Section: Application Domains

Clinical Applications

Some of the scientific challenges described previously can be seen in a general context (such as solving constraints between different types of objects, parallel computing for interactive simulations, etc.) but often it is necessary to define a clinical context for the problem. This is required in particular for defining the appropriate assumptions in various stages of the biophysical modeling. It is also necessary to validate the results. This clinical context is a combination of two elements: the procedure we attempt to simulate and the objective of the simulation: training, planning or per-operative guidance. Below are a series of applications we plan to develop. The choice of these applications is not random: the clinical procedures we target are all technically challenging, they highlight various parts of our research, and often they represent an ideal testbed for transitioning from training to planning to guidance. It is important also to note that developing these applications raises many challenges and as such this step should be seen as an integral part of our research. It is also through the development of these applications that we can communicate with physicians, and validate our results. SOFA will be used as a backbone for the integration of our research into clinical applications.

Interventional radiology

Over the past twenty years, interventional methods such as angioplasty, stenting, and catheter-based drug delivery have substantially improved the outcomes for patients with vascular disease. Pathologies that used to require a surgical procedure can now be treated in a much less invasive way. As a consequence, interventional radiology procedures represent an increasing part of the interventions currently performed, with more than 6 million patients treated every year in Europe and about 5 millions the United States. However, these techniques require an intricate combination of tactile and visual feedback, and extensive training periods to attain competency. To reinforce the need to reach and maintain proficiency, the FDA recently required that US physicians go through simulation-based training before using newly developed carotid stents. Besides simulation for training, interventional radiology is a perfect target to illustrate the potential of planning and rehearsal of procedures. As an initial step in this direction, Alcove and Magrit were partners in an ARC project (Simple) to develop a planning tool for the treatment of aneurysms using coils. This collaboration still goes on after the end of the ARC, and led to a series of papers in key conferences [5] [28] , [34] , [21] .

Interventional neuro-radiology

We will continue the development of our simulation and planning system for interventional radiology, with two principal clinical partners: Massachusetts General Hospital in Boston and University Hospital in Nancy. We have completed the integration in SOFA of improved versions of algorithms for describing the behavior of catheters, guide-wires, coils, as well as the interactive simulation of fluoroscopic images, the modeling of complex contacts. Our future efforts will focus on the development of an advanced planning system for interventional radiology, in particular for coil embolization. This will require the integration of new methods of reconstruction of vascular anatomy from medical images (in collaboration with the MAGRIT team). We will also add our recent results on blood flow simulation in aneurysms.

Interventional cardiology using radio-frequency ablation

Cardiac arrhythmias (or dysrhythmias) are problems that affect the electrical system of the heart muscle, producing abnormal heart rhythms, and causing the heart to pump less effectively. About 5% of people over 40 years old are affected by this pathology, with a rather high morbidity rate. Radio-frequency ablation is a non-surgical procedure that has been used for about 15 years to treat tachyarrhythmias, i.e. rapid, uncoordinated heartbeats. The procedure is performed by guiding a catheter with an electrode at its tip to the area of heart muscle where there is an accessory pathway. The catheter is guided under fluoroscopic imaging. When the catheter is positioned at the site where cells give off the electrical signals that stimulate the abnormal heart rhythm, a low radio-frequency energy is transmitted to the pathway. This destroys heart muscle cells within a very small area near the tip of the catheter and stops the area from conducting the extra impulses that caused the arrhythmia. In this context, a simulation system would be able to provide added value in two main areas: 1) to train physicians in the early stages of their apprenticeship and 2) to provide quantitative information during the planning phase of a complex procedure, using patient-specific data. Most aspects of this simulation will rely on components developed during our research program but we will also extend our collaboration with the ASCLEPIOS team and the CardioSense3D project on the modeling of the heart the Cadiosense3D project. This involves an important integration task, and it will also validate the reusability aspects of the code developed within SOFA.

