Real-time Error Control for Surgical Simulation

Real-time Error Control for Surgical Simulation

Real-time Error Control for Surgical Simulation

Real-time simulations are becoming increasingly common for various applications, from geometric design to medical simulation.

Two of the main factors concurrently involved in defining the accuracy of surgical simulations are: the modeling error and the discretization error. Most work in the area has been looking at the above sources of error as a compounded, lumped, overall error. Little or no work has been done to discriminate between modeling error (e.g. needle-tissue interaction, choice of constitutive models) and discretization error (use of approximation methods like FEM). However, it is impossible to validate the complete surgical simulation approach and, more importantly, to understand the sources of error, without evaluating both the discretization error and the modeling error.

Our objective is thus to devise a robust and fast approach to measure the discretization error via a posteriori error estimates, which are then used for local remeshing in surgical simulations. To ensure that the approach can be used in clinical practice, the method should be robust enough to deal, as realistically as possible, with the interaction of surgical tools with the organ, and fast enough for real-time simulations. The approach should also lead to an improved convergence so that an economical mesh is obtained at each time step. The final goal is to achieve optimal convergence and the most economical mesh, which will be studied in our future work.

The work was submitted to IEEE Transaction on Biomedical Engineering. This is the joint project between Legato team (Huu Phuoc Bui, Satyendra Tomar, Stéphane Bordas) and Stéphane Cotin (Mimesis Inria team, Strasbourg), and Hadrien Courtecuisse (ICube, Strasbourg).

Interested readers can refer to or for more details.

This work is partially supported by University of Strasbourg Institute for Advanced Study, the European project RASimAs, and the  European Research Council Starting Independent Research Grant RealTCut (Towards real time multiscale simulation of cutting in non-linear materials with applications to surgical simulation and computer guided surgery).




Useful links:

Computational Sciences at UL

Computational Engineering at UL

Computational Sciences NEWS

Chargés de mission

Uncertainty quantification for soft tissue biomechanics

Uncertainty quantification for soft tissue biomechanics

Uncertainty quantification for soft tissue biomechanics


– Assessing the effects of uncertainty in material parameters in soft tissue models.
– The sensitivity derivative Monte Carlo method provides one to two orders of magnitude better convergence than the standard Monte Carlo method.
– Complex models with only few lines of Python code (DOLFIN/FEniCS).


– Stochastic FE analysis.
– Uncertainty quantification (material properties, loading, geometry, etc.).
– Random variables/fields.


Two realisations (log-normal distribution)

– Global and local sensitivity analysis.
– Biomechanical modeling, simulation and analysis with random parameters.

fig_brain   ci


– Monte Carlo and quasi Monte Carlo methods (Caflisch, 1998).
– Accelerating Monte Carlo estimation with sensitivity derivatives (Hauseux, Hale, and Bordas, 2016).
– Non-intrusive multi-level polynomial chaos expansion method.
– Multi Level Monte Carlo methods (Giles, 2015).


– UFL (Unified Form Language) (Logg, Mardal, and Wells, 2012).
– Automatically deriving tangent linear models with FEniCS !
– Parallel computing (Ipyparallel and mpi4py).
– Python package for uncertainty quantification (Chaospy, SALib).


Accelerating Monte Carlo estimation with derivatives of high-level finite element models: Hauseux P., Hale J. and Bordas S.
Image to analysis pipeline: single and double balloons kyphoplasty: Baroli D., Hauseux P., Hale J. and Bordas S.
Bayesian statistical inference on the material parameters of a hyperelastic body: Hale J., Farrel P. and Bordas, S.

Computational Sciences at UL

Computational Engineering at UL

Computational Sciences NEWS

Chargés de mission

Real-time error controlled adaptive mesh refinement in surgical simulation: Application to needle insertion

Real-time error controlled adaptive mesh refinement in surgical simulation: Application to needle insertion


submitted to the International Journal for Numerical Methods in Bioengineering. This is collaborative work with here at Legato (on ERC RealTCut), @phuoc who starts with us in a few weeks and was funded by my Strasbourg Institute of Advanced Studies Fellowship and the team of Stéphane Cotin (Inria MEMESIS) and Hadrien Courtecuisse (former post-doc now at ICube in Strasbourg). Congratulations to everyone. This paper shows is the result of a long-lasting collaboration between Mathematics, Computer Science and Engineering and shows that error estimators can be useful also in real-time simulations through an example in liver surgery.


