AI tool could help doctors ID breast cancers vulnerable to Enhertu

For decades now, doctors tasked with diagnosing breast cancer have peered into microscopes and pored over slides, hunting for signs the tumor cells they’re studying harbor a protein known as HER2.

A positive result meant they could deploy powerful drugs like Herceptin or Perjeta to knock the tumor into remission, vastly improving their patients’ prognosis. No HER2 expression, on the other hand, would guide them to choose other treatment. Until recently, that diagnostic work was relatively straightforward. Doctors bucketed cancers as either “HER2 positive” or “HER2 negative,” and went from there.

Pathologists’ jobs have become more complicated since the arrival of an antibody-drug conjugate called Enhertu, which delivers a tumor-killing toxin directly to cancerous cells positive for HER2. Testing by Enhertu’s makers, AstraZeneca and Daiichi Sankyo, proved it’s beneficial in as many as two-thirds of tumors that normally would have been classified as HER2 negative.

As a result, doctors must now determine whether tumors might be “HER2-low” or “HER2-ultralow,” and their patients eligible to receive Enhertu. Identifying these tumors is a challenge, though, as it’s easy to miss a scattering of HER2-positive cells among the panoply of others contained within a tissue sample.

“Many HER2-low or HER2-ultralow tumors are misinterpreted as HER2-null cancers, so these patients potentially miss access to effective therapy,” said Marina De Brot, a professor and associate breast pathologist at the A.C. Camargo Cancer Center in São Paolo, Brazil.

Artificial intelligence could help. Research by De Brot and others, unveiled Thursday and set to be presented at the American Society of Clinical Oncology’s upcoming annual meeting, show that an AI support tool can improve physicians’ identification of HER2-low and HER2-ultralow samples.

De Brot shared the researchers’ findings with reporters on a call held by ASCO ahead of the data’s release publicly.

“Our study provides the first multinational evidence that artificial intelligence can help close a critical diagnostic gap and open the door to new therapies like antibody-drug conjugates for a majority of patients who, until recently, had not been offered these options,” De Brot said in a statement provided by ASCO.

In the study, De Brot and her colleagues organized a series of sessions with 105 pathologists from 10 different countries. During the sessions, the pathologists were asked to evaluate 20 breast cancer cases across three different exams. In the third exam, they were given use of AI software developed by AstraZeneca and a diagnostic company called Mindpeak. Their diagnoses were compared to consensus HER2 scoring done by a panel of expert pathologists.

Researchers found that, with AI assistance, the pathologists’ scoring more often agreed with the central reference scores, rising from 76% of the time on average to about 90%. Notably, AI support reduced by more than 25 percentage points the number of HER2-ultralow cases that were misidentified as HER2-negative.

“These findings shed light on the promising role for AI in oncology, not as a replacement for the physician, but as a powerful tool to help us work smarter and faster to deliver high-quality, more personalized care,” said Julian Hong, an associate professor and medical director of radiation oncology informatics at the University of California, San Francisco, in an ASCO statement.

The researchers plan to share the AI support tool with more pathologists from additional countries. They also aim to conduct an “implementation” study that places the tool alongside routine diagnostics, so they can measure whether it leads to changes in treatment or enables patients with HER2-low and HER2-ultralow breast tumors to receive targeted therapy more quickly.

More broadly, studies like these could help raise awareness of the new classifications and encourage greater familiarity with their nuance.

“We’ve had to adapt ourselves,” De Brot said on the call with reporters. “Everything is changing quickly because of this new approach of HER2 assessment in breast cancer.”

This post has been syndicated from a third-party source. View the original article here.