For patients with acute medical conditions, early diagnosis can mean earlier treatment – and a better chance of recovery. But what does that have to do with data storage?
The answer is, surprisingly, quite a lot.
Artificial intelligence (AI) teams worldwide are analyzing medical images – such as x-rays, CT scans, and mammography scans – to train deep learning (DL) models that save lives. The DL models can look at medical scans and identify symptoms that human eyes might miss. Even when those eyes belong to a highly experienced doctor.
By using AI tools, medical teams can analyze a greater number of scans in less time. Scans can be analyzed immediately, with less human effort, so that doctors are free to spend more time with their patients. Outcomes for patients are better all round.
But where does data storage come in?
World’s Top 10 Most Innovative Health Company Relies on Zadara
At the forefront of this technology is Zebra Medical Vision, a medical imaging team named by Fast Company as one of the 10 most innovative health companies of 2020.
The company empowers radiologists to make new leaps in patient care. It’s All In One (AI1) solutions puts powerful AI tools in the hands of radiologists, making medical image diagnosis faster, more accurate, and automated.
Zebra-Med solutions are helping patients around the world. The company is partnered with Oxford University Hospitals’ osteoporosis team and has developed a COVID-19 solution that analyses chest CT scans to diagnose, triage, and monitor patients.
But to train the AI models at the heart of its solutions, Zebra-Med must analyze and re-analyze millions of high-resolution medical image samples. And that means Zebra-Med, and other medical imaging companies, cannot do this important work without high performance, highly scalable data storage.
A storage challenge unlike any other?
When Zebra-Med started to face significant storage challenges, the company partnered with Zadara to solve them.
One of the company’s biggest obstacles was data growth. As medical scanners become more advanced, they produce scans at higher resolutions and with bigger files sizes. These scan samples never become obsolete, and Zebra-Med can never delete or even archive them – in fact, all data must be kept “hot” for continuous use in AI training.
Another challenge was throughput. Zebra-Med’s team has grown significantly since it was founded in 2014, and researchers now need to run multiple storage-intensive research projects concurrently.
The answer was the Zadara Storage Cloud, a managed on-premises solution with unlimited scale and throughput of up to 8GB/s read and 2GB/s write.
“With Zadara, whatever scale we need, whatever throughput we need, we can get to it,” says Eyal Toledano, Co-Founder and CTO at Zebra Medical, in our new case study. “That’s the killer feature for us.”
But these were not the only data storage challenges faced by Zebra-Med and solved by Zadara Storage Cloud.
The company also needed to:
- Gain cloud-like cost-efficiency while keeping storage on premises, close to its GPUs
- Integrate on-prem storage with AWS cloud storage
- Scale storage quickly at petabyte scale, without loss of performance
- Reduce storage management burden on its research teams
Read the full story on why these challenges are important to medical imaging companies like Zebra Medical – and how Zadara Storage Cloud helped to solve them.