Stéphane Bordas

Today’s Legato Themes: Wang Tiles and Computational Biomechanics

The planning for today’s group meeting:

1. At 0930, we will have a one hour discussion with Australia to put together the working plan for the visit of Grand Joldes and the PhD students he co-supervises with Karol Miller. They will visit us early June for 10 days.

2. At 1030, Jan Novák from Prague will present, along with his PhD students, jointly with Jan Zeman the most innovative and fascinating approach they have devised to generate arbitrary microstructures based on Wang Tilings.

— Stéphane Bordas


Mechanics of (sterile) needle insertion into Human skin

Mechanics of (sterile) needle insertion into Human skin

Several researchers have studied the force required to penetrate solid and some of constitutive models have been developed for the penetration of a soft solid but it seems non of them dealt with sterile needles insertion into human skin.


Understanding the total complexity of bio-mechanical/mechanical properties of the human skin and developing an advanced computational model (e.g. the Finite element skin models) that not differ (or not so much differ!) from experimental data would provide information which could be very useful for surgical training and practical use (special in this project, the goal is to inform the development of optimized device which can be used for effective and reproducible skin penetration in the clinical setting. This project will also provide and make it possible of generating a robust computational and physical model and an excellent technique for measuring skin deformation and in combination with advanced computational/mechanical methods it will also offer many possibilities for in vivo measurements).

For evaluation of simulation of needle insertion into human skin the development of the multilayer cutting will provide the required underlying basis.

Based in scientific and industrial progress,the multilayer cutting of soft solid is scattered over a considerable extent of interests and usefulness.As an application example is the remote robotic surgery or using created computer simulation program for surgical training.

The complex physics and mechanics of cutting process which requires advance knowledge of the fracture mechanics , deformation and friction additionally,can be modeled(3D Mixed-Mode Cracks Model) and imitated using different numerical computation methods such as the extended finite element method (XFEM)[a numerical technique based on the generalized finite element method (GFEM) and the partition of unity method (PUM)] or cohesive elements method.In this study,the method used for simulation of cutting process is based on cohesive elements method.

Literature Review

In considering the complexity of non-linear mechanical behaviour of human skin and the Advanced Measurement Approaches, S Evans* and C A Holt School of Engineering, Cardiff University, UK (2008-2009)[mechanicalpropertiesofhumanskin][3], after a series of Experimental Measurements on human skin and related computational modelling which was the combination of digital image correlation and advanced Finite element modelling, found evidence to suggest,- due to reduction of the errors-,the applying stochastic optimization algorithms ,because output analysis of stochastic optimized algorithm will produce better result than Simplex algorithm and will enable the method to escape a local optimum and eventually to approach a global optimum.

In other studies by R.B. Grovesa, S.A. Coulmanb, J.C. Birchallb and S.L. Evans School of Engineering, Cardiff University, UK (2011)[Groves2012][Groves2012, hyperelasticmodelforskin][4,7], in order to optimised microneedle device designs, -which is completely depends on understanding of human skin biomechanics under small deformations-, after doing a series of optimized laboratory developed tests and using much more precise model(considering the skin as a multilayer composite)with applying multilayer finite element model( with the results of which show a remarkable degree of success) it could find out that because of not strictly accurate or precision between experimental and FEM measurements the problem with the precise approach! is still exist and it could be solved if some other materials property like viscoelasticity and anisotropy will be considered, which tends to reinforce the belief that optimum development of numerical-experimental procedure and modelling of very complex mechanical behavior of human skin, would require first the perfect understanding of dependency and independency of parts or elements of skin combined with mechanical description which can be used later for computational modeling.

