Oncoradiomics is the first company offering a reliable decision support tool for selecting the right diagnosis and treatment options of patients and patient stratification for clinical trials based on standard imaging. We have developed a portfolio of RADIOMICSTM Plug-in software products and services which can be integrated into software packages of large establishments in the fields of radiology, radiation therapy, PACS and other healthcare IT software institutions.
RADIOMICSTM provides: quality of diagnosis, prognosis, ease of use and cost efficiency:
Quality of diagnosis:
- It provides probabilistic decision support tools to select the right diagnosis (e.g. isolated lung nodule…)
- In cases of uncertain diagnosis it advises the doctor of the next step and/or the follow-up
Quality of prognosis:
- It provides predictive and prognostic decision support tools to deliver the optimal treatment, i.e. the highest cure rates and minimal side effects
- It is based on an extensive clinical database obtained from internationally renowned institutions
Ease of Use
- It provides non-invasive, automated and adaptable tools based on images already taken
- No additional procedures are necessary since images are already taken for diagnostics, planning or guidance purposes
- It avoids overkill of examination for diagnosis
- It avoids over- and under-treatment
- It offers efficient clinical trial management
- It offers patient selection for new and expensive targeted therapies
RADIOMICS™ aims to offer through a dedicated service (including SaaS) model. Images can be uploaded to our web-based clinical grade SaaS service and can be analyzed by our clients themselves. We also provide tailored support with generating the data and a report with information on the tumor status. In addition, further options are customized releases on the base of customer needs.
At OncoRadiomics we offer tailor made consultancy support which will provide you with our expertise and tools to facility your Radiomics analyses. WE prefer to be involved at the writing stage of your protocol to ensure proper image acquisition. In addition and if required, we are also able to undertake the complete Radiomic analysis for your organization or for your specific project.
Radiomics for clinical use
Our first clinical grade application will to help the Doctor choose between curative treatment with or without adjuvant chemotherapy or palliative treatment in inoperable, non metastatic Non Small Cell Lung Cance.
For further information on any of our products or service, please contact us directly.
The Virtual Patient Avatar:
a digital guardian angel in the cloud
The future of medicine will be Personalized, Preventive, Predictive, and Participatory. This is the so called Precision Medicine.
However, there are at least three key challenges:
- The number of biomarkers, which aid the choice of best treatment, is increasing exponentially, and are developed independently of each other.
- There are an ever increasing number of therapeutic options.
- Due to the increased complexity of the medical decisions, patient involvement decreases and medical practice becomes more paternalistic.
We need a conceptual break through to tackle these issues. We believe it will be the virtual patient avatar or VPA which will be a synthetic entity of standardized biomarkers linked to each other through an ontology. The VPA will be fed from data generated directly from the general practitioner, the hospitals and the patient via smartphone and multiple nanosensors.
The VPA will have at least two levels of complexity: One data aggregation system- version for doctors, and one smart phone interactive system - version for the patient. The system will first focus on wellness and make the patient responsible for his health. It will calculate probabilities of diseases through simple dashboards, will proposes preventive actions and monitor them. It will be a type of digital guardian angel in the cloud, if you wish.
In case of a disease, we will emulate the clinical scenario by virtually treating the customized digital guardian angel in a virtual hospital with all the different types of treatment available. The best probable outcome will be calculated using advanced mathematical models updated continuously with privacy-preserving distributed machine learning approaches.
The patient will control the information he gets and he will be encouraged to participate in the medical decision using patient decision aids tools which will help him to experience the different treatments using virtual reality.
A Virtual Patient Avatar promises to sharply reverse the ever escalating costs of health care, to empower patients, leverage all existing medical knowledge, and to improve patient outcomes.
It will make ‘Big Data’ in health care small and useful for you, so that you can take control of your own destiny.
DISTRIM: a tool allowing Big Data for imaging
DISTRIM is currently under development. More information about DISTRIM can be found below.
Biomarkers including Radiomics-based QIBs strongly rely on the large independent data sets. There is a need to continuously update and refine the radiomics signatures which increase the predictive power of these algorithms. In order to do so, there is a need to have access to many different and large data sets. However, these data sets are distributed among many different centres and access is restricted due to privacy reasons. This means that these data sets cannot leave the hospitals.
Distributed learning of medical data (DISTRIM™) is a new approach to use distributed medical data sets from all over the world without the need of data leaving the hospital firewalls. DISTRIM™ will enable us to mine large amount of images without the images leaving the firewall of the hospitals. Distributed learning is an emerging topic in the field of machine learning; it was created and studied in order to achieve performance optimisation and solving optimisation problems with a large number of variables and items as a typical implementation of parallel computing. This approach can be exploited both to the aim of preserving data privacy and to protect data property. The added values of Distributed Learning are: 1) the privacy issues resolutions: the data do not leave each centre, less organisational and bureaucratic issues, with possibility to exploit validated standard. It will facilitate the participation of more centres in the data sharing and 2) Scalability of the computation processes: map-reduce approach.