AI-based reading solutions: Pointing the way to the cloud

Digitalizing healthcare is in full swing and the number of solutions using AI is increasing.
10:37 AM

Photo: Siemens Healthcare GmbH, 2022

The pace of digitalization in healthcare is accelerating. AI-based reading solutions that support radiologists in managing their growing workload and delivering confident diagnoses will become particularly more important over the next years.

Usage of AI is on the rise

In healthcare, the demand for diagnostic services is steadily growing – while the number of available experts is decreasing. Additionally, diagnostics and treatment are becoming increasingly complex. As a result, software solutions using artificial intelligence (AI) are on the rise. In fact, AI in the healthcare market is projected to grow from USD 6.9 billion in 2021 to USD 67.4 billion by 2027 – translating into a compound annual growth rate (CAGR) of 46.2% from 2021 to 2027.1 Especially in diagnostics, where radiologists must examine ever larger amounts of data, AI has demonstrated its value by supporting radiologists in image reading – pushing healthcare more to the cloud and cloud computing.

Cloud computing becomes essential

The dramatic increase in AI software calls for the implementation of cloud computing. It is essential for tapping the full potential of AI algorithms and maintaining flexibility. Some countries or even individual clinical institutions, however, have strict regulations that prohibit sending data to the cloud because patient data are particularly sensitive. This remains the case even though cloud computing has proved to be safe. So, does this mean that some radiologists will be left out in the cold, unable to use AI-based reading software?

Making full use of AI algorithms with a hybrid solution – an example

Offering automatic postprocessing of imaging data sets through AI-powered algorithms, AI-Rad Companion from Siemens Healthineers is an analytical tool that uses patient data to support radiologists in fast reading and confident diagnostic decision-making. To support customers in making full use of AI algorithms without having to go to the cloud, Siemens Healthineers offers the Edge functionality of its teamplay digital health platform. With this hybrid computing solution, patient data stays on the local server of the clinical institution, while only the AI algorithms of the reading software are managed from the cloud, so they can be reliably updated and maintained.

Deeper insights into how the Edge functionality works

This hybrid computing solution combines essential capabilities of the cloud with the need for local data storage. By activating the Edge functionality, a closed environment is downloaded that is controlled by Siemens Healthineers via the cloud. This allows Siemens Healthineers to fully manage its applications like AI-Rad Companion locally, update AI algorithms within the defined regulatory framework and share data that are relevant for updating the AI based on the preferences of the clinical institution.

The draw of the hybrid solution is that the Edge functionality is a one-way data street. It allows the cloud to send data into the closed environment only to interact with the AI algorithms – and allows the cloud to retrieve only the data needed for servicing the AI, leaving the privacy settings up to the customer and keeping patient data on premises.

Taking a first step toward cloud computing in healthcare

Clinical institutions that are restricted in using cloud computing may have to look for other solutions that allow them to stay current with innovations. With new hybrid computing solutions like the Edge functionality, radiologists have the chance to fully benefit from AI-based reading solutions while complying with strict data protection regulations.



Visit our site teamplay digital health platform

For more visual insights, please view this short video:

Want to get more stories like this one? Get daily news updates from Healthcare IT News.
Your subscription has been saved.
Something went wrong. Please try again.