Expected Deliverables

With a 2 year project, it is important to ensure that there are clear expectations for what can be delivered in these timescales. We have defined the following as key goals and deliverables for the project:

The primary goal of the eDiaMoND project is to develop a prototype working system by the project end date where a working system is defined as having the following characteristics:

• It has a significantly large distributed database of mammograms (400 cases per site with a majority annotated).
• It is scalable and is designed in such a way that it could scale to cope conceptually with millions of images spread around the 90+ Breast Care Units in the UK.
• It is effective in that it is fast, it is useful to the clinicians in the areas of screening, training, epidemiology and computer aided detection, and it is intuitive for the users.
• It must be built such that upgrades of platform or image analysis software are graceful.
• It is reusable, in that the platform could be used as a foundation for other e-health projects.
• It is based on Grid architecture.

What will the system look like?

To the users, the system will appear to be simply the workstation. It will be showing images and functionality as well as the IBM screens. Behind the scenes, the system will comprise the following :

• Digitised images in SMF format as well as raw anonymised images.
• Information about the case and the images to enable retrieval.
• Additional application specific data e.g. training case information and annotation data.

The prototype workstations will be as good as existing visual methods with the grid middleware able to be used with other workstation manufacturers.

What will the system do?

The system will be deemed a success if it is demonstrated to be:

• Manageable and deployable within the clinical environments
• Robust in that it requires very little daily support from systems administration staff

A key aspect of the project is winning clinic approval, that is to say that the measure of success would be that, for the application areas selected, and with the project constraints imposed, the clinical partners recognise the benefits of the prototype and see it as a must for the future.

With this in mind, we must demonstrate that we have understood the working practises of this field and implemented the workstations in sympathy with them.

The management of expectations will be an essential aspect of this success factor in ensuring that the definition and delivery of the prototype versus the blueprint for the future are clearly communicated. It must also demonstrate the following application areas:

1. Data Acquisition

The project will delivery and deploy data acquisition and annotation capability to enable the project to capture and store approximately 400 anonymised clinical mammograpy cases for the eDiaMoND federated database. This process will enable data to be collected and stored in DICOM format, for display not only by the eDiaMoND applications but also general medical imaging applications and viewers.

2. Screening

The screening workstation will show how this function could be performed using a digital system and will simulate the process of performing a reading of 100 cases in an hour with preloaded data.

This will demonstrate the retrieval of data through queries and the multiple annotations of cases.

3. Training

The Training workstation will demonstrate a computer aided training solution for radiologist using prepared training cases of special interest as well as a selection of normal cases.

The solution will enable the retrieval of cases for test rollers as well as the sitting of tests and subsequent marking.

The success of this application is that it is useful to clinical partners at the end of this project.

4. Epidemiology

The project will show how the infrastructure could be used for epidemiology with the ability to demonstrate the querying of data and subsequent analysis for data joined between NHS databases and the images and base data on the Grid.

In addition, the project will show how studies over the whole federated database could be performed.

5. Computer Aided Detection

Techniques for computer aided detection will be evaluated during this project with clinicians and research staff. The measure of success for this aspect of work would be a demonstration that CADe could be incorporated into the screening process.

6. Image Standardisation

The project will evaluate the effectiveness of standardisation techniques, initially SMF processing, and temporal vectors using clinical input. A measure of success of the SMF processing would be a demonstration that benefits could be achieved through performing this processing.

A Blueprint document will be developed which details what the team consider important for the delivery of such a system in a real world situation, taking into consideration the domain constraints discovered in the project.

Our final deliverable is that we ensure full adherence to the legal and ethical constraints for using data originating from patient records and will seek to discover the processes required to enable eDiaMoND in a real clinical setting with raw patient data.