We engage and closely work with medical professionals, academia, governments, NGOs, and industry to solve relevant and significant problems which plague healthcare systems. We work in the following six domains which approximately map to six research teams which comprise Sano’s research core.

Clinical Data Science

…conducts research into improvements in decision making at all stages of patient interaction with the healthcare systems: analysis of data to provide differential diagnoses with predictions of uncertainty, data-driven estimation of risks of future morbidity, utilisation of patient-specific data to determine the optimal path of treatment considering also predictions of treatment side effects, optimising risks and benefits of hospitalizations, and treatment monitoring – analysis of therapy effectiveness indicators or outcomes in treatment management. Cooperation with the Personal Health research strand will bring advances in real-time treatment data analysis to provide recommendations about effectiveness and the likelihood of success based on patient-specific data.Read more

The decision support AI will use data science machine learning and deep learning models based on a combination of one or more of the following patient-specific data types: patient presentation, medical history, laboratory results, clinical notes, medical imaging data, molecular biology data, socioeconomic factors, patient-specific simulation data, and other available relevant data.
Modelling and Simulation
Algorithmic Decision Science

Computer Vision Data Science

…improves upon already impressive advancements of computer-aided interpretation and analysis of visual information in healthcare. In recent years, convolutional neural networks and deep learning have revolutionized computer vision (CV) applications and they are positively impacting medical practice, following success in many other trades. Automatic interpretation of radiologic images improves patient-specific accuracy and reduces radiologist burnout, either as support to reach accurate and consistent expert-level decisions, or as a second opinion about a potential pathology. Applications also include image interpretation during procedures such as endoscopy, AI-assisted surgery, automatic interpretation of digital pathology.Read more

AI systems can analyse thousands of hours of surgical videos to develop patient-specific surgical procedures, by, for instance, detecting anatomical structures such as major nerves and helping with preventing nerve damage during the procedure. Feeds from classical or infrared cameras can be used in conjunction with computer vision methods for patient safety, logistical optimization, monitoring the spread of infections, health assessment, mobility, and similar. Image-based information can also be combined with other types of medically relevant information to provide synergistic decision support based on visual and non-visual cues. The interrelations between the phenotype, determined on the basis of medical imaging, molecular biology, and other data are also of high relevance to precision medicine.

Personal Health Data Science

…concentrates on research aimed at shifting healthcare philosophy from reactive to proactive, as an integral part of the Healthcare 2050 vision. Most people are now constantly “online”, consuming and generating ever growing amounts of data, and this trend is expected to continue. Undoubtedly, strong potential exists for creating a positive change in how people manage and influence their health through these data. The long-term goal of this effort is to create a personal health tracker that collects personal health information and provides prediction of our future wellbeing. Examples of such information include longitudinal medical data, genetics, behavioural information, stress levels, dietary habits, environmental and socio-economic factors and other. Apart from computing existing risks, such a tracker could also predict them if the input changes, thus empowering individuals to positively affect their own health, prevent disease and, in consequence, live longer and in better health.Read more

Collaboration with healthcare insurers will bring access to historical data about interaction of a vast number of patients with the healthcare system, census, death records, environmental data – which will be used for analytics and to create predictive machine learning models of personal health. Sano will also work with medical institutions to obtain detailed medical data, including EHR imaging and genetics, of carefully selected cohorts around Poland. By strategically combining limited, but representative samples of population, our researchers will develop detailed machine learning models at the national level. This could then be recreated for a different country, or even on a larger scale (like, e.g. the EU).

Data Science
In Silico Techniques

Modelling and Simulation

…expands research from the Virtual Physiological Human initiative. Unlike fully data-driven machine learning data science approaches a rich assortment of complementary modelling approaches will be utilized, including 3D (e.g. Navier-Stokes solvers, Lattice Boltzmann approaches), 1D and 0D models of fluid mechanics, Finite Element Analysis, Growth and Remodelling frameworks and Agent-Based Models of structural mechanics and others. This research area will be concerned with the development of fundamental modelling tools and workflows, by simulating physiology and predicting multi-morbidity through integration with data science technologies.

Read more
This team will also investigate applications of in-silico clinical trials which concerns modelling and simulation of devices, drugs and interventions to be evaluated over trial populations covering complete target treatment populations, hugely reducing the time and cost of traditional clinical trials through the use of computing. Significant research questions remain around these approaches, particularly over missing and inaccurate data, which could be addressed by combining modelling and simulation approaches with Data Science and Healthcare Informatics for input data distribution and accuracy. Furthermore, biology, biochemistry, physics models can be developed in the attempts to explain AI decisions, to improve clinical acceptance largely black-box deep learning approaches of data science.

Health Informatics

…manages the use of patient healthcare information and deals with the resources, devices, and methods required to optimize acquisition, storage, retrieval, and use of information in medicine. This team concentrates on a new generation of approaches to medical communication and incorporation of output of computational methods (data science, in-silico methods) in medical workflows. New models of information exchange between existing decision agents (patients, their families, doctors, care teams) are considered and investigated. Intuitive interfaces for improved understanding of health-related data and AI insights will be an important part of this effort.

Read more
This team will, among others, investigate new computer interfaces for enhanced perception of multi-modality, multi-source medical data, including 3D virtual interactive environments. Using such environments for advanced remote interactive communication without the need for physical co-presence will improve today’s telemedicine, also in the context of pandemic threats. For smaller medical centres, and rural area care, these solutions will allow for remote virtual interaction between patients and specialists, and advanced, data-driven specialist consultations. The Health Informatics team will also investigate optimal decision-making when heterogeneous information sources – machine recommendations, expert human knowledge, and patient preferences – are combined during the health-related decision-making process.

Healthcare Informatics
High Performance Computing

Extreme-scale Data and Computing

…tackles the new fundamental challenges for computer science and engineering research, posed by the unique combination of Modelling and Simulation with Health Data Science and Health Informatics. The computational and data processing needs of Sano research will push the boundaries of current state-of-the-art infrastructures for AI, HPC, big data and cloud computing. This includes alignment of traditional HPC systems with big data analytics and ML/AI workloads, using CPU, GPU, many-core, hybrid, virtualized and containerized environments with the computational needs of systems required to deliver patient-specific care at timescales appropriate for clinical use. Moreover, health data science will benefit from novel approaches in distributed computing and security research, such as Federated Learning, Blockchain, Differential Privacy or Encrypted Computation, which can be applied to medical data in a secure and privacy-conscious manner.

Read more
New developments in computer hardware, programming models, cloud computing and emerging services will influence development, deployment and execution of computational and AI models at extreme scale, requiring constant evaluation of new technologies and platforms, experimenting with novel approaches, and prototyping new solutions. Development of systems operating in clinical, research, HPC, Big Data/AI environments will require novel, transparent techniques, and the delivery of state-of-the-art in health data science and in-silico techniques will require exascale computing resources. Sano will become a driver for developments within EU-level HPC and cloud initiatives, including PRACE and EuroHPC.

Research teams are currently being formed and leaders recruited