Naturally all these studies were carried out in laboratory conditions with parallel load(Evans and Holt 2009) and perpendicular load( Grovesa, Coulmanb, Birchallb and Evans 2011) to the human skin surface.

Without being affected by complex nature of soft solid penetration, it is worth to say that the existing literature unfortunately provides not much insight the underlying mechanismus of penetration.Generally they indicate the deep penetration involves deformation and cracks and in most case without taking into account the existance of (sliding)friction.

There are two main studies which aim to help to develop this project.The first one is the study by Oliver A.Shergold and Norman A. Fleck (2004)[JonathanWainwright][9]with  development of the deep penetration of a soft solid by a flat-bottomed and by a sharp-tipped cylindrical punch with using one term Ogden strain energy function and considering the skin as an incompressible hyperelastic,isotropic solid and the second one is the study by Mohsen Mahvash and Vincent Hayward (2001)[vincenthayward][16] by developing the haptic rendering of cutting with a clarifying of the gemotry and mechanism of interaction of tools and sample.


[1] Enzo Berardesca. Bioengineering of the skin : methods and instrumentation. CRC series in dermatology. CRC Press, Boca Raton, 1995. lc95005294 edited by Enzo Berardesca … [et al.]ill ; 25 cm. Includes bibliographical references and index.

[2] Nuttapong Chentanez. Interactive simulation of surgical needle insertion and steering.

[3] Groves. Quantifying the mechanical properties of human skin to optimise future microneedle device design. Comput Methods Biomech Biomed Engin, 15(1):73–82, 2012. Groves, R B Coulman, S A Birchall, J C Evans, S L eng England 2011/07/14 06:00 Comput Methods Biomech Biomed Engin. 2012;15(1):73-82. doi: 10.1080/10255842.2011.596481. Epub 2011 Jul 12.

[4] Groves. An anisotropic, hyperelastic model for skin: experimental measurements, finite element modelling and identification of parameters for human and murine skin. J Mech Behav Biomed Mater, 18:167–80, 2013. Groves, Rachel B Coulman, Sion A Birchall, James C Evans, Sam L eng Netherlands 2013/01/01 06:00 J Mech Behav Biomed Mater. 2013 Feb;18:167-80. doi: 10.1016/j.jmbbm.2012.10.021. Epub 2012 Nov 19.

[5] A. N. Guz, V. M. Nazarenko, and V. L. Bogdanov. Combined analysis of fracture under stresses acting along cracks. Archive of Applied Mechanics, 83(9):1273–1293, 2013.

[6] F.M. Hendriks. Mechanical behaviour of human skin in vivo.

[7] Holt and Evans. Measuring the mechanical properties of human skin in vivo using digital image correlation and finite element modelling. The Journal of Strain Analysis for Engineering Design, 44(5):337–345, 2009.

[8] Richard D. Wood Javier Bonet. Nonlinear continuum mechanics for finite element analysis.

[9] UK) 2367 2001 Feb 22 07:16:13 Jonathan Wainwright (T&T. Mechanisms of deep penetration of soft solids, with application to the injection and wounding of skin. 2004.

[10] Wen-mei Hwu Li-Wen Chang. A scalable, numerically stable, high-performance tridiagonal solver for gpus.

[11] Ronald Marks, P. A. Payne, and European Society for Dermatological Research. Bioengineering and the skin : based on the proceedings of the European Society for Dermatological Research symposium, held at the Welsh National School of Medicine, Cardiff, 19-21 July 1979. MTP, Lancaster, 1981. lc81014288 edited by R. Marks, P.A. Payne. ill ; 24 cm. Includes bibliographical references and index.

