• AI
  • Molecular Imaging
  • CT
  • X-Ray
  • Ultrasound
  • MRI
  • Facility Management
  • Mammography

Advanced Machine Learning in Imaging: The Zebra Approach

Article

A machine-learning radiology startup takes on cutting costs and improving quality of care.

Think about your busiest days – ones filled with dozens of images flying at you from referring physicians, as well as the emergency room. Wouldn’t it be nice to take those images, feed them into a computer program, and have the program spit back a diagnosis within minutes?

Your time savings would be significant, and you might even be able to catch clinically significant issues that you would otherwise overlook. The good news is that this capability largely exists and is currently available in a growing number of radiology areas.

Making diagnosis easier via an automated engine is the founding principle behind Zebra Medical Vision, said Elad Benjamin, the company co-founder and chief executive officer.

“We can provide organizations a significant benefit by giving them an engine that is completely and automatically scalable,” he said. “One that can run through hundreds, millions of studies.”

Zebra pulls together images from research endeavors and other clinical environments to create an extensive imaging library to provide radiologists with the most comprehensive diagnostic results.

In fact, Benjamin said, the company’s name reflects its efforts to identify as many clinically significant findings as possible. An existing maxim states when you hear hooves, first think of horses – not zebras. The same concept applies to this service. The diagnostic engine searches deeply to pinpoint all obvious issues (horses), as well as less common ones (zebras).

How It Started
From the beginning, said Zebra Medical Advisory Board member Eliot Siegel, MD, Zebra’s intent was to tackle challenges created in the health care industry from the absence of a comprehensive imaging database. By designing such an engine, the company hoped to improve diagnostic abilities, as well as create possibilities for practitioners to contribute to research.

These goals appeared in 2013 with the real development work materializing in 2014. Earlier this year, the company’s doors finally opened, and practices and hospitals are putting Zebra’s services to work.

Initially, company creators, Benjamin said, designed algorithms to mine and analyze the myriad of included and submitted images. They also spent significant time indexing each study, ensuring the search engine would be high-speed.

Ultimately, he said, Zebra wants to improve the quality of care and help reduce overall costs. The company also hopes its services will help radiologists better manage their time and provide capabilities to help providers hone in quickly on where problem areas might be.

“The type of organizations that knock on our door are ones that are looking to increase their efficiency, as well as look at population health management,” he said.

Who Are Zebra’s Customers?
According to Benjamin, Zebra has only been in full operation since the beginning of the year, so it has just a few clients on board taking full advantage of its services. Much of the focus has been on the U.S. market, but attention is slowly expanding to global opportunities, as well.

“Outside the United States, there are millions of people in the middle class who are now starting to consume health care services, including imaging,” he said. “But, today, there’s not a deep enough pool of qualified radiologists to read images.”

By providing an automated diagnostic service, Zebra can fill a gap that could greatly improve international patient care, he said.

In the current domestic health care climate, there’s also a wide variety of facilities that can benefit from Zebra. For example, integrated health management organizations, accountable care organizations, and those organizations focused on the cost-side of providing health care are finding Zebra to be a useful partner.

Growing Popularity[[{"type":"media","view_mode":"media_crop","fid":"48968","attributes":{"alt":"machine learning","class":"media-image media-image-right","id":"media_crop_5221948258616","media_crop_h":"0","media_crop_image_style":"-1","media_crop_instance":"5882","media_crop_rotate":"0","media_crop_scale_h":"0","media_crop_scale_w":"0","media_crop_w":"0","media_crop_x":"0","media_crop_y":"0","style":"height: 134px; width: 201px; border-width: 0px; border-style: solid; margin: 1px; float: right;","title":"©Ai825/Shutterstock.com","typeof":"foaf:Image"}}]]
Although many clinical environments are still identifying how best to take advantage of Zebra’s services, the industry is largely starting to take note of the possibilities, Benjamin said. There’s a paradigm shift toward using more automated services to positively impact clinical activities and patient care, so more hospital executives are considering tactics for integration.

“The bigger the shift toward risk-based, capitated medicine, the more organizations will be looking into preventive care,” he said. “The more they look to identify people’s conditions earlier on, the more they’re going to need tools like ours.”

A major driving force in the popularity and utility of predictive analytics, he said, will be the continued rise of health care costs.

In addition to augmenting clinical services, said Siegel, who is also a diagnostic radiology and nuclear medicine professor at the University of Maryland School of Medicine, some academic medicine practices and clinics could opt to use Zebra’s services to strengthen any research efforts that involve diagnostic imaging.

What Services Are Available and How It Works
Currently, Zebra is implementing its first few capabilities. The engine can identify individuals at risk for osteoporosis, fatty liver disease, and emphysema. The capability to make an identification is so strong, Benjamin said, that human involvement isn’t needed.

“We don’t provide the diagnosis – we raise the flag for people that seem to be at high risk for the disease,” he said. “We tell physicians, or whoever might need to have the information, that this is something they need to be aware of.”

In essence, rather than a physical product, Zebra provides a service that is an advanced clinical decision support tool. It assists in providing population health management, and soon, Zebra will be able to confidently identify pulmonary embolisms and recognize hypertension.

Clients who contract Zebra’s services give the company the authority to access and connect with their PACS. That simple integration enables providers to easily submit an image for analysis. For example, Benjamin said, if a practice sends in a CT abdomen, the appropriate algorithm will search like-images to determine any clinically significant findings before sending the results back.

“The destination for results depends on the practice or the hospital,” he said. “Some want results sent back straight to the electronic medical record – others to the RIS or radiologist’s workstation.”

To analyze the images, Zebra uses its algorithms that compare new images provided by referring physicians and hospitals to those housed within its large library. Providing this type of analysis, Siegel said, will also help bring together the radiology research and clinical practice worlds to improve and advance patient care.

“The concept and direction in which Zebra is heading will result in major improvements in how we bridge the chasm that exists between excellent research and routine clinical care,” he said. “It will all allow for safer, more accurate, productive diagnoses.”

Recent Videos
Current and Emerging Insights on AI in Breast Imaging: An Interview with Mark Traill, Part 1
Addressing Cybersecurity Issues in Radiology
Computed Tomography Study Shows Emergence of Silicosis in Engineered Stone Countertop Workers
Can an Emerging AI Software for DBT Help Reduce Disparities in Breast Cancer Screening?
Skeletal Muscle Loss and Dementia: What Emerging MRI Research Reveals
Magnetoencephalopathy Study Suggests Link Between Concussions and Slower Aperiodic Activity in Adolescent Football Players
Radiology Study Finds Increasing Rates of Non-Physician Practitioner Image Interpretation in Office Settings
Assessing a Landmark Change in CMS Reimbursement for Diagnostic Radiopharmaceuticals
Addressing the Early Impact of National Breast Density Notification for Mammography Reports
2 KOLs are featured in this series.
Related Content
© 2024 MJH Life Sciences

All rights reserved.