interestingly, in cases where the artificial intelligence (AI) disagreed with the human pathologist’s determination, a second test proved the AI ​​correct.

“This is a momentous achievement,” explained Prof. Kimmel.

“The changes detected by the computer are indistinguishable by the human eye.

Cells are arranged differently whether they have PD-L1 or not, but the differences are so small that even an experienced pathologist can’t identify them for sure.

Now our neural network can do it.”

“It’s an incredible opportunity to bring artificial intelligence and medicine together,” said Dr. Shamai.

“I love math, I love developing algorithms.

Being able to use my skills to help people, to advance medicine, is more than I ever expected when I started majoring in computer science.”

Now he leads a team of 15 researchers, who are taking this project to the next level.

“We hope that AI becomes a powerful tool in the hands of doctors,” shared Prof. Kimmel.

“Artificial intelligence can help make or verify a diagnosis, it can help tailor treatment to the individual patient, it can offer a prognosis.

I don’t think it can, or should, replace the human doctor. But it can make some elements of a doctor’s job easier, faster and more accurate.”

Want to learn more about Israeli technology?

And answer “and” Also add us on Twitter.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Recent News

Editor's Pick