[12] Robert M. Nerem. Tissue engineering the science, the technology and the industry, 2007. : Robert Nerem. Animated audio-visual presentation with synchronized narration. Title from title frames. Contents: Historical perspective – Biomedical devices and diagnostics industry – Medical implant industry – Approved tissue products – Dermagraft – Tissue engineered skin substitutes – Cell source – Matrix – Immune tolerance – Off-the-shelf availability – Embryonic stem cells – Scaffolds – Bioreactor technology – Integration into the living system – Therapeutic products – Key industry trends – Advances envisioned. Mode of access: World Wide Web. System requirements: Operating System: PC Windows 2000+, Mac OSX+ 3.2. Browser Compatibility: IE6+, Firefox 2+, Opera 9+, Safari 2+ 3.3. Browser settings: enable JavaScript, enable popups from the Henry Stewart Talks site. 3.4. Required Browser Plugins & Viewers: Adobe (Macromedia) Flash Player 7+, Adobe Acrobat Reader 6.0+. Henry Stewart talks. * Cardiff University Internet Electronic Seminars.

[13] R. Radovitzky, A. Seagraves, M. Tupek, and L. Noels. A scalable 3d fracture and fragmentation algorithm based on a hybrid, discontinuous galerkin, cohesive element method. Computer Methods in Applied Mechanics and Engineering, 200(1-4):326–344, 2011.

[14] James R. Rice. Mathematical analysis in the mechanics of fracture.

[15] David Roylance. Introduction to fracture mechanics.

[16] vincent hayward. Haptic rendering of cutting: A fracture mechanics approach. 2001.

[17] M.T. Hayajneh V.P. Astakhov, M.O.M. Osman. Re-evaluation of the basic mechanics of orthogonal metal cutting: velocity diagram, virtual work equation and upper-bound theorem.

[18] H. U. I. Wang and Qing-Hua Qin. A fundamental solution-based finite element model for analyzing multi-layer skin burn injury. Journal of Mechanics in Medicine and Biology, 12(05):1250027, 2012.

[19] T. TYAN* YANG’ and WE1 H. Analysis of orthogonal metal cutting processes.


Hybrid lattice continuum approach to interactive cutting in soft tissues.

Hybrid lattice continuum approach to interactive cutting in soft tissues.



The outcome of cutting, tearing, needle insertion and similar operations which require topological changes, or contact detection, is significantly affected by the microstructure of the material (discontinuities, holes, interfaces) remaining some of the most difficult surgical gestures to simulate. We are interested in the development of a numerical tool capable of the interactive simulation of surgical cutting using a multi-domain lattice-continuum approach. Around the cutting region, a mesoscopic discrete lattice approach suitable for initiation of cuts and subsequent tears is used. The remaining regions can be modeled by a continuum approach or through model reduction approaches based on pre-computations. The algorithms are implemented within the SOFA framework which is  targets  real-time computations, with an emphasis on medical simulation and the work is being performed in collaboration with the group of Dr Hadrien Courtecuisse and Dr Stéphane Cotin in Strasbourg.

The final goal of this project is to simulate in real-time the cutting of heterogeneous of soft-tissues using two-scale model instead of using one macroscopic model as in Courtecuisse, H., Allard, J., Kerfriden, P., Bordas, S. P. a, Cotin, S., & Duriez, C. (2014). Real-time simulation of contact and cutting of heterogeneous soft-tissues. Medical Image Analysis, 18(2), 394–410. doi:10.1016/ [Download].

The work is partially funded by USIAS – University of Strasbourg Institute for Advanced Study. Details can be found here.

The work is being performed by Huu Phuoc Bui within Legato team led by Stéphane P.A. Bordas in direct collaboration with Stéphane Cotin, Hadrien Courtecuisse and Michel de Mathelin in MIMESIS and AVR teams in Strasbourg.

Coupling robotics and medical simulation for automatic procedures – CONECT

Coupling robotics and medical simulation for automatic procedures – CONECT

Coupling robotics and medical simulation for automatic procedures.– CONECT”.

The objective of this project is to develop a robotic system, controlled by simulation, for inserting medical needles in the context of interventional radiology for the treatment of cancer. Regarding this work, the aspect of our research can be split into two parts.The first concerns the improvment of the calculations of the simulation time in order to transmit this information quickly enough to the simulated robotic system. The second concerns the coupling of robotics for addressing  problems such as the simulation accuracy and  the error estimation.

