Periodic Reporting for period 2 - DESIREE (Decision Support and Information Management System for Breast Cancer)
Reporting period: 2017-08-01 to 2019-07-31
Breast cancer is the most common and most deadly type of cancer affecting woman in the EU countries, with more than 460,000 new cases and 130,000 deaths in 2012. Multidisciplinary Breast Units (BUs) were introduced in order to deal efficiently with breast cancer cases, setting guideline-based quality procedures, clinical decisions on cases based on consensus and a high standard of care.
Despite the evident advances, daily clinical practice and case presentation in the BUs is hampered by the complexity of the disease, the ever-growing amount of patient and disease data available in the digital era, the difficulty in coordination, the pressure exerted by the system and the difficulty in deciding on cases that guidelines do not reflect.
The amount of data generated for every case may be overwhelming. A single case usually lasts for months or years, with repeated cycles of diagnostics and treatments. The associated digital information generated is increasing exponentially. It includes non-strucuture rich information data sources, such as complex medical imaging datasets, which may include 3D modalities and genomic sequencing data providing crucial information allowing to characterize the tumor (tumoromics) or the possible reaction of the patient to the drug (pharmacogenomics) which in turn determines therapeutic strategies. The potential of exploiting this heterogeneous information and comparing it with other cases is enormous. However, the disparity of sources, inconsistency in representation and lack of tools for retrospective analysis prevent exploiting all this information in an agile manner for continuous care or discovery.
The advent of the BUs has had an important impact in oncology practice, but may drown in an intractable amount of data. This is where DESIREE comes to the rescue, by providing a set of software tools, models and processes to deal with the efficient case representation of BUs, the exploitation of rich sources of information such as imaging or genomics and the review and learning from retrospective cases. Altogether, we call this intelligent software system a Decision Support System, aiming to provide timely information and support for case review in the Breast Unit, based on clinical guidelines, experience from previous cases, advice for non-compliant decisions and exploitation of retrospective cases.
Furthermore, DESIREE develops imaging biomarkers, characterizing breast tissue and tumor and ultimate real-time visualizations based on image segmentation and analysis that provide insight into the disease and its progression, helping decision making. Last, but not least, DESIREE has developed a novel tool, based on a complex physiological model, for predicting the aesthetic long-term outcome of the Breast Conservative Therapy intervention (resection of the minimal amount of tissue necessary instead of complete mastectomy), which is useful both by the the patient and the surgeon.
Our ultimate goal, is to provide intuitive and intelligent software tools that improve the case review and management of the breast units and ultimately are capable to impact the patients by providing the necessary information to ultimately optimize treatments and outcomes.
Despite the evident advances, daily clinical practice and case presentation in the BUs is hampered by the complexity of the disease, the ever-growing amount of patient and disease data available in the digital era, the difficulty in coordination, the pressure exerted by the system and the difficulty in deciding on cases that guidelines do not reflect.
The amount of data generated for every case may be overwhelming. A single case usually lasts for months or years, with repeated cycles of diagnostics and treatments. The associated digital information generated is increasing exponentially. It includes non-strucuture rich information data sources, such as complex medical imaging datasets, which may include 3D modalities and genomic sequencing data providing crucial information allowing to characterize the tumor (tumoromics) or the possible reaction of the patient to the drug (pharmacogenomics) which in turn determines therapeutic strategies. The potential of exploiting this heterogeneous information and comparing it with other cases is enormous. However, the disparity of sources, inconsistency in representation and lack of tools for retrospective analysis prevent exploiting all this information in an agile manner for continuous care or discovery.
The advent of the BUs has had an important impact in oncology practice, but may drown in an intractable amount of data. This is where DESIREE comes to the rescue, by providing a set of software tools, models and processes to deal with the efficient case representation of BUs, the exploitation of rich sources of information such as imaging or genomics and the review and learning from retrospective cases. Altogether, we call this intelligent software system a Decision Support System, aiming to provide timely information and support for case review in the Breast Unit, based on clinical guidelines, experience from previous cases, advice for non-compliant decisions and exploitation of retrospective cases.
Furthermore, DESIREE develops imaging biomarkers, characterizing breast tissue and tumor and ultimate real-time visualizations based on image segmentation and analysis that provide insight into the disease and its progression, helping decision making. Last, but not least, DESIREE has developed a novel tool, based on a complex physiological model, for predicting the aesthetic long-term outcome of the Breast Conservative Therapy intervention (resection of the minimal amount of tissue necessary instead of complete mastectomy), which is useful both by the the patient and the surgeon.
Our ultimate goal, is to provide intuitive and intelligent software tools that improve the case review and management of the breast units and ultimately are capable to impact the patients by providing the necessary information to ultimately optimize treatments and outcomes.
DESIREE has developed a set of software providing decision support for the breast units. It allows multidisciplinary case review before, during and after the breast unit meetings. The DESIREE Information Management Systems, displays all the relevant case information in a structured manner and according to medical specialities, allowing to review and prepare the case during the meetings and organize the clinical sessions.
Developing a DSS for the breast unit has required an extensive amount of knowledge modelling of the breast cancer domain. The DESIREE knowledge model incorporates such knowledge in a digital model implemented as an ontology.
The core of the DESIREE Decision Support System interface is intended to be used during breast unit meetings for multidisciplinary case management. It shows only the relevant clinical and visual information (i.e. after medical image processing) for the current case and therapeutic scenario.
