Introduction

Microscopic images of laser/tissue interaction, coagulation produced with a conventional CO2 laser
© dkfz.de

VIS-A-VIS STELA is a adaptation of our basic research platform VIS-A-VIS to a new kind of minimally invasive therapy in neurosurgery. In this new therapeutical approach a ultra-short pulsed laser beam should be used to remove gently deep-seated tumour tissue in direct neighbourhood of critical structures.

Laser Neurosurgery

Microscopic images of laser/tissue interaction, coagulation produced with a conventional CO2 laser
© dkfz.de

The use of conventional laser beams in neurosurgery has some important limitations. The laser coagulates tissue but does not ablate it. Due to the diffusion of heat into surrounding healthy tissue severe side-effects were observed, and therefore the application areas are very restricted. In addition, the coagulated tissue remains in situ and can cause dangerous necrosis.

Together with MRC Systems Heidelberg and the Clinic for Stereotactic and Functional Neurosurgery of the University in Cologne we develop a new therapeutical approach to remove deep-seated tumours with minimal irritation of the surrounding healthy tissue, by using an ultra short-pulsed laser (ND:YLF). The interaction mode of ultra short laser pulses with tissue is called electromechanical. Because mechanical effects dominate this interaction mechanism and because ultra short pulse duration does not allow conduction of heat to the surrounding tissue, there is no thermal injury of sensible structures close to the target volume [1]. In-vitro studies with calf brain have shown the possibility to ablate tissue with an extremely high precision without histopathological damages to adjacent structures [2].

Microscopic images of laser/tissue interaction, ablation with ultra short-pulsed laser. Bottom images: electron microscope.
© dkfz.de

Because of the high energy densities of the laser, glass fibres are unsuitable for transmission of laser pulses into the operation field. Therefore a rigid stereotactic probe with a diameter of 5.5 mm was constructed (Fig. 2), where the laser is transmitted by small deflecting mirrors. The probe is a three tube system, where the single tubes can be moved and rotated by step motors [5]. By moving the tube up and down and rotating it, the target volume can be ablated in cylinders from inside to outside.

By means of a pressure-controlled irrigation/aspiration system the ablated tissue fragments (debris) will be removed out of the operation cavity. The irrigation/aspiration system should control and preserve the inter-cranial pressure and the anatomical relations in the operation area. Nevertheless, a monitoring system is required, to observe the operation field and to detect possible tissue displacements.

Microscopic images of laser/tissue interaction, ablation with ultra short-pulsed laser. Bottom images: electron microscope.
© dkfz.de

We prefer ablation of tumour tissue instead of coagulation. Yet, if there are vessels in the operation field, coagulation is still needed. Otherwise the ablation laser might cause dangerous bleedings. Therefore an additional coagulation laser is integrated inside of the probe. An important task of the planning system is the pre- and intra-operative optimisation of the laser control, to switch between the different laser modes.

Treatment Planning

The Laser Probe
© dkfz.de

The ablation or resection strategy must be planned an optimised preoperatively for various reasons:

Since the possible ablation volume is restricted, it might be possible, that a position and orientation of the probe must be defined, where first healthy tissue must be ablated, before the laser focus can reach tumour tissue. To treat the patient with care as much as possible, the amount of resected healthy tissue must be minimised. On the other side, removing healthy tissue will increase operation time substantially, since the laser works in very fine steps with a high precision.

An integrated rinsing system shall remove ablated tissue fragments and shall preserve the intra-cranial pressure and this way the geometric relations between different anatomical structures.

Since the laser does not coagulate tissue it can cause bleedings, if the wall of a vessel is injured. Therefore an additional coagulation mechanism must be added, and a strategy must be developed, to ablate after previous coagulation in regions where no risk of thermal side effects exists (e.g. in the centre of the tumour) and to use only the ablation laser near the tumour border and in regions with high risk of damaging vital structures.

The planning and optimisation of a laser neurosurgical treatment consists of several steps:

  • Multi-modal definition of therapy relevant structures. This step is similar to other kinds of image guided therapy planning. For more information on implemented strategies please refer to the VIS-A-VIS and VIRTUOS pages.
  • Definition of trajectories, defined by target point and trepanation. Since the resection range is restricted, it might be necessary to define several trajectories.
    The resection might cause brain shifts. In this case, the precalculated ablation strategy will become invalid and must be repeated intra-operatively. To detect tissue shifts, the operation region must be monitored continuously with intra-operative image modalities.
    Intra-operative imaging can be done for example with MR or by using 3D ultrasound systems. Both modalities imply restrictions with regards to possible trajectories. In an interventional MR scanner the space for placing the laser probe is limited. If an ultrasound device is applied through an additional trepanation the position of the trepanation must be planned preoperatively in the same way as other approaches, to guarantee, that the operation region and important landmarks are enclosed by the image field.
  • Calculation of resection strategy, with special consideration of regions where coagulation must be applied respectively where thermal side effects must be avoided categorically.
  • Verification of the calculated strategy, based on the acquired image sequences.

VIS-A-VIS - Planning of laser surgery: Due to the limit range of the it is necessary to define three approaches from the same orientation with different target points probe to cover the complete target volume (indicated by three different coloured cylinders around the target volume). A fourth approach is defined for placing the US probe in a way that the complete operation area could be monitored. Besides standard orthogonal slices through the cube there are three additional sections defined by the orientation of the trajectories. Coloured rectangles in the Observer’s View indicate their position. Trajectories are displayed in 2D as orthogonal projections, small circles indicate the cross sections with the displayed slice.
© dkfz.de

These four steps must be repeated iteratively, until a sufficient treatment strategy is developed.