This project is funded by Labex funding to Dr Hadrien Courtecuisse in collaboration with Michel de Mathelin and Stéphane P. A. Bordas


Statistical Parameter Identification

Statistical Parameter Identification

Core Team Members: Jack S. Hale, Team Legato, University of Luxembourg.
Collaborators: Patrick Farrell, Oxford University.

« Supported by the Fonds National de la Recherche, Luxembourg (#6693582) »


In this project we are developing computational methods to understand the uncertainty in the recovered material parameters from limited and noisy displacement observations of a hyperelastic material body. This has important applications in developing patient specific models for surgical simulation when direct observations of material properties cannot be made. The addition of statistical information allows the user understand limitations in experimental methodologies, equipment and poor model selection.

Using any suitable medical imaging method we calculate a displacement field of the body and then perform an adjoint based PDE-constrained optimisation problem to recover the material parameters.

Using Bayes’ theorem, we recover the statistical information about the posterior distribution (the probability that we have the material properties given the limited and noisy observations) by forming a low-rank decomposition of the Hessian matrix of the minimisation functional evaluated at the maximum aposteriori point.

Trailing Eigenvector of Hessian evaluated at the maximum aposteriori point.

Trailing Eigenvector of Hessian evaluated at the maximum aposteriori point. The truth value is a circular inclusion and limited and noisy observations are available only on the boundary of the domain. This eigenvector corresponds to the parameters that are least constrained by the given observation data, e.g. across the interface and inside the inclusion.

Mechanics and Dynamics of Needle Insertion Into Soft Tissue

Mechanics and Dynamics of Needle Insertion Into Soft Tissue



Reliable surgical simulators remain elusive. Computer-based simulators could allow surgeons to practice surgical operations without any danger to the patients, to plan difficult interventions and could also help guide surgeons during the operations themselves. This could lead to major improvements in surgical training, decrease risks and ultimately raise ethical standards in surgery. The central stumbling block in surgical simulation is the need to simulate in real-time the interaction of the surgeon with a model of the body. Any numerical model of cutting is as adequate and realistic as the fundamental (phenomenological) models of the underlying physical phenomena that occur during the biological tissue cutting process. The importance of having relevant phenomenological process models of cutting is a motivation for this research. These models will also be able to account for the tissue constitutive behavior. The majority of existing FEM simply ignore the existence of the cutting forces, assuming that the instrument ‘always’ cuts the tissue, thus moving away from the reality of the true physical situation and making force feedback impossible. The FEM models are likely to be physically incorrect (due to the computational necessity), while the existing energy balance based models can be physically correct, but with a very limited application usage. Thus, creating a feasible synergy between the FEM and the analytical energy methods is another important goal of this research.


The major aim of this research is to create feasible analytical models of dynamics and mechanics of soft tissue cutting process and experimentally verify those models. Another goal is to be able to connect the suggested cutting process models with the corresponding FEM models.


(a) Derivation of the cutting process dynamics model. The 1-D cutting process dynamics model will be based on a conventional spring/mass/damping type approach to model the needle/tissue system by means of the ‘lumped’ parameters, derived using a novel velocity-controlled formulation. The usage of energy methods will allow to simulate virtually any cutting instrument (geometry, stiffness, etc.), and any biological tissue interaction process (law). The dynamics model will be able to capture, for example, the system effects related to the cutting equipment, needle motion controller, varying cutting conditions, possible chatter vibrations that can affect the penetration force and can be difficult to separate from the actual cutting process-related parameter oscillations (e.g., stick-slip friction, etc.). The model will be powered by an additional tissue constitutive model in the form of either a strain rate function derived from the corresponding constitutive equation (e.g., Mooney-Rivlin model or other) or by introduction of an additional dynamics equation that characterize the tissue properties (from the corresponding needle insertion tests).

