AI Goes Live for Clinical Pathology

A healthcare provider in Israel has begun to use what its vendor claims is the first-ever artificial intelligence-driven (AI) pathology diagnostic tool to go live in a clinical setting, according to an announcement.

Maccabi Healthcare Services, which runs a pathology institute responsible for 160,000 histology accessions every year, recently started leveraging Ibex Medical Analytics’ Second Read system, the tech developer said yesterday. The launch came after a pilot period in which the AI tool “identified isolated major errors” regarding prostate core needle biopsies, which were incorrectly diagnosed benign, according to Ibex.

Although this is not the first clinical application of AI, Ibex said it is the first in this particular space, helping patients with prostate cancer and adding to a growing portfolio of the technology’s diagnostic capabilities.

>> READ: Leo Celi and the ‘Holy Grail of Personalized Medicine’

“We are excited to be the first company to ever deploy an AI-based system in a clinically active pathology lab, leveraging the enormous potential of [AI] to make a real impact on human lives,” said Joseph Mossel, MS, co-founder and CEO of Ibex Medical Analytics.

The technology is already paying off. Not long after its launch, the system pinged a prostate core needle biopsy that a human pathologist marked as benign hours earlier. The AI forced a second look at the case, resulting in a diagnosis of prostate cancer, according to Ibex.

“It is extremely pleasing that our system has already positively affected a patient,” Mossel added.

Ibex’s Second Read software uses machine learning to pinpoint different kinds of cells and other markers in images of prostate core needle biopsies, grading cancerous glands and more. The algorithm was trained on thousands of images of prostate core needle biopsies, according to Ibex. The goal is to prepare pathologists to churn out “more efficient, metric-driven, objective, and accurate diagnoses,” the company noted.

“The complexity of prostate cancer diagnosis, together with the considerable shortage of pathologists, makes a second-read system like this extremely useful for diagnostic accuracy and safety,” said Judith Sandbank, MD, who leads the pathology institute and serves as Ibex’s chief medical officer.

The announcement comes at a time when AI for healthcare is heating up in the United States and elsewhere. Last week, for instance, the FDA approved a first-of-its-kind AI system to detect diabetic retinopathy. FDA Commissioner Scott Gottlieb, MD, followed that news up with a tweet blast and a speech in which he promised new regulatory approaches for AI and next-generation sequencing, further opening the door to a technology that some believe will transform medicine and others consider worthy of great skepticism.

As recently as January, industry observers have said there was no known clinical application of AI in pathology, though the field’s readiness for such tech has been clear for some time.

Get the best insights in healthcare analytics directly to your inbox.

Related

Charting Changing Communities With Health Tech

What’s the Difference Between AI, Machine Learning, and Good, Old Analytics?

How to Hold Public Algorithms Accountable