Requirements Federated Learning and mUlti-party computation Techniques for prostatE cancer
0.1.0 - ci-build
Requirements Federated Learning and mUlti-party computation Techniques for prostatE cancer
0.1.0 - ci-build
Requirements Federated Learning and mUlti-party computation Techniques for prostatE cancer, published by HL7 Europe. This guide is not an authorized publication; it is the continuous build for version 0.1.0 built by the FHIR (HL7® FHIR® Standard) CI Build. This version is based on the current content of https://github.com/hl7-eu/flute-requirements/ and changes regularly. See the Directory of published versions
A conceptual/logical model of the data that has to be exchanged SHALL be specified.
Context: The artificial intelligence developed in the FLUTE project will need to train on data exposed by the hospital members of the project. In order to have a clear vision of what is to be exchanged, it is necessary that a logical model of what is to be exchanged is shared between the different members of the project. Such logical model can then be shared using a FHIR StructureDefinition.
In order to guarantee the correct exploitation of data, Privacy Policies SHOULD be modeled and exchanged using the HL7 FHIR standard.
In Particular, permission to access healthcare Data SHOULD be modeled and exchanged using the HL7 FHIR Permission resource.
Context: In the project, aggregate information is to be exchanged. Each hospital will define its privacy policy to define how private the information is to be. Information on this privacy policy is to be modeled and shared using the HL7 FHIR standard so it can be exploited. Tracking which institution or user has the right to access the data, or which information can be publicly accessed can be done using the FHIR Permission resource.
The prediction of whether a biopsy is need SHOULD be modeled and exchanged using the HL7 FHIR standard.
Context: The project goal is to have an AI capable of detecting whether a patient is at risk of developing a prostate cancer using patient information and medical images. In order to have the possibility for it to be integrated in a standardized way in the clinical patient care, the information of whether the patient is at risk should be shared using the FHIR standard. In particular, the RiskAssessment Resource can be of interest.
The FHIR exchange capabilities of each system SHALL be modeled and exchanged using the HL7 FHIR CapabilityStatement resource.
Each Hospital SHALL expose its non-imaging data using the HL7 FHIR standard.
Context: The FLUTE project is a multicentric, multi-organization project. It is of the utmost importance that each actor knows the data exchange capabilities of its partners and that all partners are aligned on the data exchanged standard to be used. The HL7 FHIR standard offers several data exchange capacity, in particular on displaying a system capacity to exchange information using the standard.
The FLUTE project SHOULD explore the possibility to model AI models using the HL7 standards FHIR and/or CQL.
Context: The question of how to share an AI model in a standardized way has been without answer for now. We estimate that digging into the possibility of modeling AI using the FHIR and/or CQL standards can be of interest for the project an AI studies in general. HL7 organization has a working group studying the possibility to use HL7 standards for AI modeling.
The definition of the different actors of the platform SHOULD be modelled and exchanged using the HL7 FHIR ActorDefinition resource.
The definition of the different requirements of the platform SHOULD be modelled and exchanged using the HL7 FHIR Requirement resource.
The example scenarios of the platform usage SHOULD be modelled and exchanged using the HL7 FHIR ExampleScenario resource.
The testing of the different requirements of the platform SHOULD be modelled and exchanged using the HL7 FHIR TesScript, TestPlan and TestReport resource.
Context: Sharing and exploiting specification as well as being able to develop interoperable solutions is a hurdle when working in a multi-partner project. Having clear and shareable expectations of how systems should behave is critical in diminishing the costs of developing a system in such conditions. FHIR does propose a few Resources that can help.