Skip to main content
Exploring what’s next in tech – Insights, information, and ideas for today’s IT and business leaders

Creating better stroke treatment using AI and blockchain technology

Every year, 10,000 people die due to a stroke in the Netherlands alone. AI and blockchain are part of the innovative technologies being used to improve the outcome.

One in six people will suffer from stroke in their lifetime. Of the estimated 15 million victims worldwide, 6 million die every year and another 6 million are permanently disabled. Worse, the incidence of strokes is increasing, especially among people under age 55. By 2050, the number of strokes will have more than doubled, according to the American Stroke Association. Beyond the impact on individual patients, there’s also a financial burden, with costs depending heavily on "outcomes." Annual costs of stroke in the European Union alone are estimated at $53.1 billion.

That’s the bad news. As with so many other things, the good news is that scientists are employing technology to minimize the immense burden of this devastating disease. Physicians, who need to quickly evaluate brain images to initiate treatment as soon as possible, are turning to some of the newest technologies to improve the outcome of stroke: AI and blockchain.

First, some medical basics. To function properly, the brain (like other organs) needs oxygen and nutrients. If the supply of blood is restricted or stopped, brain cells begin to die within minutes, which can lead to brain injury, disability, and death. In most strokes (85 percent), the blood supply is stopped because of a blood clot in one of the vessels. It is crucial to restore the blood flow as fast as possible.

Reducing time to treatment saves lives

Stroke patients lose an average of four healthy living years due to the lack of oxygen: That is 25 days per minute, until the blood flow is restored. The chance a patient can live independently after a stroke drops by 10 percent per hour.

Medical specialists’ impression on clinical and radiological findings has a direct effect on delivering fast, objective treatment. This appears to be difficult: Some studies show that only 50 percent of medical specialists come to the same conclusion for a given CT scan. And assessing the medical scans manually and deciding on the best treatment can take hours—when time to treatment has a significant result on the long-term outcome.

Blockchain guide: IDC explains how to drive business transformation with distributed ledger technology and smart contracts.

Treatment depends on the severity of stroke, including which part of the brain was affected and how much brain tissue can be saved. Strokes are usually treated with medication, including medicines to prevent and dissolve blood clots, reduce blood pressure, and reduce cholesterol levels. In some cases, a neurointervention procedure may be required to remove blood clots mechanically. This procedure, called endovascular thrombectomy, can be performed only in specialized hospitals.

Improving stroke protocol

Efficient stroke workflow is critical. Currently, more than 50 percent of stroke patients arrive at their nearest hospital; when endovascular thrombectomy is indicated, they require extra transit time to get to an intervention center where the operation is performed. In this process, accurate assessment of brain scans is extremely important to properly select patients for endovascular thrombectomy—and in a timely manner—by making trade-offs between treatment-related benefits and risks.

In the current process, physicians assess brain scans “with the naked eye,” which leaves room for human error. A reliable manual assessment is a tedious and time-consuming task, and often expert opinion is needed. Images are transported to specialists electronically or by CD-ROM, and valuable time is lost.

AI ensuring better decision-making

Here’s where technology steps in. AI has already made meaningful changes in drug discovery and psychological health. Now, AI is suggesting promise for stroke treatment.

Nico.lab, a medical AI spin-off of the Academic Medical Center in Amsterdam is collaborating with Tymlez, a provider of enterprise blockchain technology. The companies are starting a pilot project this summer with a number of hospitals, aiming to provide superior treatment of strokes. The team received funding from the Dutch Ministry of Economic Affairs and Climate to support the pilot project.

The algorithms learn and improve based on data collected, making them both fast and precise. Nico.lab also makes its AI solution, StrokeViewer, available to hospitals via the cloud, which ensures high-speed results within minutes on any device.

The deep learning StrokeViewer algorithms “learn” based on thousands of CT scans and, as such, may capture features the human eye fails to perceive. For example, the algorithm automatically identifies blood clots in 88 percent of cases, versus 68 percent recognized by experienced neuroradiologists on the same scan.

The anonymized and encrypted CT scans are sent to the Nico.lab cloud via an Internet connection. After the AI analysis is complete, the medical specialist can log in remotely to view the result. The aim is to have the AI-enabled analysis available to the specialist anywhere, anytime, with the time from CT scan upload to receipt of the biomarker report for treatment decision under three minutes.

Blockchain technology guarantees patient data confidentiality

Naturally, confidentiality and data integrity must be guaranteed when exchanging patient data. As with all medical records, this is an exchange of privacy-sensitive data in a chain of medical professionals.

The power of blockchain can provide security in data exchange. In this case, blockchain technology provides fast and secure image exchange between the primary and intervention stroke center. Furthermore, the blockchain technology provides a shared immutable audit trail of all events.

Many blockchains are known for cryptocurrency such as Bitcoin, but Tymlez—here, working with Nico.lab—can ensure that all parts in the chain of people and machines are connected to the system via an unbreakable and safe process. The brain CT scans are uniquely coded so they can be shared with only authorized chain participants, fully guaranteeing the security of patient data.

One cause of blockchain implementation delays is its (perceived) technological complexity, which has a negative effect on its adoption. Among the technical challenges are legacy integration, labor-intense deployment, and scaling, as well as insufficient application performance and data protection. To minimize this burden, Tymlez, SUSE, and Hewlett Packard Enterprise are partnering to provide Nico.lab and its partners each with a blockchain-in-a-box setup; the hardware and software are pre-integrated into a ready-to-go blockchain platform.

Strokes can have an immense impact on individuals and the families who love them. Anything that can be done to improve the likelihood of survival and recovery is a welcome thing.

Better treatment of stroke: Lessons for leaders

Recently, several heads of neurology from hospitals in the Netherlands expressed their support for AI-supported neuroimaging assessments. The reasons are clear:

  • Enable faster treatment time: Faster treatment for stroke victims dramatically improves the patient outcome and reduces the number of hospitalization days.
  • Ensure the right treatment: Selecting more patients for the right treatment, based on better quantitative results acquired through AI algorithms, also leads to improved outcomes.
  • Prevent adverse drug reactions: Avoiding selection of patients for the wrong treatment reduces unnecessary costs and surgery risks.
  • Save lives: Healthcare providers want to pay for AI-supported neuroimaging assessments because it reduces costs, ensures the right treatment, and saves lives.

With special thanks to Merel Boers, PhD, Nico.lab chief research officer, and Reinier van der Drift, co-founder of Tymlez.

Related links:

Know where you do NOT want to use blockchain technology and why

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.