Minimally-invasive surgery

Laparoscopic hepatic resection

The liver is one of the major organs in the human body. It is in charge of more than 100 vital functions. Because of its many functions, its pathologies are also varied, numerous and unfortunately often lethal. This is for instance the case of hepatitides which today affect about 300,000 people in France for hepatitis B and 600,000 people for hepatitis C. The most advanced state of evolution of these pathologies is generally cirrhosis followed by cancer, which represents the third cause of cancer related death. In 2005, 14,267 liver cancer cases and 20,497 cirrhosis cases have been diagnosed in France. The surgical solution remains the option offering the best success rate for these pathologies. More than 7,000 surgical interventions have been carried out on the liver in 2005 and partial resection of the liver remains the most common approach. In this context, the ability to train surgeons, and to be able to plan complex procedures using computer-based simulations, would be a formidable help to the current apprenticeship model: “See One, Do One, Teach One”. Right now, only a few commercial systems are available to the medical community, and they are limited to basic skills training. Developing a realistic simulation system that could be used to plan and rehearse procedures would be a very important step in the introduction of new training paradigms in medicine. This is the main objective of the PASSPORT european project in which we are actively contributing at two levels. First, our research results on biomechanical modeling of solid organs and on coupling will be used to propose a realistic model of the deformation of the liver and its vascular network. Second, SOFA has been chosen in this project as the software for integrating all results from the different partners. Both aspects will help validate our models, test SOFA and obtain feedback from the clinicians.

Ophthalmology and cataract surgery

A cataract is an opacity in the natural lens of the eye. It represents an important cause of visual impairment and, if not treated, can lead to blindness. It is actually the leading cause of blindness worldwide, and its development is related to aging, sunlight exposure, smoking, poor nutrition, eye trauma, and certain medications. The best treatment for this pathology remains surgery. Cataract surgery has made important advances over the past twenty years, and in 2005, more than 5 million people in the United States and in Europe underwent cataract surgery. Most cataract surgeries are performed using microscopic size incisions, advanced ultrasonic equipment to fragment cataracts into tiny fragments, and foldable intraocular lenses to minimize the size of the incision. All these advances benefit the patient, but increase training requirements for eye surgeons. At the end of 2007, we started the development of a new training system for cataract surgery. The main objectives of this simulation are to reproduce with great accuracy the three main steps of cataract surgery: 1) capsulorhexis 2) phacoemulsification and 3) implantation of an intraocular lens. We have already started the development of this simulation. The main research effort went in the choice of appropriate deformable models for the lens and lens capsule. An important effort also went into the development of topological changes corresponding to the capsulorhexis and phacoemulsification [20] . The modeling of the intraocular implant and its deployment in the capsule has been published to the major conference in medical simulation [26] .

Neurosurgery and deep brain stimulation

Deep brain stimulation (DBS) is a neurosurgical treatment which stimulates the brain with low electrical signals. The signals reorganize the brain's electrical impulses (similarly as what was presented above for radio-frequency ablation for cardiac problems). This results in major improvements in several pathologies such as Parkinson disease. The principle of the procedure is the following: a thin, insulated wire lead with several electrodes at the tip is surgically implanted into the affected area of the brain. A wire runs under the skin to a battery-operated pulse generator implanted near the collarbone. The generator is programmed to send continuous electrical pulses to the brain. To implant the electrodes, a neurosurgeon uses a stereotactic head frame and magnetic resonance or computed tomography imaging to map the brain and pinpoint the problem area. The main difficulty in this procedure comes from the deformation of the brain (small brain shift when the skull is opened, and local deformation of the brain due to the insertion of the electrode) and the deflection of the electrode itself during and after the procedure. This results in a difference between the planned target and the location of the end effector of the electrode. Our main objective is to use our work on soft tissue deformation, vascularized structures, as well as our recent results on constraint solving between soft tissues and flexible devices [29] . This work will be done in collaboration with the VISAGES team and we will dedicate an important effort in validating our results, analyzing post-operative medical images, and interacting with surgeons. This project has a strong potential as DBS is being increasingly used yet most research groups only consider non deformable planning systems (geometrical planning). Our proposal could make a important difference in the accuracy of the planning as it takes into account the biophysics of the brain.