Extended Abstract: A Framework for Virtual Surgery

Surgical simulation for medical education and preoperative planning has attracted more and more attention in recent years and a number of such virtual environments have been developed and validated. However, almost all of them are focusing on simulating the surgical scene and haptic interaction to provide users with freedom to perform surgery in the virtual environment. We propose that critical surgical procedures and motion path could be guided by an information intensive process model, especially those being trained in such a virtual surgical environment. In this paper, we outline the virtual surgery framework and the design of the software environment for suturing procedure. The preliminary system that incorporates the above functionality with realistic surgical scene and haptic interaction is still on the development stage. Potentially, the information based modeling process of the surgical motion could also help automate the process of the procedural operation of surgical robot. THE CONCEPTUAL FRAMEWORK The general view of our framework architecture is shown in Figure 1, where the whole system is composed of the hardware part and the software part. In the hardware side(refer to Figure 2), besides such tradition input and output device like keyboard, mouse, haptic device and display, we leave room for sensors to record surgical motion as system input or drive micro assembly work cell to validate and measure our virtual environment in precisely simulating the microsurgery process quantitatively. The software architecture is generally composed of the following three modules. Fig. 1. Framework Architecture A. The IDEF-0 Parser In our approach, the emphasis is on the creation of an Information Intensive Process Model using the IDEF-0 modeling methodology. This model will be used to identify critical and non critical categories of information encompassing the core steps in a neurosurgical (diagnosis) and surgery process; these will possibly include the key or driving assumptions, information inputs, skill constraints, the intermediate ’attribute’ outcomes between various steps or stages of the process in reference as well as the crucial performers (which can range from the medical personnel involved in the diagnosis and surgery D/S itself to the medical assisting devices which play a key role in the outcome of various steps in this D/S process). This information model will not only capture the functional relationships among related tasks at various levels of abstraction but will also enable the representation of temporal precedence constraints among sub-tasks. They will provide a valuable insight into the process of micro surgery; our background in engineering and IT and our prior work in such modeling activities will be helpful in (a) Haptic: Phantom Premium (b) Microassembly Work Cell Fig. 2. Hardware Setup developing such a model. A lower level decomposition of this model is shown in Figure 3. Fig. 3. IDEF-0 Model (A2 Level Decomposition) We seek to develop a robust understanding of these surgical processes using information modeling strategies that have proven to be successful in understanding functional relationships within complex processes that range from performing micro devices assembly to designing virtual reality based simulation environments for satellite assembly (among others). B. Surgical Mode Selection After the framework reads and parse the customizable IDEFL-0 model file, the user will be having the option to further customize the system mode, which is composed of training mode, planning mode, and evaluation mode. The default mode is the virtual environment without any constraints. Those three modes in Figure 1 are described as: • Training and Evaluation Mode: students are able to practice surgery following those constraints defined in the information intensive process model (IIPM) by instructors or experienced surgeons. Meanwhile, the system could evaluate students’ performance based on the comparison of their operation with those constraints defined in the model. • Planning Mode: surgeons could first freely explore different potential surgical strategies in the virtual environment, and then decide a couple of final procedural surgical plans and define those procedural constraints in the IIPM for future validation. • Validation Mode: this is where users could validate those surgical operations in the virtual environment by driving the physical validation system. In our system, since we are focusing on microsurgery, a microassembly work cell (in Figure 2(b)) is used as validation purpose. C. The Virtual Reality Engine This is where the graphics visualization, collision detection and haptic rendering happens, and the FEM Analysis module in VR Engine is used to simulate the soft tissue deformation. Meanwhile, the surgical path generator module is used to define haptical constraints and visualize critical path. The red line in Figure 6 in section IV illustrates this idea. CASE STUDY: VIRTUAL SUTURING In this section, we give a overview of the model and development process of our virtual suturing system based on the above framework. During the early modeling stage, we got our first hand suturing information by learning the microsurgery lab manual and direct clinical microsurgery experience one of our author of this paper has over 20 years clinical microsurgery experience, and the real sutured scene in Figure 4 was taken by a micro lens camera after he was done with the suture. Fig. 4. Real Suturing Scene A. Information Intensive Process Model and Constraints After collecting all those real world information, we are able to model the whole process. The IDEF0 diagram in Figure 3 gives a general model of the real surgical Suturing process. For simplicity of this paper, we will not give the further decomposition here. Meanwhile, two critical steps in performing suturing is illustrated in Figure 5, where Figure 5(a) shows the critical path when the needle starts to insert into the vessel the needle has to be as perpendicular as possible to the target surface, and Figure 5(b) shows the force constraints of how to guide the needle go through the vessel wall. B. Preliminary Results From Figure 5, we can notice that the surgical path during needle insertion and going through vessel wall are critical steps of a successful suture, so to speak, they are important evaluation criteria in the training and evaluation mode, potential procedural constraints in the planning mode, as well as (a) Insertion Angle (b) Shear Force Fig. 5. Suturing (Courtesy: Internet) Fig. 6. Linear Constraints precision motion control in the validation mode. Because the framework is still in the early development stage, in order to illustrate the idea how we incorporate constraints for those critical steps, we developed a much simplified scene to show how the linear constraint guide a surgical cut procedure in Figure 6, where the red line represents the linear constraints -when the scalpel is about to cut the blood vessel (represented by a hollow cylinder), the user will be able to feel a force that guide their movement along the line. RELATED WORK Due to the huge volume of recent research papers on virtual reality system for surgical training and preoperative planning, [8],[6] and [12] have given a relatively detailed review of surgical simulation applications and technology. We limit our survey on the state of art of virtual reality system for training and complex microsurgery preoperative planing, and those key modeling and visualization techniques that enabled such virtual reality system. Surgical Training and Planning System:Though technology has been evolved a lot for the past decades, it is still a challenging task to efficiently simulate the complex surgical operation environment. To avoid the situation of starting an over ambitious project that includes everything but eventually get nothing done, most of the virtual surgery system focus on creating a realistic surgical scenario for some specific purpose, e.g. suturing training [8] [7][19] [1][13] or preoperative planning [21][18] [9][16][17]. Soft Tissue Modeling: Simulating the soft tissue deformation under surgical operation is not trivial, even a simple procedure involves great effort, eg. [3][11] [4][10] were specifically devoted to simulate the cutting procedure. [14] gives a detailed survey of the real-time deformable models used for surgery simulation. Among those models mentioned in [14], MassSpring model and FEM model are the two dominant models currently used in the research community. For its simplicity and low computational cost, the heuristic Mass-Spring model has attracted lots of research and was extensively optimized and deployed in the early years, such as in SPRING system developed by Kevin et all [15]. However, due to its lacks of realism, recent research work are shifting to center around the later continuum mechanics based FEM model as in [1] [5][2] for its accuracy and realistic mechanic behavior, though it is computational costly and needs lots of optimization to achieve real time performance.