Frequently Asked Questions.
Here are all the answers to the questions you're probably asking yourself. Or at least they're the ones we're asked most often by our customers.
Which AI Solutions can I get access to?
Already available on ‘deepcOS’ are over 70 carefully selected and clinically validated AI algorithms, developed by globally leading AI companies. Currently supported modalities are CT, MRI, X-Ray, Mammography, PET and SPECT. Take a look at our Product page to learn more.
What do the AI outputs look like in my workflow?
The outputs from the AI solutions can be displayed in different ways:
DICOM objects (such as secondary capture, GSPS, etc.) stored directly in your PACS, utilizing heat-maps, lines / arrows, bounding boxes, regions of interest, circles and/or volumetric segmentations
In text form where detected findings can be integrated into applications including PACS work lists and structured reports, or automatic calculations e.g. of lesion size and volume.
If you would like to learn more about how this could look for you in your systems, please contact support@deepc.ai.
How do I know which AI solutions generate the best value?
deepc provides clinical value consulting to discuss the unique needs of your institution and to find the AI solutions that create the most value. The deepc platform is vendor neutral, meaning we objectively evaluate and partner with providers of many AI solutions on the market, providing you with the widest choice.
How do I know that the AI solution is accurate?
deepc only partners with regulatory cleared AI engines. To get regulatory clearance for an AI engine, medical device companies need to demonstrate the safety and effectiveness of their products to competent authorities prior to market release. Our AI partners based in Europe are CE marked, while AI partners based in the USA are FDA-cleared.
Do I need to initiate the AI analysis or is it automatic?
As soon as the study has entered the PACS, ‘deepcOS’ is automatically triggered and the study is routed to the relevant AI solution, meaning no additional clicks are required. However, we also provide the ability to manually initiate specific AI-engines if desired
How long does it take until the results are available in my PACS?
Depending on the individual AI solution and the use case, processing times can be between a few seconds and a few minutes. For screening use cases, deepc and our partner AI vendors process the studies in the background as soon as the study has entered the PACS, without requiring interaction to initiate the analysis.
Is deepc compatible with any scanner?
Yes, deepcOS is compatible with all scanners as well as all PACS providers. There may, however, be some AI solution specific requirements which we will inform you of during the AI selection process.
Can I host my own AI models on the deepc platform?
For productive clinical use, we are only partnering with AI solutions that are regulatory cleared. deepc does however collaborate on projects with universities and research institutions to support AI-related projects, and for this purpose we support non-regulatory cleared engines. If you would like to discuss with us the opportunities to host your AI solution on deepcOS, please reach out to us here.
How does ‘deepcOS’ work?
deepcOS will process selected studies either automatically or manually from your PACS with the AI applications of your choice. The results of AI processing are made available where radiologists need them, for example within the PACS, modality workstation, RIS or other downstream application. deepcOS operates in the cloud with highest data privacy standards. For further information, visit our product section.
How long does it take to get access to additional AI solutions?
After signing an appendix to the existing commercial contract, gaining access to additional AI solutions can take as little as a few hours. All that is required from our side is to add an orchestration path from the ‘deepcOS’ gate, our on-premise component, to the new AI solution.
Is the deepc platform secure?
Our infrastructure is designed so that personal data will never leave the institution. All patient data is pseudo-anonymized before being processed by the AI solutions, and any personal data is deleted and not stored after AI processing. Our cloud services are also secure, and only provided by ISO 27001 certified suppliers.
What regulatory clearance do the AI solutions have?
To be hosted on deepc, all AI-engines must be regulatory approved medical devices and therefore have at minimum either the CE mark or FDA clearance. To learn more, visit our AI portfolio.
Where is the cloud from deepc hosted?
Depending on your location, we are using the closest servers to host our infrastructure. We are cloud-agnostic and can easily switch to other cloud providers for specific installations upon request.
Does deepc have a Data Protection Agreement with each AI vendor?
deepc has an AVV or DPA in place for each subcontracted AI Vendor as required by the GDPR.
What are the benefits of joining a platform like deepc?
Deploying AI solutions via the deepc platform offers numerous advantages including:
One time integration effort with deepc with the potential of numerous installations.
Reduced contractual, legal, administrative and customer support requirements as these are provided by deepc.
Shorter time to revenue & increased market reach by leveraging our existing customer base and partnerships with leading PACS providers.
Does deepc support non-regulatory cleared engines?
deepc allows only productive usage of CE and FDA-cleared engines. However, deepc also collaborates on research projects with universities or research institutions to support AI-related projects.
How can we integrate with deepc?
The simplest method is to forward AI results via DICOM secondary captures (SC), DICOM Structured Reports (SR) or DICOM PDFs, which are then stored in the PACS. The preferred integration method however is through our SDK, which allows you to leverage our deep integrations with PACS vendors and, for example, populate the radiologist’s worklist directly with your AI outputs. Integrating via the SDK also allows the opportunity to gain real-life feedback on the performance of your AI solutions.