(b) Objective 2: Derivation of the cutting process mechanics model. We aim to derive a unique mechanics model of the tissue/cutting instrument interaction based on a conventional metal cutting mechanics approach. This will be the first attempt to combine the universal model of the effective rake angle of an infinitesimal cutting edge of a needle tip active cutting edge, derived using a novel derivative based approach from the metal cutting mechanics and the conventional fracture force model (for a single-point cutting tool), applied to describe the needle tip cutting edge in this research. The novelty of the suggested analytical model lies in the fact that this model is more comprehensive, that is, it can be applied to any needle tip cutting edge geometry (e.g., bevel, conical, cylindrical, helicoidal, etc.) due to the usage of the derivative based approach.

(c) Verification of the suggested tissue cutting process models. An experimental verification of both proposed models will be carried out by means of a series of needle insertion tests into an artificial (phantom) tissue (silicon gel or candle gel) .

We would like to underline that the experimentation tissues will not be human nor animal ones.



(a) Derive the process model that simulates only penetration force without non-linear friction, and plastic deformation effects considered. The model will take advantage of classic spring/mass/damper type modelling approach formulated in a velocity-controlled form. The model will also contain a tissue constitutive model in the form of an additional strain rate function in the cutting process dynamics equation. Perform testing of the model, using a commercially available software, e.g., Matlab/Simulink, under various cutting conditions. Upgrade the cutting process model with linear semi-experimental stiffness, linear damping, and friction models, based on experimental tissue cutting data available.

(b) Derive the corresponding fracture force model for an infinitesimal cutting edge and for the whole active cutting edge of a needle. The cutting edge is represented by a single point cutting tool characterized by an effective rake angle. Afterwards, derive a corresponding fracture force model for the whole active cutting edge, being a distribution of the infinitesimal cutting edges. A tissue fracture toughness will be obtained from the corresponding needle insertion experiments on an artificial tissue of constant thickness and properties.

(c) A typical experimental test case will consist of a needle being inserted inside and retracted from a phantom tissue at various insertion speeds (1…10 mm/s). During this insertion, a feedback force from the needle will be measured by means of the force sensor. The measured data will be recorded using the LabVIEW. During these experiments, different needle tip geometries and various artificial tissue types will be tested under various cutting conditions.



(a)  The deliverable of the 1-D dynamics model will be a cutting force ‘signature’ model of the cutting process in question (i.e., needle insertion).

(b) The deliverable of the mechanics model will be an analytical fracture force expression.

(c) The deliverable will be a series of needle insertion verification tests providing us with experimental cutting force ‘signatures’ to verify the suggested mechanics and dynamics models .


Future work:

The future work will aim to answer the following questions:

1) Does the 1-D model, resulting in 1-DOF process description, adequately describe the tissue cutting or must extra DOFs be added? What are the plausible process effects to be considered significant in the process model: e.g., friction, damping, chatter, etc.?

2) How well does the derivative based metal cutting mechanics approach describe the needle edge geometry? Does the fracture force mostly depend on the needle tip’s cutting edge geometry?

3) How to create a feasible synergy between the analytical and FEM soft tissue cutting process models?

Eventually, we will integrate the obtained force ‘signature’ with the available FEM models of the tissue cutting processes.

 Academic collaborators:

INRIA (France), Northwestern University (USA)


Brain shift modeling

Brain shift modeling

Deep Brain Stimulation


3D model representing a unilateral DBS system on a patient, with the three components (neurostimulator, extension and electrode)

DBS is the electrical stimulation of a specific area located deep into the brain tissue, but also refers to the surgery resulting to the stimulation. Two surgeries are necessary to implant the complete DBS system, which consists of three components: the stimulating electrode, a neurostimulator and an extension connecting the electrodes and the neurostimulator.

The electrical stimulation of a specific part of the brain can treat different diseases such as movement (Parkinson’s disease for example) and affective disorders. 94% of the DBS treating PD targeted the subthalamic nucleus (STN), 3% the GPi and 3% the VIM.

