This project has a 48-months duration and is structured in 10 Work Packages (WP):
- WP1: CCRI will lead WP1, which is focused on clinical scenarios and use requirements. The overall objective of WP1 is to design a Decision Support System (DSS) for cancer management with advanced functionality and usability, under a user-centric approach guided by our clinical partners (European Key Opinion Leaders in Paediatrics Oncology), aligned with their current work flows and aimed to conquer trust and gain acceptability by the clinical practitioners.
- WP2: The objective of the WP2 is to develop a hybrid open datacloud and processing middleware which comprises the use of open public and private clouds, and specific processing resources, in particular High Performance Computing (HPC) and cloud, to support predictive tools for assisting diagnosis, prognosis, therapies choice and treatment follow up, based on the use of novel imaging biomarkers, in-silico tumour growth simulation, advanced visualisation of predictions with weighted confidence scores and machine-learning based translation of this knowledge into predictors for the most relevant, disease specific, Clinical End Points. These tools (imaging biomarkers, models for in-silico medicine research, advanced visualisation tools) will be validated in the application context of two paediatric cancers, Neuroblastoma (NB, the most frequent solid cancer of early childhood) and the Diffuse Intrinsic Pontine Glioma (DIPG, the leading cause of brain tumour-related death in children).
- WP3: Implementation of extraction/de-identification/routing pipeline for retrospective data (ME), to be also used for prospective studies and interoperable data repositories (UNIPI, ME, Clinical partners).
- WP4: La Fe University and Polytechnic Hospital and La Fe Health Research Institute – Biomedical Imaging Research Group and the Paediatric Oncology Unit (Spain) will lead WP4 with participation from QUIBIM (Spain), Universita di Pisa (Italy), St. Anna Kinderkrebsforschung e.V. Children’s Cancer Research Institute (Austria) and Uniklinik Koeln (Germany). The goal of WP4 is to deliver the imaging and molecular biomarkers in patients with neuroblastoma and DIPG. The subtasks of this work package are the following: (1) Selection of biological biomarkers. Molecular, genetic and biochemical biomarker panels are defined for the diagnosis of neuroblastoma and DIPG, including genomics through Next Generation Sequencing, serum-urine-CFS biomarkers (liquid biopsy), pathological data, MRD data, immunological profile, metabolomic and pharmacogenetic data; (2) Selection of imaging biomarkers. Methodologies for MR, PET/CT and MIBG scan image analysis, clustering and visualisation are defined, and recommendations for imaging data collection and annotation that facilitate extraction of reusable and meaningful imaging biomarkers are established. Models for automated segmentation of organs, lesions contained and the quantification of imaging biomarkers in the acquired images are developed by using convolutional neural networks. Artificial intelligence algorithms are evaluated to create models to assist radiomics analysis with the objective to assist diagnosis and prognosis. These models are interoperable with clinical PACS, for images acquired by the principal commercial systems (Philips, Siemens, General Electric and Toshiba). Recommended imaging protocols for data collection and real-time quality control procedures are generated, and a rigorous protocol conformity and quality assurance is expected to be implemented in order to assess confidence level achievable; (3) Implementation of diagnosis module. The PRIMAGE platform module for personalised diagnosis is delivered by integrating the use of imaging biomarkers with the biological biomarker panels (genetic, molecular and biochemical biomarkers) and delivering weighted confidence scores for supporting decisions by clinical practitioners. The platform allows for the data mining of all the radiomic features extracted through a specific environment which is also compatible with Excel, SPSS and R file formats for statistical analysis.
- WP5: The aim of WP5 is to deliver in-silico models of solid tumour growth for Neuroblastoma and Diffuse Intrinsic Pontine Glioma using multiscale simulation frameworks to couple model at subcellular scale, to cell and tissue in order to complete organ models, enabling assessment of a radiotherapy and chemotherapy treatment on tumour progression for a specific patient.
- WP6: WP6, “Translation to usable clinical knowledge,” combines retrospective patient data (WP3) with novel imaging biomarkers (WP4) and innovative In-Silico models (WP5) allowing medical doctors and other domain experts to gain insight into this Big Data. WP6 creates a visual interactive platform for exploratory analysis and sense-making integrating the combined Big Data and the domain users expertise. Generating knowledge requires human involvement (figure: The “Knowledge Generation Model for Visual Analytics” describes the interaction between the human and the machine to generate knowledge from data.
[D. Sacha, A. Stoffel, F. Stoffel, B. C. Kwon, G. Ellis, and D. A. Keim; IEEE Transactions on Visualization and Computer Graphics (Proceedings Visual Analytics Science and Technology 2014), 20(12):1604 – 1613 , DOI: 10.1109/TVCG.2014.2346481 , 2014.]):
- WP7: QUIBIM is a worldwide developer and provider of advanced computer vision and artificial intelligence algorithms for radiology through on-premise and cloud solutions. The company has extensive knowledge on requirements for interoperability on the clinical settings and requirements as a cloud-based infrastructure. Furthermore, as a company serving the healthcare sector, development is oriented to fullfil clinical user requirements promoting that platform operations are well aligned with the actual workflows in hospital environments, particularly in oncology departments and related diagnostic services.
QUIBIM will lead WP 7, focused on the architecture, the integration and the final design of PRIMAGE platform, including an exhaustive lab testing of the beta version ready to be validated in the context application of paediatric cancers such as neuroblastoma (NB) and diffuse intrinsic pontine glioma (DIPG). In addition, the platform will include advanced visualisation tools, providing visualisation of uncertainty and prediction reliability of in-silico models, to facilitate clinical decision-making.
- WP8: Clinical validation, with leading paediatric oncology departments at hospitals in Spain, Germany and Austria. (WP8 needs to run during the whole project to recruit the expected cohort; the prospective data will be used for platform validation only during year 4).
- WP9: WP9 encompasses the communication, dissemination and exploitation needs of the PRIMAGE Project. SIOP Europe, together with the partners, will develop the dissemination and communication tools. Furthermore, networking, dissemination and communication actions will be designed to ensure that PRIMAGE is networked with relevant national and international projects, initiatives and networks and that project results are fully disseminated to all relevant audiences. MATICAL, together with the partners, will oversee innovation management, including ownership, access rights, decision making procedures, publications, and IP management, as well as business, sustainability and exploitation planning.
- WP10: This WP comprises the steering of the day to day running of the project, ensuring that the work programme adheres to its schedule, fulfills its objectives, and the project meets the needs and expectations of all the partners. The Coordinator together with the Project Manager will be active in steering the evolution of the project by closely monitoring the work performed by each partner. Both will assume responsibility for ensuring that the project management procedures are being correctly implemented and that all deadlines and obligations are met.