Origin of the Hanarth Fonds
The Hanarth Fonds was founded on September 28, 2018, created using the legacy of Arthur de Prado, founder and former CEO of ASM International. He had a desire to support scientific research aimed at the cause and cure of cancer, partly because his first wife died of lung cancer. More information on Arthur de Prado can be read here.
Aim of the Hanarth Fonds
To promote and enhance the use of artificial intelligence and machine learning to improve the diagnosis, treatment and outcome of patients with cancer. In this context the Hanarth Fonds supports scientific research concerning (primarily, but not exclusively, rare types of) cancer.
Cancer is on the rise. Ten years ago, 12.7 million people worldwide were living with this disease. By 2030, 21.7 million people are expected to be affected, mainly because of a growing and ageing population as well as lifestyle changes. Despite clear progress in early detection and treatment of the disease there are certainly still unmet needs in oncology.
Artificial intelligence in oncology
Medical knowledge in the field of oncology is expanding rapidly with novel therapies at hand continuously, based on advances in immunology, genetics and system biology. This leads to an exponential accumulation of electronic health record data. Furthermore, there is a demand for personalized treatment of cancer. Every patient represents a big data challenge, with vast amounts of information on current state and past trajectories. All this makes medical decision-making increasingly complex and decisions are frequently sub-optimal.
In the near future, medical decision-making can clearly benefit from big data analysis with artificial intelligence and machine learning. Machine learning is a subfield in computer science that explores the study and development of algorithms which can learn from and make predictions on the basis of data. It is the science of getting computers to act without being explicitly programmed. The number and variety of available data is still growing in all areas, like in common or rather rare adult cancers.
Rare types of cancer
Of all new human cancers, 22% are rare cancers, defined as an incidence of less than six new cases per 100,000 people per annum. These rare cancers are linked with worse survival rates than common types. This is mainly because of a delay in obtaining an accurate diagnosis, inadequate treatments given as initial treatment in curative phases and the restricted opportunities for patients to participate in clinical trials, because of the lack of financial support from both academic and industrial sponsors for dedicated/specific trials for this group. The increasing fragmentation of all cancers into molecular subgroups with their own specific therapeutic strategies implies a substantial increase in the number of rare adult cancers and their associated socio-economic burden.
Know-how from common cancers may be crucial as input to improve the analysis of rare cancers. Currently computational processing is becoming more powerful and data storage is becoming more affordable. This enables a quicker development of algorithms and models that can analyze bigger, more complex data and deliver faster, more accurate results. It is time to develop tools based on this knowledge to support clinicians to interpret this data rapidly and use it in daily practice.
- automatic image processing for classification, detection and radiomics;
- predictive and prognostic models that can be used for personalized medicine;
- models based on different input parameters, which can help signaling safety issues;
- identification of rare disease patients and outcome related factors from electronic health records;
- application of machine learning on connected databases for different rare cancers with the potential to help extract latent knowledge and patterns present in these databases. (such latent knowledge and patterns of outcome are often missed by clinicians, because they typically rely on patterns learned from patients with frequently occurring cancers encountered in their own clinical settings).
New players are needed
A multi-disciplinary team enriched with some new players is necessary to apply machine learning tools to established data sets of rare cancer subtypes and to use the results in daily practice. These new players are clinicians or scientists* with profound background or interest in machine learning (or artificial intelligence in general) and a partner (or partners) with background in machine learning (or artificial intelligence in general) and interest in medicine, who can contribute meaningfully to the development and evaluation of algorithms.
The current medical education system and clinical practice do not fulfil these needs and fail to educate doctors in data science, statistics, or behavioral science necessary to develop, evaluate and apply algorithms in a clinical practice.
Medical or non-medical