Medical Imaging Informatics: Towards a Personalized Computational Patient - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Journal Articles IMIA Yearbook of Medical Informatics Year : 2016

Medical Imaging Informatics: Towards a Personalized Computational Patient


Medical Imaging Informatics has become a fast evolving discipline at the crossing of Informatics, Computational Sciences, and Medicine that is profoundly changing medical practices, for the patients' benefit. Keywords Medical Imaging Informatics, personalized computational patient, computational anatomy, computational medicine Yearb Med Inform 2016;xxx Published online xxx In 2002, my preface to the IMIA Yearbook was entitled " From Digital Anatomy to Virtual Scalpels and Image-Guided Therapy ". It was announcing a revolution in medicine brought by the extensive use of medical image computing to better assist the diagnosis and therapy of the patient. Today, the promised revolution is here: Medical images are omnipresent at the hospital, and Medical Imaging Informatics is required more than ever to exploit their flood of information. All around the world, medical image computing is used to extract the clinically relevant information from medical images, and to present this information in a way that is clinically useful to the physician. This is mainly done through the construction of a computational and personalized model of the patient. Building a computational model of the human body requires dedicated algorithms that take into account a thorough knowledge of the human anatomy and physiology. Huge progress has been made during the last decades to describe and simulate the structure and functions of organs thanks to advanced mathematical, biological, physical, and chemical models of the living tissues at various scales from the nanoscopic (molecular) to microscopic (cellular), mesoscopic (tissue), and macroscopic (organic) scales. Computational models of the human body rely on a set of parameters that allow, for instance, to specify the structure and function of organs. Generic models are based on average parameters estimated over a population. Confronted to in vivo anatomical and functional images and signals of a singular patient, those parameters are adjusted by efficient personalization algorithms in order to reproduce more precisely the observed structures and functions leading to the personalized computational model of this particular patient. The personalized computational model of the patient is then used to provide quantitative and objective measurements on the patient's condition to better assess the diagnosis. It is also used to predict a pathological evolution resulting in a better assessment of the prognosis. Finally, the computational model of the patient is extensively used to plan and simulate the effect of a therapy, in order to optimize its actual delivery. These three steps-computer aided diagnosis, prognosis, and therapy-announce the fast development of the computational medicine at the service of the physician. The tremendous progress of Medical Imaging Informatics also accompanies the evolution of normative and reactive medicine towards a more personalized, precise, preventive, and predictive medicine. This progress relies on numerous algorithmic advances in medical image analysis and inverse problem solving. It also relies on continuous advances in the modeling of human anatomy and physiology. It benefits from the improvement of medical image acquisition techniques, and from the introduction of new imaging modalities at various scales. It is supported by the regular


Medical Imaging
Fichier principal
Vignette du fichier
Ayache-IMIA-Yearbook-2016-v6-preprint.pdf (62.62 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-01320985 , version 1 (30-05-2016)



Nicholas Ayache. Medical Imaging Informatics: Towards a Personalized Computational Patient. IMIA Yearbook of Medical Informatics, 2016, 25 (Suppl. 1), pp.S8-S9. ⟨10.15265/IYS-2016-s002⟩. ⟨hal-01320985⟩


306 View
275 Download



Gmail Facebook X LinkedIn More