Skip to main content
European Commission logo print header

Big Data and models for personalized Head and Neck Cancer Decision support

Article Category

Article available in the following languages:

Big Data and visual analytics for head and neck cancer diagnosis and therapy

Head and neck cancers (HNCs) have a high death rate mainly due to diagnosis when the disease has reached an advanced stage. The BD2Decide project has harnessed cloud computing and new models to save lives.

Health icon Health

Risk factors for HNCs are smoking, excessive alcohol consumption and, as highlighted more recently, infection with the human papillomavirus (HPV). As the majority of patients are still detected in advanced stages (Stages III and IV), their treatment usually requires a variety of treatments, including surgery followed by post-operative radiotherapy and chemoradiation. Treatment for HNCs as it stands can cause disfiguring facial asymmetries, impact on speaking, swallowing and eating, and, due to its toxicity, severe morbidity as well as greatly deteriorate patients’ quality of life. Tito Poli of the BD2Decide project, coordinator and associate professor at the University Hospital of Parma, emphasises: “Clinicians are therefore faced with the need to choose the best curative treatment while at the same time minimising side effects and impacts on patients’ quality of life.”

An integrated solution with Big Data at its heart

BD2Decide has developed a solution to these as-yet unmet needs of physicians: a validated integrated clinical decision support system (CDSS). “This framework links together population-specific epidemiology, behavioural and environmental data, patient-specific multiscale data from genomics, pathology, and clinical and imaging data with available multiscale prognostic models and graphical visualisation tools,” outlines Poli. Research efforts by BD2Decide have resulted in various innovations. The researchers have introduced radiomics and combined genomics into the diagnostic and treatment decision process and into prognostic models. Considering all the multiscale factors involved in HNC progression and population data, the project has integrated and recalibrated different models. Image analysis from computer tomography, CT, and magnetic resonance imaging, MRI, has resulted in radiomic feature extractor tools. These can be utilised for combined radio-genomic prognostic modelling and the discovery of new personalised prognostic signatures. For all this new wealth of information, the project harnessed Big Data techniques to explore cloud computing and analysis techniques to uncover unexpected prognostic indicators. Set-up of Big Data and cloud infrastructure to collect and homogenise data was done in compliance with state-of-the-art standards.

Patient input vital

For patient management, the introduction of new collaborative concepts applied to decision-making through data sharing, prognostic modelling and data investigation was a major innovation with a resulting interactive patients’ co-decision aid. For clinical use, BD2Decide partners have implemented a ‘transparent’ data presentation and personalised prognostic modelling output. In addition to the CDSS tool, the team also developed a knowledge management system. BD2Decide has made significant progress beyond the state of the art by developing new accurate methods for HPV determination in oropharyngeal cancer. Identification and validation of a new genomic signature for oral cavity cancer may result in a patent.

Future direction on the research map

“Through more accurate and personalised prognosis prediction and treatment decisions, consistent with up-to-date clinical knowledge and patients’ preferences, BD2Decide is expected to contribute to reducing healthcare and social costs of HNCs,” concludes Poli. Moreover, care delivery will be optimised and unnecessary actions avoided. With the future of HNC research firmly fixed in their sights, BD2Decide clinical partners are negotiating with existing EU research infrastructures – the Italian node of the ELIXIR network. The aim is to systematically deposit the BD2Decide data set in a permanent repository, to share it with the scientific community and to foster further HNC research, along the lines of the EU’s European Open Science Cloud principles. Branching into the academic and training realm, the University of Parma is planning to add a course on digitalisation in health and Big Data analytics in clinical settings, the content of which is rooted in BD2Decide experience and results.

Keywords

BD2Decide, Big Data, cloud, head and neck cancer, HPV

Discover other articles in the same domain of application