Skip to main content

Personalised Decision Support for Heart Valve Disease

Periodic Reporting for period 2 - EurValve (Personalised Decision Support for Heart Valve Disease)

Reporting period: 2017-08-01 to 2019-01-31

The primary objective of EurValve was to produce potentially-useable (after regulatory approval) clinical Decision Support System (DSS) tools that could assist clinicians in difficult areas where extra data might assist the decision. EurValve developed a computational tool that provided extra, derived, 'biomarkers' (quantifying aortic stenosis and mitral regurgitation) obtained from 3D analyses of the patient's cardiac geometry. These data were added to an existing commercial valve-sizing system that is currently in clinical use.

The target user will be the healthcare professional, who will make the decision on the nature and timing of the intervention. The major advance of this system over current practice is that it integrates and interprets all heterogeneous data available about the patient, integrates population data where needed, and provides a consistent, repeatable, quantitative and auditable record of the information that contributes to the decision process. The DSS system will allow for in silico simulation of different treatment options and thus allow comparison of their immediate haemodynamic outcome.

The project has constructed a DSS for Heart Valve Disease. It includes a number of important developments and innovations. The DSS includes presentation of clinical guidelines, computational model-based diagnostic and prognostic indications and case-based reasoning. The DSS provides quantitative measures of disease severity and patient impairment, complementary to those gathered in the current clinical pathway, and predicts changes under candidate interventions. The potential of the DSS for improving the management and interventional planning was confirmed by a group of independent clinical users in a control experiment. There has been particular clinical interest in the use of activity monitoring to provide a measure of the degree to which the disease might affect everyday living, and to inform the physiological state in the core computational models.

The design of a practical tool for use by cardiovascular clinicians has been a significant step forward in bringing VPH technology close to the clinic. From the translational perspective, there are significant innovations in the integration of clinical data into the computational workflow and the personalisation of model parameters, and in the streamlining of the analysis process so that it is operable, with acceptable accuracy, in timescales of a few minutes. The process from medical image through data integration to computational results is automated, although there is the facility for inspection and manual correction of valve anatomies for difficult cases. Each case of the 120 cases in the EurValve cohort was processed in approximately four minutes.
The onward exploitation trajectory for EurValve is well-identified, and the prospects for a product with appropriate regulatory approval reaching the market in the foreseeable future are good. It is the intention of the developers to continue with this work, and to bring an enhanced EndoSize product with Decision Support to the market.
WP2 Infrastructure
1. File Store service implemented as the Data Warehouse's architecture.
2. The File Store component implemented a security recommendation of encrypting all contents on the underlying storage resources
3. Dedicated portal was created
4. Web-based visualization component presenting 3D meshes of heart models as output of the segmentation process has been implemented.

WP3 Software Components –Models
1. Estimation of 6 parameters and influencing cardiac output. Estimation based on patient info and ICD codes, model trained with simulated data.
2. Segmentation tools for AV in CT + US TEE/TTE images available via webservice or stand alone cmd + GUI tools.
Parametric model that describes patient with anatomically meaningful parameters + allows 0D embedding and generation of synthetic patients
3. 0D circulation model for heart + tests, parametric 2D model of AV, initial 3D AV model that can be inferred from segmentation
4. Sensitivity analyses and uncertainty quantification toolkit + sensitivity analysis of 0D model. Simulation of test data for (1)
5. Reproduction of detailed ODE cell model, numerical analysis of improved cell excitation and calcium diffusion model. 54 samples measured
6. ROM for parametric 2D AV model, validation and testing of ROM. Automatic generation of iso mesh over heart cycle. Representation of temporal behaviour with SVD modes.

WP4 - Digital Patient Definition, Data collection
Two clinical trials conducted. A Retrospective Study with data from a large patient group, to facilitate the development of the mechanism to infer missing data, and to provide evidence for the generation of the rule sets that will drive the detailed decision support process to inform a ‘machine learning’ process.
The Prospective Study had 120 patients in 3 clinical centres. The computer model was used to simulate each patient’s heart before and after valve replacement, and the results compared with the actual results of the intervention.
Activity monitoring was performed using the Bristol Sphere kit in Sheffield and the Philips Health Watch in all 3 centres.

WP5 Decision Support System
The DSS is built on an existing decision support tool, and has integrated different concepts and patient-specific measures to improve the management of heart valve disease.

WP6 DSS Operation - Clinical
1. Ethical approvals obtained
2. A protocol implemented
3. Imaging standards implemented.
4. Amalgamated set of data defined allowing the storage of clinical parameters in conjunction with analyses performed and modelling output
5. Patients underwent activity monitoring both before and after intervention in STH, CATH and DHZB

WP7 Exploitation
WP7 exploitation ensured that the highest possible impacts of the practical and exploitable outputs are secured. Three reports on exploitation produced.
Dissemination of EurValve includes a website (; 5 newsletters; a flyer; 6 publications and a slide deck. A number of outreach events were attended. A LinkedIn and Twitter group active.
The ambition of EurValve was to develop a Decision Support System (DSS) which went well beyond the state of the art. The DSS developed in EurValve represents a significant advance on current practice, because it combines the power of mechanistic physiological modelling with a level of personalisation that has not hitherto been achieved.

The primary advances beyond the state of the art that were achieved in EurValve are:
•The personalisation of the model using all of the information that is available in the clinical record and in the literature.
•The representation of the physiological envelope of the individual, extrapolating beyond the measurement condition.
•The exploration of the process of inclusion of data from the monitoring of the activity of the individual. This will exploit the rich data that is increasingly becoming available from self-monitoring of lifestyle activity

The EurValve proposal included a categorisation of impacts into multiple areas of applicability, including societal, users, the economy, the EC Work Programme and the call. In general the wider impacts at higher levels of abstraction will naturally benefit from any moves to make adjustments at these more practical levels.