30 Mar 2020

CF team retains over 1.25M euros for Artificial Intelligence for Health Imaging project

The Champalimaud Foundation is pleased to announce that ProCAncer-I: An AI Platform integrating imaging data and models, supporting precision care through prostate cancer’s continuum, submitted to the Horizon 2020 DT-TDS-05-2020 topic, has been retained for funding with an amazing score of 15 out of 15 points.

CF team retains over 1.25M euros for Artificial Intelligence for Health Imaging project

The Champalimaud Foundation is pleased to announce that ProCAncer-I: An AI Platform integrating imaging data and models, supporting precision care through prostate cancer’s continuum, submitted to the Horizon 2020 DT-TDS-05-2020 topic, has been retained for funding with an amazing score of 15 out of 15 points.

ProCAncer-I is led by Fondazione del Piemonte per l’Oncologia, Italy, and involves 20 partners from 11 different countries and a total budget of almost 10 Million EUR over 48 months. 1,246,125 EUR of this budget will be retained by the Champalimaud Foundation’s branch of the project, led by Scientific Manager of the Consortium and Team Leader Dr. Nickolas Papanikolaou, and involving Dr. Jorge Fonseca (Urology Unit), Dr. António Beltran (Pathology Unit), Dr. Celso Matos (Imaging Department) and the Hardware platform.


Nickolas Papanikolaou. The Scientific Manager of the ProCAncer-I Consortium

The main objectives of the project are to develop an imaging repository comprising of more than 17,000 MRI examinations of patients with prostate cancer and to harness these imaging data to develop AI models that can address 8 important clinical questions related to treatment efficacy, toxicity and management of patients with prostate cancer. State of the art AI methods will be developed based on vendor-specific and vendor-agnostic methods using federated learning.

The Champalimaud Foundation is a core partner in the project that will contribute imaging data as a clinical partner, but most importantly will coordinate the AI model development of the consortium.
Upon receiving the news that the project had retained funding, Dr. Papanikolaou said: “I’m very happy to realise that the efforts of the last 4 years produced such an outstanding outcome. If I’m not mistaken, this is the first large scale AI project in the medical imaging domain obtained by an institution based in Portugal. I want to acknowledge the CCIG [Computational Clinical Imaging Group] for materialising, through hard work, the vision and transmission of the message that young Portuguese researchers can be competitive when working in the proper environment. Last but not least, I want to thank the Champalimaud Foundation Board for entrusting us and helping us by providing the means to achieve such a remarkable result.”

—————————————————————————————————————————-
Project abstract

In Europe, prostate cancer (PCa) is the second most frequent type of cancer in men and the third most lethal. Current clinical practices, often leading to overdiagnosis and overtreatment of indolent tumors, suffer from lack of precision calling for advanced AI models to go beyond SoA by deciphering non-intuitive, high-level medical image patterns and increase performance in discriminating indolent from aggressive disease, early predicting recurrence and detecting metastases or predicting effectiveness of therapies. To date efforts are fragmented, based on single institution, size-limited and vendorspecific datasets while available PCa public datasets (e.g. US TCIA) are only few hundred cases making model generalizability impossible. The ProCAncer-I project brings together 20 partners, including PCa centers of reference, world leaders in AI and innovative SMEs, with recognized expertise in their respective domains, with the objective to design, develop and sustain a cloud based, secure European Image Infrastructure with tools and services for data handling. The platform hosts the largest collection of PCa multi-parametric (mp)MRI, anonymized image data worldwide (>17,000 cases), based on data donorship, in line with EU legislation (GDPR). Robust AI models are developed, based on novel ensemble learning methodologies, leading to vendor-specific and -neutral AI models for addressing 8 PCa clinical scenarios.

To accelerate clinical translation of PCa AI models, we focus on improving the trust of the solutions with respect to fairness, safety, explainability and reproducibility. Metrics to monitor model performance and a causal explainability functionality are developed to further increase clinical trust and inform on possible failures and errors. A roadmap for AI models certification is defined, interacting with regulatory authorities, thus contributing to a European regulatory roadmap for validating the effectiveness of AI-based models for clinical decision making.

Role of the Champalimaud Foundation in the project

The Computational Clinical Imaging Group will be responsible for the development of a Master Model using the bulk of retrospective imaging data (WP5), while it will develop one of the Vendor specific models (Philips) (WP6). In addition, it will participate in the distributed learning to develop the final Vendor neutral model. Individual AI models will be developed to detect cancer with high accuracy, to characterize it according to its biological aggressiveness and to predict the risk of tumor recurrence, of treatment response in case of radiation therapy and of radiation-induced urinary toxicity. To help developing the AI models, the Urology and Pathology Units will provide clinical, pathology and imaging data, as well as clinical outcomes of approximately 300 patients with prostate cancer. These models will then be validated and tested by CCC’s units, using approximately 150 prospective cases per year. CCIG in collaboration with Urology, Radiology and Pathology units of CF will be the only partner in the consortium to run a subproject concerning 100 patients that will undergo radical prostatectomy, where imaging and digital histopathology will be compared side by side using individualized molds produced with a 3D printer, to perform biological validation of the models using digital histopathology as the ground truth. In addition it will be involved in management and coordination (WP1), scientifically coordinating the project, in ethical and legal aspects (WP2), in dissemination and communication actions (WP3), in the infrastructure design (WP4) and has a minor role in the exploitation of the project (WP8).