The DSS incorporates three complementary decision support paradigms: 1) The guideline-based DSS, 2) the experience-based (EXP-based) DSS and 3) Case-based DSS.
The DESIREE visual analytics and pattern analysis tools are intended to be used for retrospective case evaluation, allowing exploitation of generated real-world evidence.
DESIREE has developed novel methods for breast and tumour tissue characterization from digital mammography (DM), digital breast tomosynthesis (DBT) and magnetic resonance imaging (MRI) allowing to assess risk and prognosis.
DESIREE is also connected to the Genesystem platform from Sistemas Genómicos, allowing access to a large genomic variant database for consultation.
Another relevant development is the multi-scale modelling of breast conservative therapy (BCT) which provides a simulation of the breast contour after healing. The resulting visualization provides both the patient and surgeon a feedback of the possible cosmetic appearance (i.e. perceived symmetry), defects and scar after wound healing.
After integration of the different subsystems, extensive technical validation has been performed in order to prepare the system for clinical validation.
Dissemination activities have been very relevant during the project life. Scientific impact has been achieved by different journal publications, some in the first quartile.
Developing a DSS for the breast unit has required an extensive amount of knowledge modelling of the breast cancer domain. The DESIREE knowledge model incorporates such knowledge in a digital model implemented as an ontology.
The core of the DESIREE Decision Support System interface is intended to be used during breast unit meetings for multidisciplinary case management. It shows only the relevant clinical and visual information (i.e. after medical image processing) for the current case and therapeutic scenario.
The DSS incorporates three complementary decision support paradigms: 1) The guideline-based DSS, 2) the experience-based (EXP-based) DSS and 3) Case-based DSS.
The DESIREE visual analytics and pattern analysis tools are intended to be used for retrospective case evaluation, allowing exploitation of generated real-world evidence.
DESIREE has developed novel methods for breast and tumour tissue characterization from digital mammography (DM), digital breast tomosynthesis (DBT) and magnetic resonance imaging (MRI) allowing to assess risk and prognosis.
DESIREE is also connected to the Genesystem platform from Sistemas Genómicos, allowing access to a large genomic variant database for consultation.
Another relevant development is the multi-scale modelling of breast conservative therapy (BCT) which provides a simulation of the breast contour after healing. The resulting visualization provides both the patient and surgeon a feedback of the possible cosmetic appearance (i.e. perceived symmetry), defects and scar after wound healing.
After integration of the different subsystems, extensive technical validation has been performed in order to prepare the system for clinical validation.
Dissemination activities have been very relevant during the project life. Scientific impact has been achieved by different journal publications, some in the first quartile.
We have already set the grounds for the first known decision support system operating in a breast unit. DESIREE is pioneer in developing DSS paradigms to exploit this information.
The GL-based DSS provides different recommendations based on clinical guidelines of choice.
The Exp-based DSS is a pioneer concept, which accumulates real-world evidence based on the complex DESIREE knowledge model. It provides additional recommendations based on similar cases with non-compliant decisions.
The Case-based DSS incorporates novel paradigms for case comparison based on similarity metrics and a novel rainbow-box visualization tool, showing similar cases with different decisions and variables in common with the current case.
Due to the good feedback provided by clinicians and hospital managers, the interest of exploitation by different companies and the emerging related softwares, we believe that such a tool may have considerable impact.
DESIREE has also performed extensive work on breast image analysis, including novel methods for breast tissue segmentation and characterization in DM, DBT and MRI, characterization of masses in MRI and the most extensive validation of a method for fully automatic pectoral muscle segmentation.
A first outline and prototype implementation for real-time augmented image visualization during case-review has been carried out. This approach differs from current image-based softwares more focused on diagnostic by radiologist than on case review and exploration by the BUs.
Most of the work being done in the BCT physiological modelling part is a pioneering work in terms of predictive visualization of multi-scale physiological processes after breast conservative therapy.
With respect to radiobiological modelling, a specific tool has been developed to predict and assess the toxicity impact of the treatments.
The GL-based DSS provides different recommendations based on clinical guidelines of choice.
The Exp-based DSS is a pioneer concept, which accumulates real-world evidence based on the complex DESIREE knowledge model. It provides additional recommendations based on similar cases with non-compliant decisions.
The Case-based DSS incorporates novel paradigms for case comparison based on similarity metrics and a novel rainbow-box visualization tool, showing similar cases with different decisions and variables in common with the current case.
Due to the good feedback provided by clinicians and hospital managers, the interest of exploitation by different companies and the emerging related softwares, we believe that such a tool may have considerable impact.
DESIREE has also performed extensive work on breast image analysis, including novel methods for breast tissue segmentation and characterization in DM, DBT and MRI, characterization of masses in MRI and the most extensive validation of a method for fully automatic pectoral muscle segmentation.
A first outline and prototype implementation for real-time augmented image visualization during case-review has been carried out. This approach differs from current image-based softwares more focused on diagnostic by radiologist than on case review and exploration by the BUs.
Most of the work being done in the BCT physiological modelling part is a pioneering work in terms of predictive visualization of multi-scale physiological processes after breast conservative therapy.
With respect to radiobiological modelling, a specific tool has been developed to predict and assess the toxicity impact of the treatments.