Optimal treatment for patients with solid tumours in Europe through Artificial Intelligence

Demands of cancer care in Europe continue to increase significantly, the number of incident cancer cases in Europe is projected to increase by 14.1% by 2030 [1]. This leads to a growing demand for innovative cancer treatments among patients, payers, physicians, and society. At the same time, the complex biology of cancer is getting more deciphered and as a result pharmaceutical companies are developing a multitude of new therapeutic agents. This includes, but is not limited to, novel kinase inhibitors, immunotherapy combinations, and cell therapies.

This trend for new, effective therapies creates more treatment options for patients. However, it confronts physicians with an exploding amount of potential therapeutic options, which each need to be understood and adopted effectively. Numerous factors such as genetic analysis, specific tumour biology, and biomarkers have a growing influence on clinical decision-making. To become familiar with the huge volume of available information, physicians need to continuously learn about medical guideline changes and marketed treatments. In conclusion, future decision-making processes will become ever-more complex such that the treatment choice could be sub-optimal or
even wrong. Furthermore, some patients have disease characteristics where evidence of guideline recommendations is scarce and physicians lack information about real-world treatment outcomes. Hence, the challenges to be addressed are assisted guideline-based decision making and the discovery of knowledge about treatment outcomes in real world settings. As the latter challenge requires analysis of large data sets, the application of Artificial Intelligence (AI) will be a key technology.

To ensure the challenges can be properly addressed, and ensure the innovations reach the physicians and
patients, a public-private partnership is necessary, including the following actors:

  • patient organizations and regulatory authorities to specify the requirements and boundaries of AI-driven
    data processing, data security and privacy as well as individual data ownership
  • medical societies to provide the network of participating in- and outpatient clinics to enable data access
  • medical experts/institutions to specify AI approaches, validate the decision support and set the
    requirements for general acceptance
  • life-science companies to contribute study data for the evaluation of therapeutic approaches, as well as
    expertise in data mining and data-set merging
  • SMEs for infrastructure set-up, data management and data security, AI-driven data processing and merging
    of unstructured information, visualization and user experience design.


The scope of this call topic is to establish guideline-based decision support and platform solutions to generate knowledge discovery for breast, lung and prostate cancer with applicability to other indications, in several European ‘model’ regions. The results obtained from these model regions are expected to be of relevance to countries with different socioeconomic backgrounds. The funded action will focus on breast, lung and prostate cancer to ensure a high number of cases per year, a high unmet medical need, multiple available therapeutic options and a fastevolving treatment environment. The three core objectives of this call topic are as follows:

Objective 1: Establish a guideline-based decision support for prioritized indications
Development of a decision support tool that automatically extracts relevant clinical information from electronic health records (EHRs) and facilitates guideline-compliant treatment approaches for the defined solid tumours.

Objective 2: Establish a structured and interoperable data platform to unlock real-world-data potential in an
oncology network

A major requirement for the provision of patient-specific treatment is the availability and the harmonization of extensive patient data across in-patient (e.g. academic centers, teaching hospitals) and out-patient (community and private practices) settings - stored in a structured format, ready to be used and interoperable. The successful consortium should address this need by involving relevant and available regional/national networks of in/outpatient clinics providing access to their data, for instance with the inclusion of medical societies.
Easy-to-use new platforms that enable the gathering and granular storage of clinical data to offer a foundation for data analysis and knowledge discovery need to be established. The real-world data platforms should include prospective data from electronic health records, structured data from (non)interventional studies provided by members of the pre-identified industry consortium as well as potentially registry data.

Objective 3: Leverage the real world data gathered by the action to establish an AI-knowledge base and support treatment decisions for prioritized indications
The funded action will develop a disease-specific (breast, lung and prostate cancer) AI system that facilitates the discovery of novel medical knowledge. This includes hypothesis generation about optimal treatment sequences and prognostic features that can be validated in clinical research. The output will strongly support building the European health data space and improve the quality and acceptance of AI-generated evidence in decision making in research and healthcare delivery. It will also set the foundation for explainable AI approaches necessary for personalized treatment.

Read more here.