The first planning step is similar to conventional planning of stereotactic interventions with some exceptions. As already mentioned, several approaches must be defined which have close relations. Not only the tissue directly on the access path must be considered but a large deep-sited resection area below other structures.

The last two steps are unique to laser neurosurgery. Special modules for the calculation and visualisation of the resection strategy are needed. VIS-A-VIS supports especially these steps, but it can be used for conventional neurosurgical therapy planning, too.

Calculation of the resection strategy

© dkfz.de

o calculate the resection strategy, a precise three-dimensional model of the patient’s anatomy is needed. Therefore all relevant anatomical structures must be delineated before planning and optimisation can start. The process which will do the calculation relies on the definition of a triangulated surface model. Based on this information, it will subdivide the target volume in small resection primitives. The resection of each primitive can be executed separately. The result of this calculations is stored in a series of three-dimensional matrices, the resection cubes. Each voxel of that cubes represent some kind of information related to the corresponding primitives, e.g. order of resection, coagulate before resection etc.).

Since this information is represented in the same way as dose distributions in radiotherapy planning, the idea was to suit methods already available for these new task and actually due to the large number of the resection step, it makes sense to examine the order of the single resection primitives based on a colour-coded map in the same way as radiotherapeutical dose distributions.

There are several exceptions, the most important one is, that there are several resection cubes for one plan, in opposite to conventional radiotherapy planning, where only one dose distribution exists, which describes the final treatment strategy.

Monitoring of the Intervention

VIS-A-VIS – Intra-operative Monitoring In Sagittal View 2 a section through an intra-operative 3D ultra-sound data set is displayed. In Sagittal View 3 a section through an preoperatively acquired data set is displayed, in the lower image a section through an intra-operatively acquired data set is displayed. Registration was done by applying a mix of stereotactic and interactive methods. The other three windows show three orthogonal sections through the cubes in one of the available fusion techniques. MRI information is displayed with a red colour scale US by using a green colour scale
© dkfz.de

The exact planning and optimisation of the ablation strategy is an important prerequisite to speed up the intervention. But the pre-calculation of the laser control assumes static spatial relations between the anatomic structures. Even if the irrigation/aspiration system could preserve the intra-cranial pressure, it might be possible, that the ablation process will result in a spatial displacement of tumour and brain tissue. Therefore it is necessary to monitor the intervention in real-time.

Since US is a cheap intra-operative imaging modality and widely used, we have first investigated US for monitoring the intervention. Depending on the histology intra-cranial tumours can be delineated quite well in US images. The transducer will be applied through an additional bore-hole. In most cases ablation slices will not have the same orientation as the image plane of a 2D US transducer. Therefore we have examined several 3D US systems [4]. 3D sequences have the additional advantage, to allow monitoring of the whole ablation volume, this is an important prerequisite to detect tissue displacements promptly.

Figure 4 demonstrates the possibility to register 3D US Images properly to pre- and intra-operatively acquired MR sequences. The displayed images were acquired in a conventional neurosurgical operation at the Clinic for Neurosurgery of the University of Heidelberg. The registration was done post-operatively by a mix of stereotactic and manual registration methods. That means, we have determined the orientation of the probe relatively to the patient by using the available navigation system. Based on this information it was easy to align the cubes manually. A comprehensive universal method must still be developed.

Acknowledgements

The presented work was founded by Deutsche Krebshilfe project no. 70-2049-Schl 2 from 1997 – 1999.

The microscopic images and the figure with the laser probe were kindly supplied by MRC Systems Heidelberg.

The pre- and intra-operative MR and US images were kindly supplied by Dr. Bonsanto and Dr. Staubert, University Clinic for Neurosurgery Heidelberg

Future Plans

The work on the STELA mode of VIS-A-VIS is still in progress. The current and future activities will be focused on the implementation of fast and reliable monitoring facilities including the development of elastic registration methods to be able to track and consider possible tissue movements correctly.

The laser probe was tested successfully in several animal trials. The clinical application of the laser system depends on the completion of a new operation theatre at the site of our clinical partner in Cologne.

References

[1] Isner JM and Clarke RH (1987) The paradox of thermal ablation without thermal injury. Lasers Med Sci 2: 165 - 173

[2] Suhm N, Goetz MH, Fischer JP, Loesel F, Schlegel W, Sturm V, Bille J, Schroeder R (1996) In vitro ablation of nervous tissue with short pulsed laser: Histopathological examination of the lesions. Acta Neurochir 138: 346-349

[3] Bendl R, Dams J, Suhm N, Lorenz A, Bille JF, Schlegel W (1997) A Planning System for Stereotactic Laser-Neurosurgery. In: Lemke HU, Inamura K, Jaffe CC, Felix R (eds.) Computer Assisted Radiology. Proceedings of the International Symposium CAR 97, Elsevier Science: 778 - 783

[4] Suhm N, Dams J, van Leyen K, Lorenz A, Bendl R (1998) Limitations for Three-Dimensional Ultrasound Imaging Through a Bore-Hole Trepanation. Ultrasound in Medicine and Biology 24 (5): 663-671

[5] Goetz MH et al. (1999) Computer-guided laser probe for ablation of brain tumours with ultra-short laser pulses. Phys. Med. Biol. 44: N119-N127

[6] Bendl R, Dams J, Fischer S, Götz M (2000) Planning and Optimisation of Stereotactic Laser-Neurosurgery. In: Schlegel W and Bortfeld T (eds) Proceedings of the XIIth International Conference on the Use of Computers in Radiation Therapy, Heidelberg: 350 - 352

to top