The physical basics of magnetic-guided capsule endoscopy of the stomach and results of a feasibility study in the porcine stomach

We read with interest the gastric capsule paper by Rey et al. [1]. We would like to report on similar investigations that we undertook back in 1999 with a project group from the University of Erlangen-Nürnberg led by T. Rupprecht and M. Zapke. The investigations defined the technical specifications of a prototype for magnetic-guided endoscopy in children and adults, which were mainly to develop a system that did not require external cables and that allows freely adjustable movement of the capsule in three dimensions. These results have not yet been published.

In the first theoretical attempt in 2000, a computer simulation study was performed to determine the feasibility of these specifications and to define the physical needs of a hardware prototype (Patent with T. Rupprecht). The calculations were performed using Maple V and the physical data of a M2A capsule (dimensions 11 × 23 mm, weight 4 g; Given Imaging, Yoqneam, Israel), which was modified by incorporating a small permanent magnet. For free movement of the capsule to be possible while also providing the necessary rotation control, a mathematical system was described by different variables such as parallel capsule orientation to the magnetic field, force determined by gravity, and the Maxwell equations. Accordingly, the system had 8 degrees of freedom and therefore 8 magnetic gradients had to be generated requiring at least 14 magnetic coils. Of the resulting multiple coil arrangements, one was chosen for the further simulation study and is represented in [Fig. 1].

Fig. 1 Schematic drawing of the simulated coil arrangement to generate all 8 field gradients. If rotation and orientation of the capsule (dBx/dy .. dBz/dy) are ignored, only 6 coils (green, blue, and red on figure) are necessary. This fact leads to two possible prototype models – one with 6 coils, and a more sophisticated second one with 14 coils for full movement control. In the experiments described in this paper, the simpler design using 6 coils was chosen.

In the next step, the resulting coil currents were simulated mathematically by a solution of the Maxwell equations (Maple V) for the prototype system, under the assumption of a capsule mass of 8 g (including a 4-g permanent magnet) and a capsule dimension of 27 × 11 mm (originally an M2A capsule).

The simulation study resulted in the following conclusions: magnetic degrees of freedom; full control (including rotation and orientation) requires all 14 coils; 3 D control and navigation is possible with 6 coils (see [Fig. 1]); for free flight (hovering), a magnetic field of less than 200 mT is necessary, which can be generated by a coil current of between 1 and 10 amps (therefore no superconductive coils can be used); stable free flight requires an exact localization of the capsule and position-dependant current regulation, which is currently not possible.

Following these studies, the prototype according to the schematic representation shown in  [Fig. 1] was built using conventional copper coils (1 – 10 amps) and a pulse width modulation control for all 6 coils. A 3 D computer software interface for the navigation of the capsule was designed and an M2A capsule was modified by incorporating a small (4-g neodymium) permanent magnet. The prototype system had no possibility for exact localization of the capsule, and therefore only visual control of capsule navigation in all three dimensions could be achieved rather than stable free flight.

Using this prototype device, an experiment was performed using 10 porcine stomachs, which were filled with water and cut open on the lesser curvature for inspection. The prototype capsule was introduced and movement was tested in all directions. In all 10 animal models, a complete inspection of the stomach was possible. The capsule reached a maximum speed of ~ 10 cm/second in each direction. Movement control was judged to be simple by three independent investigators. [Fig. 2] and [ Video 1 ] show typical navigation experiments.

Fig. 2 Typical navigation experiment. The porcine stomach was ligated distal to the pylorus and at the esophageal orifice. It was cut open along the lesser curvature for visualization of the capsule. The magnetic gradients were switched to 3 D, and the experiment included navigation through the whole organ. In this image the capsule is located beneath the esophageal opening, which shows an uncommon fat agglomeration in the mucosa (yellow region).

Video of magnetic-controlled capsule navigation. The capsule is directed by 3 field gradients. Orientation of the long axis of the capsule follows the gradient direction, a feature that results in a view selection along the movement axis. In this animal the paraesophageal mucosal fat is much less pronounced than in [Fig. 1]

Based on our conclusions, the system design was further developed. We considered live animal experiments not to be necessary and the next step in the use of such a system was taken by Rey at al. in their recent study [1]. We congratulate the authors for their important study in the development of this technique, which will be the technology of the future in many human endoscopic examinations.

• References

• 1 Rey JF, Ogata H, Hosoe N et al. Feasibility of stomach exploration with a guided capsule endoscope. Endoscopy 2010; 42: 541-545