gms | German Medical Science

The paranasal sinuses are composed of the ethmoid, maxillary, sphenoidal and frontal sinuses. Common surgery of the paranasal sinuses is the functional endoscopic sinus surgery (FESS), which is a minimal invasive procedure that uses endoscopes.

The disadvantage of conventional FESS for the surgeon is the restriction to one hand for manipulation while the other hand is holding the endoscope. We develop an automatic anticipating robot assisted endoscope guidance during FESS. The robot system should enable the surgeon to operate with two instruments while the robot moves the endoscope. The robot control is based on several components as preoperative segmentation of the paranasal sinuses, force control, instrument tracing and trajectory planning.

There are various reasons for determining forces during surgery, ranging from the development of surgical simulations [1], [2] to robot assisted surgery systems [3], [4], [5], [6].

Previous approaches of force measurement differ in the experimental set-up, including the measuring technique, the type of tissue properties determined, in vivo or ex vivo conditions, the use of human [7], animal [8], [9] or even phantom [10] tissues.

We performed continuous force measurements during conventional cadaver dissection and conventional patient FESS. The result will be used for force controlled robot assisted endoscope guidance.

Throughout the entire operation forces were continuously recorded by custom software. One part of the experiment was carried out on five formalin fixed human cadaver heads, the other on 20 patients who have given their consent. Further we received a positive note of the ethics commission of the University of Bonn.

Simultaneously to the force measurement we noted the corresponding surgical steps. For the gaging we used the force/torque sensor Nano 43 SI-36-0.5 (Schunk GmbH, Lauffen/Neckar, Germany). Force/torque sensors are tactile devices, which determine forces acting on the instrument in six degrees of freedom. The sensor measures translation forces in x-, y- and z-direction ( F _x , F _y , F _z ) and three moments (T _x , T _y , T _z ). In order to find out the amount of forces occurring during FESS we analyzed the total force (F _ges ) which can be computed by the following equation, . The sensor is mounted between an endoscope (Karl Storz GmbH, Tuttlingen, Germany) and a video camera (Figure 1 [Fig. 1]) and measures forces and torques applied by the endoscope. The standard endoscope (No.: 28162AA) has the following hardware properties: 30° optics, 4 mm diameter and 18 cm length.

Figure 2 [Fig. 2] shows the experimental setup which consists of two components that should avoid harm to the patient. First there is an isolating transformer (Toennies Medizinische Elektronik, Freiburg, Germany) which realizes a galvanic separation between socket and patient. Secondly a sterile slipcover is used to wrap the cables connected with the sensor and the endoscope (Figure 1 [Fig. 1] and Figure 2 [Fig. 2]). The set-up for the cadaver FESS is similar to the one which is presented in Figure 2 [Fig. 2] except the isolating transformer and the slipcover. We began our force measurement with five cadaver heads. Here it was possible to operate all paranasal sinuses.

The surgical steps of 20 patients were depended on the location of the pathology. Ethmoid sinuses were operated 20 times, maxillary sinuses 18 times, frontal sinuses 11 times and sphenoidal sinuses 4 times.

Figure 3 [Fig. 3] shows the continuous force measurement of an exemplary patient. During the surgery an average force of 2.7 N and a maximum of 13.7 N appeared. The surgery lasted for about 67 minutes. Analyzing all patients we found out that forces are lower than 7 N for 94.9 percent of the whole operating time. In Figure 4 [Fig. 4] the average and mean forces of all patients are presented. There are high peak forces that reached a maximum of 25.2 N in one case. Forces are measured 64 times per second and peak forces as 25.2 N occur only for few measurement points, for example for 1/10 seconds. The mean forces are in the range of 1.8 N to 6.3 N, so normal forces during surgery are much lower than the maximal forces.

Furthermore not only the maximal and mean forces of the whole operating time especially the maximal and mean forces for each surgical step are of great interest.

In Figure 5 [Fig. 5] the mean forces of several surgical steps are presented. In general it takes lower forces to operate on ethmoid sinuses compared to maxillary and frontal sinuses.

Force-based control systems have been described in the context of a functional robotic milling procedure for otoneurosurgery with a force-based speed control [6]. In our design the force control system will not be used for an active robot that performs operation steps self-contained (for a critical review of the literature on autonomous operating robots see [11]). Instead, we will use the sensor based control for an assistance system that supports the surgeon by moving the endoscope during FESS. Relieved from holding the endoscope the surgeon can more easily focus on the pathology. The introduction of our system will contribute to reduce the complication rate and the operation time.

While force data can not be matched directly with other groups due to differences in the experimental set-ups, the conclusions from the results are comparable. Yamauchi et al. [12] used an urethane head dummy to evaluate surgical skills by recording force data and operation time for a FESS training system. They showed that exercise can reduce exerted forces and operation time. Yamauchi et al. [12] used fixed force sensors beneath the platform holding the dummy whereas in our experiments forces on formalin fixed cadaver heads were measured directly at the endoscope. For robot assisted endoscope guidance we expect constant low forces as they are not depending on the surgeons’ experience.

The high peak forces show that it is sometimes necessary to exert high forces in order to reach an optimal position in paranasal sinuses. It is our aim that the robot can guide the endoscope to every position. In further experiments we have to prove whether 7 N is a useful general threshold and whether it is possible that the robot reaches each position without exceeding 7N.

In order to define thresholds for each surgical step we must analyze more patients as it is presented in Figure 5 [Fig. 5]. We are able to define mean forces for each surgical step if there is a definite separation between the surgery of different sinuses. In some cases it is not possible to define these mean values, especially if surgery is complex, if there are many changes between left and right side and if there are many changes in the sequence of operating ethmoid, frontal, maxillary and sphenoidal sinuses.

For local thresholds we know that they must be lower for ethmoid and sphenoidal sinuses. They must be higher, e.g. 7 N, for operating maxillary and frontal sinuses. The reason is based on the location and the access of the different sinuses.

These local thresholds must also be adapted to earlier experiments [13], [14]. We determined forces that were adequate to damage structures as the lamina papyracea. It will also be necessary to evaluate the thresholds by robot experiments in order to analyze whether the thresholds are useful.

One problem of our measurements is caused by the sensor functionality. The sensor gages the sum of all forces which exerts the endoscope to nasal and paranasal tissue. So far we did not differentiate between forces within the sinuses and those which occur at the entrance of the nasal cavity. In our further work we are going to analyze the measured moments in order to localize the forces.

As a next step tracking the instrument and endoscope position will be realized during FESS. We will measure forces as we did in our current experiments. That will enable us to get more information about the endoscope position at times of high forces. We can also get information about the relative position of the instruments and the endoscope which can be used for trajectory planning.

Our data provide an experimental base to determine local force thresholds for robot assisted endoscope guidance. There are typical changes of force data depending on the surgical step. It takes higher forces to perform endoscopy of the maxillary sinus and the frontal sinus than the ethmoid sinuses. Consequently we need local force thresholds for the robot assisted endoscope guidance.

Forces exerted by the robot to nasal and paranasal tissue must not exceed 7 N at any time during FESS. We will divide nasal cavity and paranasal sinuses into subregions and we will assign one local force threshold to each of the subregions.

The measured force data are adequate to damage sensitive structure in paranasal sinuses as we found out in earlier experiments [13], [14]. The robot assisted endoscope guidance needs more than one security feature to preserve the patient’s integrity. In future further control of robotic endoscope movement will be obtained by developing a multisensor system.

Conflicts of interest: none declared.