Operating protocol

  1. Medical imaging without frame
  2. Stereotactic frame placement
  3. Medical imaging with head frame
  4. Pre-surgical target planning
  5. Pre-surgical trajectory planning
  6. Prepare patient for OR, including draping
  7. Stereotactic arc fixation (to locate burr hole)
  8. Incision & making burr hole
  9. Attachment of lead fixation  base on burr hole
  10. Stereotactic arc fixation (after burr hole)
  11. Physiological confirmation of anatomical target
  12. Intra-operative clinical testing with MER system
  13. DBS lead placement
  14. Intra-operative clinical testing with DBS lead

The success of the operation relies on the electrode placement precision, which the goal is to maximize the therapeutic outcomes, and minimize the adverse effects. To do that, a pre-operative planning step determines the target coordinates to stimulate, as well as the electrode trajectory to reach it, thanks to a combination of medical images of the patient and numerical tools

Brain shift

The term brain shift describes movement and deformation of the brain in terms of its anatomical and physiological position in the skull. Brain shift can be observed after a head injury or during a neurosurgery.

During the surgery, a burr hole is drilled in the patient’s skull in order to access the brain tissue at the entry point the surgeons defined earlier. The electrode is then inserted linearly in the direction of the target with the help of an accurate robot. When the skull and dura mater are open, cerebro-spinal fluid (CSF) can leak through the hole. This fluid surrounds the brain and support its weight. A leak of CSF may cause a change of intracranial pressure, leading to a brain deformation (brain shift). This phenomenon is important as brain deformation lead to a displacement of some brain structures, in particular the structures considered during the planning (target or obstacle structures). It results in a difference between the preoperative configuration, based on which the trajectory is selected, and the intraoperative configuration. Although the target motion can be neglected because it is located in deep tissue where the magnitude of deformation is small, blood vessels can shift up to 10 mm. If a blood vessel shifts across the path of the electrode, it could lead to hemorrhage and death of the patient.


Provide a safer and a more effective surgery

Summary of past achievements

The following results are part of a PhD thesis carried out in the Shacra team from Inria, funded by the French Research Agency (ANR) through project ACouStiC (ANR 2010 BLAN 020901).

Brain-shift risk during pre-operative planning


The color on the skin surface indicates the distance to the blood vessels. Dark red is for dangerous trajectories, while blue is for safest trajectories.

The planning utilizes pre-operative data of the patient, mostly a MRI. This planning is based on the configuration of the brain and other structures at the moment the MRI was acquired. If the brain deforms between the pre-operative MRI and the moment to implant an electrode, the planning is obsolete and should be based on the deformed configuration. Brain shift could be a problem for the efficiency of the treatment (target displacement) and regarding to the safety of the patient (blood vessels displacement) with risks of hemorrhage. Therefore, it is necessary to take this phenomenon into account in order to ensure a safe surgery to the patient. The aim of our contributions is to provide pre-operative tools based on a biomechanical model of the brain to prevent risks due to brain shift.

Physics-based intra-operative registration

It is currently impossible to identify small brain structures in intra-operative imaging systems due to a poor contrast in the images. In particular, an intra-operative CT scan only shows an homogeneous voxel intensity for the whole brain tissue. Yet, the location ofsome structures on interest would help the surgeon by adapting the procedure in case of displacement compared to the pre-operative planning. For that, one might consider using an indirect method such as an image registration method. Our objective is to propose a method relying on a biomechanical model. In our opinion, this is the best direction to explore as brain shift is a physical phenomenon. To update the pre-operative brain configuration to the intra-operative configuration, we propose to use our physics-based deformation model. The goal is to estimate the parameters of the model leading to the intra-operative configuration via an inverse problem.

Post-operative electrode curvature

A post-operative electrode displacement and deformation may appear as the brain returns to its initial position when the subdural air introduced during surgery has resolved. This hinders the efficiency of the procedure because upward migration of the electrode may fail to correctly stimulate the subthalamic area.

We propose to add the insertion of the electrode in the brain tissue in our framework. It involves the deformation of the electrode(s) as well as its interaction with the brain tissue. The simulation of the complete procedure is then possible. After the simulation of the surgery, CSF is resolved so that brain can recover its original configuration. The interaction with the electrode(s) will lead to a curvature. Due to the fixation of the electrode on the skull, we can also observe a displacement of the tip of the electrode compared to the planned coordinates.