VIGOR++ Publishes its Technology Roadmap
The VIGOR++ project has completed its technology roadmapping activity producing a to inform scientists, clinicians, system integrators and other stakeholders of likely developments and where the contributions of various parties fit in helping to realise the overall vision. An extract from the executive summary follows.
The roadmap describes a future beyond the focus of VIGOR++ but building upon its achievements. Its scope extends to a model that describes other gastrointestinal (GI) diseases and some degree of interaction between this model and other related organ models. The roadmap was built through consultation with various experts and stakeholders including clinicians, scientists and engineers, patient support groups and others. There was also extensive review of published information.
Two forms of the roadmap are presented – a descriptive version and a diagrammatic version that gives a summary view but makes clearer how the various aspects or layers relate to each other. The roadmap timeline is from 2012 to 2022.
The vision for 2022 is a healthcare scenario where a patient has a personalised virtual GI tract model that can be periodically updated with new MRI scan data thus avoiding the need for frequent uncomfortable invasive investigation and allowing better disease management. For a newer generation of patients with known genetic predisposition to GI diseases there is the opportunity for earlier and easier monitoring and therefore effective disease prevention. Because the tools are based on model and imaging data they lend themselves to remote healthcare or at least a shift away from a hospital-centred approach.
A number of drivers that will affect the likelihood of this vision being realised are explored. Business or market drivers include the increasing demands and expectations placed on national healthcare systems by the need to deal with an expanding and ageing population with increased chronic health conditions. GI diseases in particular are increasingly recognised as a major problem in the developed world. Product or service drivers include a growing demand from clinicians for improved tools for better detection, diagnosis and visualisation of GI conditions. Trends towards ICT enabled healthcare, cloud computing and the need for more objective and quantitative assessments will also support this model-based approach. Technology drivers include advances anticipated in MRI image quality, image sequence protocols, image processing techniques, image registration and segmentation, machine learning applied to modelling and classification tasks, interactive visualisation through combining images and other inputs in multiscale representation, compressive imaging and cloud computing.
The expectations of business and the market set a number of targets for the roadmap over the 10 year period. Integrated digitised medical records will become the norm across major healthcare systems by 2014/15.Large-scale screening programmes for high-risk people should be being carried out by 2022. Routine Crohn’s disease diagnosis using the VIGOR++ model should be in place by 2016 followed by the model being integrated with other related organ models by 2022. The cost and time of diagnosis will be reduced from approx. €1k / 5hrs now to €300 / 1hr by 2016. Small scale systems deployment (20 users) will be achieved by the end of the VIGOR++ project (2014) and large scale deployment (600 users) by 2018. There will be an increased involvement of well-informed patients (so called Super Patients) in their own healthcare choices facilitated by ICT-based solutions.
In order to address the market or business needs described above a number of products, processes or services are expected. Workflows will be developed to accommodate large-scale screening and to enable the ‘Super Patient’ to manage their own health. Remote and offline examination will become an option and will evolve to remote real-time diagnosis. Automation at various stages of the protocol will enable increased speed of diagnosis. Integrated VPH models, linking Crohn’s / GI disease models to cancer of the bowel will be developed and these will then be extended to other cancers. The model tool will also be used for ‘in silico’ drug discovery before 2022. New finance models will be developed to pay for the new approach to healthcare with a range of service models including recognition of the opportunity to target several potential customer types.
To deliver the products and services that address the market and business needs there will have to be technological advances in several areas, particularly in image analysis, modelling and classification and interactive visualisation.
In image analysis scan speed will improve to around 10 minutes for a full abdominal scan by 2022. MRI resolution will improve to around 1-2 mm by 2022. Image quality and utility will be improved by changes to sequence design. Significant further work is needed in the area of selecting and combining features and in addressing registration problems. Image processing will be improved by energy minimisation tools and correlation to a priori information. Optimised algorithms will enable near real-time image segmentation and registration. Mutliscale image correlation will continue to be a major challenge that will not be fully addressed by 2022. DICOM will continue to be the dominant format.
For modelling and classification supervised learning will be most suited to developing the GI model. Expert user involvement to identify regions to be classified and expert feedback to improve classifiers will continue for the foreseeable future. Further work will be needed on measuring the most representative features among several possible ones such as wall thickness, textures, shapes, signal intensity or edge features. Classification speed will not be an issue for run time (training is what takes time) until genomic feature sets become incorporated. On-going research will determine the optimal size/structure of superpixels/voxels which are useful for feature extraction, capture redundancy, and reducing complexity of processing tasks. Statistical features coupled with segmentation information will help calculate the CDEIS score and compare evolution of disease amongst patients.
For visualisation significant research will be needed to address image fusion at different scales e.g. when using different medical imaging modalities. There will be combined interactive visualisations where you can move a cursor over a 3D volume to call up further clinical information – probably the CDEIS scores. New techniques will emerge to deal with the representation from huge multi-dimensional data sets e.g. the trend to compressed imaging or similar will impact in the visualisation field. Mobile device use will be enabled by generating the visualisations at data centres thus only needing to transmit the final image. Augmented- and mixed-reality methods are still in their infancy but are likely to be used in niche applications including training of new physicians and radiologists.
Advances in the area of capsule endoscopy are also explored as these have the potential to provide complementary input data to models and potentially to provide alternative approaches to diagnosis in some cases. The main anticipated improvements to this technique are in extending the measurement capabilities from simple visible imaging, adding the ability to control capsule movement, the potential for smart targeted drug delivery from the capsule and in the long term the possibility to take biopsy samples and even carry out more complicated micro-surgical procedures using the capsule as a platform.
Other areas identified for advancement include training of clinicians in the use of the tools and the inclusion of such ICT-based tools in academic curricula to prepare a new generation of healthcare professionals for the future.
Business model development will also require attention in order to both deal with the cost of implementing the new approach to healthcare and also to ensure that the maximum benefit is extracted from the opportunity it presents. Whilst it is generally considered too early yet to specify the model most likely to be sustainable, there is a clear need to work on understanding what benefits all potential stakeholders could gain from the information in this ICT enabled healthcare toolset and what value they would put on such information. These models will likely need to consider a more patient centric approach with associated issues of data ownership, privacy, consent and assurances on use. There may also be mixed business models that involve a trusted intermediary or broker hosting a service and multiple revenue streams to match the needs and resources of several very different stakeholders such as healthcare providers, governments, insurance companies, pharmaceutical companies and patients.
Finally, social media is already an important tool in providing patients with discrete information and advice on living with GI diseases. These approaches are also likely to help with growing awareness of the alternatives provided by the VPH approach to GI healthcare. They may even play a role in driving the patient centric approach forward more quickly.
The report is available to download via the following link: