Lasers Med Sci 2000, 15:6–14 © 2000 Springer-Verlag London Limited
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control M. Klingenberg1, C. Bohris2, M.H. Niemz1, J.F. Bille1, R. Kurek3 and D. Wallwiener3 1 Department of Applied Physics, University of Heidelberg, Heidelberg; 2Department of Radiology, German Cancer Research Center, Heidelberg; 3Department of Obstetrics and Gynecology, University of Tu¨bingen, Tu¨bingen, Germany
Abstract. The application of multiple fibres for the conformal irradiation of tumours by laser-induced interstitial thermotherapy (LITT) has been investigated. A study was performed to evaluate the coagulated zones produced in porcine muscle tissue in vitro. For delivering specified powers into the tissue, a multifibre system was developed which allows the simultaneous use of up to four fibres. A new quantitative method of magnetic resonance imaging (MRI) has been applied for real-time thermometry. It is based on the temperature dependence of the T1 relaxation time and the equilibrium magnetisation. The MR results were compared with the measurements of fibreoptic thermometers. Since the acquisition time of the selected MR sequence takes only 3 s per slice and the calculation of the temperature measurement could be realised within a few seconds, the temperature mapping works closely to real time. The accuracy of the temperature measurements in muscle tissue was 1.5C. Whereas single-fibre applications induced convex-shaped isotherms, concave structures were generated by a multifibre LITT. Keywords: Laser-induced interstitial thermotherapy; Multifibre application; Quantitative MRI-thermometry
INTRODUCTION The invention of the laser has o#ered a variety of new techniques for surgical procedures. The laser has become a well-accepted tool for endoscopic procedures, but also new cancer therapies are currently under investigation. The main modalities for its medical application are photocoagulation, non-thermal ablation (‘optical knife’) and photodynamic therapy. The concept of photocoagulation is based on the conversion of electromagnetic energy into heat, whereas non-thermal ablation and photodynamic therapy rely on completely di#erent laser-tissue interactions [1,2]. The ability to locally coagulate biological tissue has led to a new technique in the treatment of solid circumscribed tissue areas, called laser-induced interstitial thermotherapy (LITT) [3]. It was first introduced by Bown in 1983 [4]. LITT is currently under investigation for some benign diseases, e.g. benign prostate hyperplasia [5,6], twin–twin transfusion syndrome [7], uterine leiomyomas [8], but special Correspondence to: M. Klingenberg, Department of Applied Physics, University of Heidelberg, Albert-U } berleStr. 3-5, 69120 Heidelberg, Germany.
interest focuses on the treatment of solid tumours, liver metastasis [9,10], tumours of the brain [11,12] and tumours of the pelvis [13]. The principle of LITT is to position an appropriate laser applicator inside the tumour and to achieve irreversible tissue damage by heating it above 60C. Either Nd-YAG lasers at 1064 nm or di#erent types of diode lasers at 800–980 nm are applied, since light at these wavelengths penetrates deeply into tissue [14,15]. First applications of LITT using bare fibres induced immediate carbonisation because of the high power density at the end surface of the fibre [16]. Since the early 1990s, special di#using applicators have been developed which scatter laser light isotropically into all spatial directions by means of an enlarged and mechanically or chemically frosted surface at the distal end of the fibre [17]. Thus, large volumes can be treated with temperature gradients not as steep as those associated with conventional thermotherapy based on heat conduction only. The laser applicator usually consists of a flexible fibre with a typical diameter of 400 m and a transparent catheter [18] through which the fibre is moved into the tissue with the assistance of qualified monitoring [7,19]. The
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
application in a clinical setting requires capabilities for reliable and real-time controlling of the induced lesions, since the e#ect is not visible. One possibility is the monitoring of induced necrosis, but this is not as safe as mapping the heat development. Namely, temperatures below the 60C isotherm also induce tissue damage, depending on the exposure time. This damage cannot be seen directly during or after the treatment [20,21] as it becomes apparent a few days after treatment. Considering this, temperature monitoring allows a better assessment of the coagulation process. The complex behaviour of tissue damage can be determined by the Arrhenius equation [22], if the temperature development of the complete treated volume is known. A threedimensional and non-invasive qualitative temperature measurement can be realised with magnetic resonance imaging (MRI) [23,24]. With quantitative temperature monitoring [25] it is possible to evaluate the final lesion size in real time, e.g. during the therapy. Typical parameters of the LITT procedure are 5–10 W laser power for a period of several minutes and coagulation volumes with diameters up to 40 mm. Optional cooling of the catheter may help to prevent thermal damage to the fibre tip at even higher power densities [26,27]. The spatial extent of the damage zone primarily depends on laser power, laser exposure, geometry of laser applicator, thermal and optical tissue properties and blood perfusion. Usually, the coagulated zones are characterised by nearly ellipsoidal shapes, caused by the cylindrical applicator geometry. Recently, the use of special ‘pull-back’ techniques [12] or multifibre applications [28,29] has been suggested in order to treat larger volumes. In this paper the possibility of fitting the physically irradiated volume to the clinically planned target volume by multifibre applications is discussed. The controlled creation of an asymmetrical, but well-defined lesion would be a distinct advantage, since tumours are of a complex shape or rather partially surrounded by a critical structure. Therefore, an in vitro study was performed to investigate the coagulated zones produced in porcine tissue. A multifibre system is presented which allows the use of up to four fibres simultaneously or one after another. Furthermore, the suitability of quantitative MRI monitoring is investigated, in particular its application in multifibre LITT irradiations. The technique of temperature mapping relies on the temperature dependence
7
of the relaxation time T1 [30,31]. Excellent temperature resolution of the MR values was obtained, although the images were acquired in a very short time. To validate the MR measurements, the experiments were performed under simultaneous temperature monitoring by conventional thermometers.
MATERIALS AND METHODS Laser and Fibre Equipment All experiments presented were performed using a Nd-YAG laser system (MediLas 4060-N, Dornier Medizintechnik, Germany), which had been internal calibrated before the experiments with a bare fibre. The lesions were induced using MRI-compatible 600 m LITT-fibres (H-6191-I, Dornier Medizintechnik, Germany) which show a typical ring mode emission profile [32]. These fibres are additionally supplied with a glass dome fixed at the distal end. The purpose of this 20 mm long and 2 mm thick glass cylinder is to decrease the local power density by increasing the active surface and to preserve the fibre from injury. The power emitted by these applicators was measured by placing them inside an integrating sphere power meter (MYTEST, Hu¨ ttinger Medizintechnik, Germany) in order to determine the total output power. The external power meter was calibrated with the internal one in the experiments. Thus, the built-in power display of the laser was used to control the total energy delivered into the tissue.
Multifibre System For separating the laser beam into four rays we used plate beamsplitters with a transmittance/reflectance ratio of nearly 50/50 at a wavelength of 1064 nm. Input and output fibres were coupled by means of planoconvex lenses (Melles Griot, Irvine, USA). Rapid replacement of the fibres was enabled by simply screwing them to the positioners without adjusting those again. The input 600 m fibre (E-6210-D, Dornier Medizintechnik, Germany) was modified at the distal end with a special connector. This was important when removing the fibre from the system during the experiments in order to calibrate the laser. The fibre tip had to be cooled with compressed air to prevent charring.
8
Fig. 1. Schematic overview of the multifibre system. L: Lens; F: Filter; BS: Beamsplitter.
The system shown in Fig. 1 includes filters of various optical densities for two reasons. First, the output powers of all four channels had to be adjusted due to di#erent transmission and reflection losses by the optical components and the dissimilar T/R-ratio of the beamsplitters. This was necessary as uniform delivered energies are important to achieve reproducible results. Furthermore, the filters were used to reduce the power of selected channels.
MRI Measurements The spatial temperature distribution of the tissue was measured during and after LITT treatment by MRI. Experiments were performed on conventional 1.5 Tesla whole body MR-systems (Magnetom SP4000 and Vision, Siemens AG, Germany) using the standard loop coil. The applied MRI method of real-time temperature mapping has been described in detail elsewhere [33]. Briefly, we used a temperature-sensitive T1-weighted SaturationRecovery-TurboFLASH (SRTF)-sequence (repetition time TRL=10.2 ms; echo time TEL= 4 ms; recovery time TREC =1100 ms; acquisition matrix RML=128128; image matrix MAL= 256256; acquisition time TAL=2–3 s/slice; field-of-view FOVL=200200 mm2; slice thickness THL=5–7 mm). The temperature dependencies of the T1 relaxation time and the equilibrium magnetisation lead additively to a reduction of the SRTF signal with increasing temperature. Using a reference image acquired before laser application these signal changes can be measured and related to temperature increases. To quantify the temperature maps, the tissue dependent T1-temperature relationship has to be known. Therefore, the absolute
M. Klingenberg et al.
T1-value at the initial temperature of each muscle tissue sample was measured as described by Bohris [33]. The temperature variation of T1 was studied in former calibration measurements which showed that T1 of muscle tissue increases linearly with 14.51.5 ms/C in the range 20–40C. Although it is expected that T1 undergoes additional non-linear changes due to tissue destruction starting about 45C, it is assumed that the linear behaviour is dominant. Since the acquisition time of the employed MR sequence takes only 3 s per slice and the calculation of the temperature measurement could be realised within seconds, temperature mapping works nearly in realtime. In comparison to other MR methods, the advantages of the proposed technique lie in a reasonable robustness against motion artefacts and in the feasibility for temperature monitoring in fatty tissues.
Experimental Procedure In a first step, we investigated the suitability of MRI-thermometry by a number of 10 singlefibre experiments. Fresh porcine muscle tissue (in vitro) was irradiated with a total power of 9 W. Precise placing of LITT applicators and thermometers was enabled by a 10 mm thick plate with several drillings on top of the sample. Positioning was checked by high resolution MR images (spin-echo sequences). To monitor tissue temperature of the whole treated volume, we chose four parallel slices with 5–7 mm thickness perpendicular to the fibre axes. During and after the irradiation SRTF-images were obtained every 20 s. For reference, fibreoptic temperature measurements with three probes were performed which were positioned at di#erent radial distances with respect to the laser applicator. The accuracy of the fibreoptic thermometer (model 755, Luxtron, USA) was better than 0.5C. For comparison, MRI temperatures were evaluated in small ROIs (‘regions of interest’) comprising 8–12 pixels that were chosen next to the probes. The laser and beamsplitter system were placed outside the MR-cabin to avoid MR-image artefacts. In the next step, we made a series of multifibre experiments in porcine muscle tissue. The experimental set-up is shown in Fig. 2. All experiments were monitored by MR. Although we gained confidence in MR temperature mapping by the results presented above, fibreoptic
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
9
with an edge length of 15 mm. The fourth applicator was inserted in the centre of this pattern but deeper than the others. Thus, the fibre tips built the corners of a regular tetrahedron. A total power of 44 W was applied to the fibres for 14 min. We repeated this experiment by delivering 4 W for 18 min to each fibre.
Fig. 2. Experimental set-up of LITT multifibre application. The image stack of MR measurements is perpendicular to the applicator axes. The experiments were simultaneously controlled by fibreoptic probes.
measurements were performed simultaneously for reference. We investigated various fibre arrangements and irradiation parameters. The applicators were always guided parallel when inserting them into the tissue.
Triangle The applicators were placed at the corners of an equilateral triangle with an edge length of 15 mm. The input power was 4 W per applicator, the corresponding irradiation time was 14 min. In this experiment, the fourth fibre can be used to control the delivered power with an external power meter.
Rectangle The applicators were placed at the corners of a square with an edge length of 15 mm. We irradiated the tissue for 9.6 min with a total power 45 W until the thermometer placed right in the middle of the rectangle reached a value of 60C. This was done to ensure complete coagulation inside the applicator configuration pattern. We repeated this experiment twice by increasing the edge length to 25 mm, first with the same input power but an irradiation time of 20 min and finally with a total power of 46.5 W and an irradiation time of 12 min. The long irradiation time in the first case is due to the condition for the probe in the middle to reach the limiting temperature of irreversible tissue damage.
Tetrahedron In the third experiment we tested a more complex configuration. Three applicators were placed at the corners of an equilateral triangle
Polygon By placing the applicators on an irregular polygon with varying distances, we aspired to induce a very asymmetrical lesion. Moreover, di#erent powers were delivered to the applicators.
Trapezoid The applicators were placed at the corners of a trapezoid. The irradiation parameters were 4 W for 8 min at each channel. Due to larger induced lesions, the number of image slices was increased to nine. MR images were recorded every 30 s during LITT, whereas after the irradiation the periods of acquisitions were enlarged due to the slight temperature variations. After the experiments, the tissue slabs were sliced perpendicular to the applicator axes for evaluating the results.
RESULTS Validation of MRI-thermometry As a representative example, Fig. 3 shows the result of an LITT treatment in muscle tissue with an exposure time of 6.4 min. The decrease of signal intensity is due to a rising temperature during the procedure. After the laser had been switched o# a clearing can be observed. Thus, increase and propagation of heat around the applicator is clearly visible as well as cooling of the tissue after treatment, respectively. From this result we obtain the quantitative temperature development by means of the reference images and calibration measurements. The computed temperature maps were superimposed as colour-coded plots. Since calculation works nearly in real time, an on-line monitoring of the LITT treatment was possible. The temperature measured by the fibreoptic probes and the corresponding MR values as a function of time are also included in
10
M. Klingenberg et al.
Fig. 3. SRTF images of porcine muscle tissue during single-fibre application (a) before irradiation, (b) 2 min, (c) 5 min after starting the LITT treatment. The positions of the applicator and the fibreoptic probes are shown in the reference image. The quantitative comparisons of the MR-measurement with the fibreoptic measurement are shown in (d) and (e) for probe no. 2 and no. 3, respectively.
Fig. 3. These fibre probes were placed at 5 and 7.5 mm from the applicator, respectively. As can be seen, the values of the thermometers during the irradiation are approximately 5C higher as those of MR measurements. This is due to a falsification of the fibreoptic values because of the intrinsic light absorption of the probes [32]. Thus, after the laser turns o#, the MR values agree very well with fibreoptic measurements. Since MR-thermometry proved to be reliable from room temperature at 20C up to 60C (see Fig. 3d), the assumption of a linear and reversible T1–temperature relationship was justified for these applications even when the tissue was irreversibly damaged. The statistical error of the MR values was about 1.5C.
Various Fibre Arrangements Various applicator arrangements and treatment parameters were investigated in this study. In all cases MR measurements again proved to be reliable. Figure 4 shows the results of multifibre applications after slicing the treated muscle tissue. The coagulated area is clearly visible as a whitish colour.
Triangle As can be seen from Fig. 4(a), the coagulated zone in the plane perpendicular to the applicators is of triangular shape with rounded corners. Initially, MR-monitoring provided a temperature increase as in the single-fibre experiments. After a while it was observed that the temperature profiles of the individual fibres run into one another, building isotherms of concave shape. Finally an extended homogeneous heating of the tissue was obtained.
Rectangle As expected, the coagulated area nearly forms a rectangle with rounded corners (Fig. 4b). The edge length of this rectangle is about 33 mm for a 15 mm applicator distance. For the greater distance of 25 mm the corresponding values are 40 mm (26 W, 12 min) and 45 mm (20 W, 20 min). Figure 5 shows the temporal evolution of the temperature for the latter experiment. The initial temperature of the tissue was set at 16C. The contour plots show the temperature increments in central applicator plane induced after 5 min, 10 min, 15 min and 20 min, respectively. The location of each fibre is marked as a filled black circle. The
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
Fig. 4.
11
Photographed lesions of multifibre applications. (a) Triangle; (b) rectangle; (c) polygon; (d) trapezoid.
concave structures could be observed. Afterwards a homogeneous heating of the tissue was obtained again. The final produced lesions were of a convex shape. Due to a longer irradiation time in the second experiment, the extension of the final lesion was something larger than in the first case.
Polygon The result is shown in Fig. 4(c). From left to right the applied powers were 4 W, 3 W, 4 W and 5 W, and the treatment time was 11.2 min. The coagulated area shows convex as well as concave structures.
Fig. 5. Contour plot of the induced temperature increase for the rectangular fibre arrangement (a) 5 min, (b) 10 min, (c) 15 min and (d) 20 min after starting the LITT treatment (central applicator plane). The initial temperature of the tissue was set to 16°C.
isotherms are characterised by a rectangular shape with little deviations only. By measuring the coagulated area after slicing these specimen (central applicator plane) and including the results of the temperature maps we estimated the coagulated volume to be 30–40 cm3.
Tetrahedron In both cases the induced isotherms were copying the tetrahedral arrangement of the applicators. Until the laser had been turned o#,
Trapezoid As can be seen from Fig. 4(d), there is a concave dent of the coagulated area. Figure 6 shows the temporal evolution of the temperature for this experiment. According to the corresponding temperature maps in this figure, the dent of the isotherms in this area is not as large as expected from the photographed lesion. The initial temperature of the tissue was set to 20C. The contour plots show the temperature increments induced after 1 min, 3.5 min, 6 min and 8 min, respectively (central applicator plane). The isotherms exhibit a concave shape at the base line of the trapezoid where the distance of two fibres was stretched. The delimitation of the lesions was quite di$cult, since the change of the visible damage to the undamaged state of tissue is marked by a di#use border. Consequently, the lesion
12
Fig. 6. Contour plot of the induced temperature increase for the trapezoidal fibre arrangement (a) 1 min, (b) 3.5 min, (c) 6 min and (d) 8 min after starting the LITT treatment (central applicator plane). The initial temperature of the tissue was set to 20°C.
marks were determined with an accuracy of only 1 mm. Comparing these results of visible inspection with the MR measurements we found a correlation between the greatest enlargement of the 50C isotherm and the corresponding visible lesion margin for all fibre arrangements.
DISCUSSION Our primary results of multifibre application demonstrated the possibility of inducing lesions of rather complex shape. By suitably spacing several fibres, a huge variety with respect to the coagulation patterns is obtained. In contrast, the single-fibre applications only yield an ellipsoidal necrosis. We first investigated symmetrical fibre arrangements like triangle and rectangle because they could be more easily reproduced and validated by treatment simulations. By selecting one fibre configuration, it is possible to generate coagulation patterns of corresponding shape, but with bulged or dented borders depending on the distance of the applicators and the treatment parameters, respectively. Apart from this, the simultaneous application of multiple fibres is significantly more e$cient as a sequential irradiation of the target volume with one fibre due to the constructive superposition of temperature increase.
M. Klingenberg et al.
This study has shown the suitability of the presented MR method for temperature monitoring of LITT. The temperature maps can be computed nearly in real time, since the acquisition time of the SRTF sequence takes only 3 s per slice and the calculation of quantitative values can be performed within seconds. In porcine muscle tissue, the statistical error of MR temperature measurement (ROI evaluation) was 1.5C. The spatial resolution of 1.51.57.0 mm3 was su$cient to describe the laser-induced temperature changes, although a better resolution would be desirable to reduce averaging e#ects in regions with a steep temperature gradient. By comparison with fibreoptic probes, it has been demonstrated, that the MR results were very reliable. Thus, the temporal evolution of the threedimensional temperature distribution of the whole treated volume was provided by stacks of 4–9 temperature maps which were updated every 20–30 s. As an alternative to the temperature measurement employing T1 the temperature dependence of the proton resonance frequency (PRF) is often used [34]. Harth et al. [35] and Stollberger et al. [36] reported very good results of temperature mapping during the treatment of brain tissue by laser. The advantages of the PRF-method lie mainly in the high and, in di#erent tissue types, uniform temperature sensitivity of the PRF which simplifies the calibration very much. However, in tissues containing fat this approach is disturbed or not applicable due to temperature-induced susceptibility changes [37]. Moreover, the method is very sensitive to motion. Our results demonstrate that by the T1-based method temperature distributions of similar good quality with comparable temporal and spatial resolution could be archived. Therefore, we regard our method equivalent to PRF-methods for ex vivo studies as presented here. In clinical applications it might be an alternative, e.g. in the breast with its high portion of adipose tissue or in body regions a#ected by respiratory motion as in the liver. The temperature contour plots shown in Figs 5 and 6 provide an excellent insight into various motivations for multifibre applications. Although the rectangular fibre arrangement enables a rather homogeneous tissue coagulation, the trapezoidal fibre arrangement may result in sparing out specific tissue sections. Hence, symmetrical fibre arrangements like the rectangle will be best suited for those
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
applications where large tissue volumes need to be coagulated homogeneously. However, in all other applications that require specific irradiation planning (e.g. whenever there exist critical structures or tissues that shall not be coagulated), asymmetrical fibre arrangements like the trapezoid are the preferred method of choice. In general, most tumour geometries will be of a rather complex shape that cannot be fitted by symmetrical fibre arrangements. It is thus the task of thorough irradiation planning to find the best-suited fibre arrangement and to calculate appropriate laser parameters. Our study was restricted to the use of four fibres. Fitting the coagulated volume to a given tumour geometry is certainly improved with higher numbers of fibres applied. However, the use of any additional fibre will make the LITT procedure less minimalinvasive. Hence, a suitable compromise between invasivity and performance must be found for each specific application. There has been many approaches by other groups for enlarging the induced necrosis by multifibre LITT [8,12,28,29]. By applying 45 W in 12.3 min Albrecht et al. [28] reported on significantly greater individual lesion volumes using four fibres simultaneously. They determined the superposed necrosis in porcine liver to be 50 cm3. Ivarsson et al. [29] presented a mathematical model for the multifibre LITT and performed an experimental study that indicate the suitability of this application for the treatment of large tumours. However, all these publications focus on the generation of large lesions and not on the fitting to the irregular shape of the tumour volume itself. Only Wyman [38] investigated the optimisation of source location in multifibre treatment by plane-cut fibres. To our knowledge, the present work is the first to investigate the application of multiple fibres for the conformal irradiation of tumours including the employment of di#using light sources and MR-thermometry.
CONCLUSIONS We investigated the applicability of multifibre LITT for fitting the physically irradiated volume to the clinically planned target volume. For a successful realisation within a clinical set-up three-dimensional and non-invasive temperature monitoring is necessary. The MR technique presented in this paper gives the
13
basis for fitting individual tumour geometries: the temperature maps were calculated quantitatively and nearly in real time. Multifibre LITT under quantitative on-line temperature monitoring o#ers promising prospects for clinical LITT applications. Our approach is based on the simultaneous use of four fibres. Compared to a sequential singlefibre application, this is less time consuming and more e$cient with respect to the extension of the necrosis. Yet, the optimal tumour volume fitting has to be supported by a simulation of temperature development in advance. Thus, the development of a treatment planning programme for multiple fibre treatment is necessary to enable individual tumour fitting. We conclude from our experiments, that multifibre LITT enables the treatment of complex-shaped tumour geometries while adjacent critical structures can be preserved.
ACKNOWLEDGEMENTS The authors are especially grateful to the DFGGraduiertenkolleg ‘Tumordiagnostik und-Therapie’ at the German Cancer Research Centre for supporting this work. We are also grateful for the grant by the fortune programme of the Medical Faculty Tu¨ bingen.
REFERENCES 1. Niemz MH. Laser–Tissue Interactions. Fundamentals and Applications. Berlin, Heidelberg, New York: Springer-Verlag, 1996. 2. Niemz MH, Klancnick EG, Bille JF. Plasma-mediated ablation of corneal tissue at 1053 nm using a Nd:YLF oscillator/regenerative amplifier laser. Lasers Surg Med 1991;11:426–31. 3. Roggan A, Mu¨ ller G. Laser-induced interstitial thermotherapy. Bellingham, Washington: SPIE – The International Society for Optical Engineering, 1995. 4. Bown SG. Phototherapy of tumours. World J Surg 1983;7:700–9. 5. Muschter R, Hofstetter A, Hessel S et al. Laser induced thermotherapy of benign prostatic hyperplasia – fundamentals and clinical experiences. Minimal Invasive Medizin 1994;2:51–4. 6. Mu¨ ller-Lisse UG, Heuck AF, Thoma M et al. Predictability of the size of laser-induced lesions in T1-weighted MR images obtained during interstitial laser-induced thermotherapy of benign prostatic hyperplasia. JMRI 1998;8:31–9. 7. Sohn C, Wallwiener D, Kurek R et al. Treatment of the twin-twin transfusion syndrome: initial experience using laser-induced interstitial thermotherapy. Fetal Diagn Ther 1996;11:390–7. 8. Chapman R. New therapeutic technique for treatment of uterine leiomyomas using laser-induced interstitial
14
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
M. Klingenberg et al. thermotherapy (LITT) by a minimal invasive method. Lasers Surg Med 1998;22:171–8. Vogl T, Mack MG, Straub R et al. Percutaneous MRI-guided laser-induced thermotherapy for hepatic metastases for colorectal cancer. Lancet 1997;350:29. Germer CT, Albrecht D, Roggan A et al. Experimental study of laparoscopic laser induced thermotherapy for liver tumours. Br J Surg 1997;84:317–20. Kahn D, Harth T, Kiwit J et al. In vivo MRI thermometry using a phase-sensitive sequence: Preliminary experience during MRI-guided laser-induced interstitial thermotherapy of brain tumors. JMRI 1998;8: 160–4. Vogl TJ, Mu¨ ller PK, Mack MG. MRT-gesteuerte laserinduzierte Thermotherapie (LITT) in der Onkologie. Medizin Bild 1997;2:31–9. Wallwiener D, Kurek R, Pollmann D et al. Palliative therapy of gynecological malignancies by laserinduced interstitial thermotherapy. Lasermedizin 1994;10:44–51. Wyman DR, Schatz SW, Maguire J. Comparison of 810 nm and 1064 nm wavelengths for interstitial laser photocoagulation in rabbit brain. Lasers Surg Med 1997;21:50–8. Germer CT, Roggan A, Ritz JP et al. Optical properties of native and coagulated human liver tissue and liver metastases in the near infrared range. Lasers Surg Med 1998;23:194–203. Sturesson C. Interstitial laser-induced thermotherapy: Influence of carbonization on lesion size. Lasers Surg Med 1998;22:51–7. Schwarzmaier HJ, Kaufmann R, Kahn T, Ulrich F. Applicators for the laser-induced thermotherapy – basic considerations and new developments. In: Mu¨ ller G, Roggan A (eds) Laser-induced Interstitial Thermotherapy. Bellingham, Washington: SPIE – The International Society for Optical Engineering, 1995:249–62. Roggan A, Albrecht D, Berlien HP et al. Development of an application-set for intraoperative and percutaneous Laser-induced interstitial thermotherapy (LITT). Proc SPIE 1995;2327:253–60. Wallwiener D, Kurek R, Sohn C et al. Study of the on-line monitoring by ultrasonography of the spreading of tissue necrosis in heterogenous tissue induced by interstitial thermotherapy. In: Mu¨ ller G, Roggan A (eds) Laser-induced Interstitial Thermotherapy. Bellingham, Washington: SPIE – The International Society for Optical Engineering. 1995:363–74. Anzai Y, Lufkin RB, Hirschowitz S et al. MR-imaginghistopathologic correlation of thermal injuries induced with interstitial Nd:YAG laser irradiation in the chronic model. JMRI 1992;2:671–8. Tracz RA, Wyman DR, Little PB et al. Comparison of magnetic resonance images and the histopathological findings of lesions induced by interstitial laser photocoagulation in the brain. Lasers Surg Med 1993;13: 45–54. Schwarzmaier HJ, Yaroslavsky IV, Yarovslavsky AN et al. Treatment Planning for MRI-guided laserinduced interstitial thermotherapy of brain tumors – the role of blood perfusion. JMRI 1998;8:121–7.
23. Vogl TJ, Weinhold N, Mu¨ ller PK et al. Early clinical experience with pre-operative MR-guided laser-induced thermotherapy (LITT) of liver metastases. Fortschr Ro¨ ntgenstr 1996;164:413–21. 24. Cholewa D, Wacker F, Roggan A et al. Magnetic resonance imaging: Controlled interstitial laser therapy in children with vascular malformations. Lasers Surg Med 1998;23:250–7. 25. Schulze CP, Kahn T, Harth T et al. Correlation of neuropathologic findings and phase-based MRI temperature maps in experimental laser-induced interstitial thermotherapy. JMRI 1998;8:115–20. 26. Sturesson C, Andersson-Engels S. Tissue temperature control using a water-cooled applicator: implications for transurethral laser-induced thermotherapy of benign prostatic hyperplasia. Med Phys 1997;24: 461–70. 27. Orth K, Russ D, Du¨ rr J et al. Thermo-controlled device for inducing deep coagulation in the liver with the Nd:YAG laser. Lasers Surg Med 1997;20:149–56. 28. Albrecht D, Germer CT, Isbert C et al. Interstitial laser coagulation: Evaluation of the e#ect of normal liver blood perfusion and the application mode on lesion size. Lasers Surg Med 1998;23:40–7. 29. Ivarsson K, Olsrud J, Sturesson C et al. Feedback interstitial diode laser (805 nm) thermotherapy system: ex vivo evaluation and mathematical modeling with one and four-fibres. Lasers Surg Med 1998;22:86–96. 30. Stepanow B, Brix G, Bader R, Lorenz WJ. A novel method for fast temperature mapping. Proceedings of the twelfth annual meeting of the Society of Magnetic Resonance in Medicine, 1993:742. 31. Bohris C, Klingenberg M, Schreiber WG et al. Quantitative MR temperature monitoring during laser-induced interstitial thermotherapy (LITT). Proceedings of the sixth meeting of the International Society of Magnetic Resonance in Medicine, 1998:2001. 32. Manns F, Milne PJ, Gonzales-Cirre X et al. In situ temperature measurements with thermocouple probes during laser interstitial thermotherapy (LITT): Quantification and correlation of a measurement artefact. Lasers Surg Med 1998;23:94–103. 33. Bohris C, Schreiber WG, Jenne J et al. Quantitative MR temperature monitoring of high intensity focused ultrasound therapy. MRI 1999;17:603–10. 34. Ishihara Y, Calderon A, Watanabe H et al. A precise and fast temperature mapping using water proton chemical shift. Magn Reson Med 1995;34:814–23. 35. Harth T, Kahn T, Rassek M et al. Determination of laser-induced temperature distributions using echo-shifted turbo-FLASH. Magn Reson Med 1997;38: 238–45. 36. Stollberger R, Ascher PW, Huber D et al. Temperature monitoring of interstitial thermal tissue coagulation using MR phase images. JMRI 1997;8:188–96. 37. De Poorter J. Noninvasive MRI thermometry with the proton resonance frequency method: study of susceptibility e#ects. Magn Reson Med 1995;34:359–67. 38. Wyman DR. Selecting source locations in multifibre interstitial laser photocoagulation. Lasers Surg Med 1993;13:656–63.
Paper received 7 April 1999; accepted after revision 5 July 1999.
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control M. Klingenberg1, C. Bohris2, M.H. Niemz1, J.F. Bille1, R. Kurek3 and D. Wallwiener3 1 Department of Applied Physics, University of Heidelberg, Heidelberg; 2Department of Radiology, German Cancer Research Center, Heidelberg; 3Department of Obstetrics and Gynecology, University of Tu¨bingen, Tu¨bingen, Germany
Abstract. The application of multiple fibres for the conformal irradiation of tumours by laser-induced interstitial thermotherapy (LITT) has been investigated. A study was performed to evaluate the coagulated zones produced in porcine muscle tissue in vitro. For delivering specified powers into the tissue, a multifibre system was developed which allows the simultaneous use of up to four fibres. A new quantitative method of magnetic resonance imaging (MRI) has been applied for real-time thermometry. It is based on the temperature dependence of the T1 relaxation time and the equilibrium magnetisation. The MR results were compared with the measurements of fibreoptic thermometers. Since the acquisition time of the selected MR sequence takes only 3 s per slice and the calculation of the temperature measurement could be realised within a few seconds, the temperature mapping works closely to real time. The accuracy of the temperature measurements in muscle tissue was 1.5C. Whereas single-fibre applications induced convex-shaped isotherms, concave structures were generated by a multifibre LITT. Keywords: Laser-induced interstitial thermotherapy; Multifibre application; Quantitative MRI-thermometry
INTRODUCTION The invention of the laser has o#ered a variety of new techniques for surgical procedures. The laser has become a well-accepted tool for endoscopic procedures, but also new cancer therapies are currently under investigation. The main modalities for its medical application are photocoagulation, non-thermal ablation (‘optical knife’) and photodynamic therapy. The concept of photocoagulation is based on the conversion of electromagnetic energy into heat, whereas non-thermal ablation and photodynamic therapy rely on completely di#erent laser-tissue interactions [1,2]. The ability to locally coagulate biological tissue has led to a new technique in the treatment of solid circumscribed tissue areas, called laser-induced interstitial thermotherapy (LITT) [3]. It was first introduced by Bown in 1983 [4]. LITT is currently under investigation for some benign diseases, e.g. benign prostate hyperplasia [5,6], twin–twin transfusion syndrome [7], uterine leiomyomas [8], but special Correspondence to: M. Klingenberg, Department of Applied Physics, University of Heidelberg, Albert-U } berleStr. 3-5, 69120 Heidelberg, Germany.
interest focuses on the treatment of solid tumours, liver metastasis [9,10], tumours of the brain [11,12] and tumours of the pelvis [13]. The principle of LITT is to position an appropriate laser applicator inside the tumour and to achieve irreversible tissue damage by heating it above 60C. Either Nd-YAG lasers at 1064 nm or di#erent types of diode lasers at 800–980 nm are applied, since light at these wavelengths penetrates deeply into tissue [14,15]. First applications of LITT using bare fibres induced immediate carbonisation because of the high power density at the end surface of the fibre [16]. Since the early 1990s, special di#using applicators have been developed which scatter laser light isotropically into all spatial directions by means of an enlarged and mechanically or chemically frosted surface at the distal end of the fibre [17]. Thus, large volumes can be treated with temperature gradients not as steep as those associated with conventional thermotherapy based on heat conduction only. The laser applicator usually consists of a flexible fibre with a typical diameter of 400 m and a transparent catheter [18] through which the fibre is moved into the tissue with the assistance of qualified monitoring [7,19]. The
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
application in a clinical setting requires capabilities for reliable and real-time controlling of the induced lesions, since the e#ect is not visible. One possibility is the monitoring of induced necrosis, but this is not as safe as mapping the heat development. Namely, temperatures below the 60C isotherm also induce tissue damage, depending on the exposure time. This damage cannot be seen directly during or after the treatment [20,21] as it becomes apparent a few days after treatment. Considering this, temperature monitoring allows a better assessment of the coagulation process. The complex behaviour of tissue damage can be determined by the Arrhenius equation [22], if the temperature development of the complete treated volume is known. A threedimensional and non-invasive qualitative temperature measurement can be realised with magnetic resonance imaging (MRI) [23,24]. With quantitative temperature monitoring [25] it is possible to evaluate the final lesion size in real time, e.g. during the therapy. Typical parameters of the LITT procedure are 5–10 W laser power for a period of several minutes and coagulation volumes with diameters up to 40 mm. Optional cooling of the catheter may help to prevent thermal damage to the fibre tip at even higher power densities [26,27]. The spatial extent of the damage zone primarily depends on laser power, laser exposure, geometry of laser applicator, thermal and optical tissue properties and blood perfusion. Usually, the coagulated zones are characterised by nearly ellipsoidal shapes, caused by the cylindrical applicator geometry. Recently, the use of special ‘pull-back’ techniques [12] or multifibre applications [28,29] has been suggested in order to treat larger volumes. In this paper the possibility of fitting the physically irradiated volume to the clinically planned target volume by multifibre applications is discussed. The controlled creation of an asymmetrical, but well-defined lesion would be a distinct advantage, since tumours are of a complex shape or rather partially surrounded by a critical structure. Therefore, an in vitro study was performed to investigate the coagulated zones produced in porcine tissue. A multifibre system is presented which allows the use of up to four fibres simultaneously or one after another. Furthermore, the suitability of quantitative MRI monitoring is investigated, in particular its application in multifibre LITT irradiations. The technique of temperature mapping relies on the temperature dependence
7
of the relaxation time T1 [30,31]. Excellent temperature resolution of the MR values was obtained, although the images were acquired in a very short time. To validate the MR measurements, the experiments were performed under simultaneous temperature monitoring by conventional thermometers.
MATERIALS AND METHODS Laser and Fibre Equipment All experiments presented were performed using a Nd-YAG laser system (MediLas 4060-N, Dornier Medizintechnik, Germany), which had been internal calibrated before the experiments with a bare fibre. The lesions were induced using MRI-compatible 600 m LITT-fibres (H-6191-I, Dornier Medizintechnik, Germany) which show a typical ring mode emission profile [32]. These fibres are additionally supplied with a glass dome fixed at the distal end. The purpose of this 20 mm long and 2 mm thick glass cylinder is to decrease the local power density by increasing the active surface and to preserve the fibre from injury. The power emitted by these applicators was measured by placing them inside an integrating sphere power meter (MYTEST, Hu¨ ttinger Medizintechnik, Germany) in order to determine the total output power. The external power meter was calibrated with the internal one in the experiments. Thus, the built-in power display of the laser was used to control the total energy delivered into the tissue.
Multifibre System For separating the laser beam into four rays we used plate beamsplitters with a transmittance/reflectance ratio of nearly 50/50 at a wavelength of 1064 nm. Input and output fibres were coupled by means of planoconvex lenses (Melles Griot, Irvine, USA). Rapid replacement of the fibres was enabled by simply screwing them to the positioners without adjusting those again. The input 600 m fibre (E-6210-D, Dornier Medizintechnik, Germany) was modified at the distal end with a special connector. This was important when removing the fibre from the system during the experiments in order to calibrate the laser. The fibre tip had to be cooled with compressed air to prevent charring.
8
Fig. 1. Schematic overview of the multifibre system. L: Lens; F: Filter; BS: Beamsplitter.
The system shown in Fig. 1 includes filters of various optical densities for two reasons. First, the output powers of all four channels had to be adjusted due to di#erent transmission and reflection losses by the optical components and the dissimilar T/R-ratio of the beamsplitters. This was necessary as uniform delivered energies are important to achieve reproducible results. Furthermore, the filters were used to reduce the power of selected channels.
MRI Measurements The spatial temperature distribution of the tissue was measured during and after LITT treatment by MRI. Experiments were performed on conventional 1.5 Tesla whole body MR-systems (Magnetom SP4000 and Vision, Siemens AG, Germany) using the standard loop coil. The applied MRI method of real-time temperature mapping has been described in detail elsewhere [33]. Briefly, we used a temperature-sensitive T1-weighted SaturationRecovery-TurboFLASH (SRTF)-sequence (repetition time TRL=10.2 ms; echo time TEL= 4 ms; recovery time TREC =1100 ms; acquisition matrix RML=128128; image matrix MAL= 256256; acquisition time TAL=2–3 s/slice; field-of-view FOVL=200200 mm2; slice thickness THL=5–7 mm). The temperature dependencies of the T1 relaxation time and the equilibrium magnetisation lead additively to a reduction of the SRTF signal with increasing temperature. Using a reference image acquired before laser application these signal changes can be measured and related to temperature increases. To quantify the temperature maps, the tissue dependent T1-temperature relationship has to be known. Therefore, the absolute
M. Klingenberg et al.
T1-value at the initial temperature of each muscle tissue sample was measured as described by Bohris [33]. The temperature variation of T1 was studied in former calibration measurements which showed that T1 of muscle tissue increases linearly with 14.51.5 ms/C in the range 20–40C. Although it is expected that T1 undergoes additional non-linear changes due to tissue destruction starting about 45C, it is assumed that the linear behaviour is dominant. Since the acquisition time of the employed MR sequence takes only 3 s per slice and the calculation of the temperature measurement could be realised within seconds, temperature mapping works nearly in realtime. In comparison to other MR methods, the advantages of the proposed technique lie in a reasonable robustness against motion artefacts and in the feasibility for temperature monitoring in fatty tissues.
Experimental Procedure In a first step, we investigated the suitability of MRI-thermometry by a number of 10 singlefibre experiments. Fresh porcine muscle tissue (in vitro) was irradiated with a total power of 9 W. Precise placing of LITT applicators and thermometers was enabled by a 10 mm thick plate with several drillings on top of the sample. Positioning was checked by high resolution MR images (spin-echo sequences). To monitor tissue temperature of the whole treated volume, we chose four parallel slices with 5–7 mm thickness perpendicular to the fibre axes. During and after the irradiation SRTF-images were obtained every 20 s. For reference, fibreoptic temperature measurements with three probes were performed which were positioned at di#erent radial distances with respect to the laser applicator. The accuracy of the fibreoptic thermometer (model 755, Luxtron, USA) was better than 0.5C. For comparison, MRI temperatures were evaluated in small ROIs (‘regions of interest’) comprising 8–12 pixels that were chosen next to the probes. The laser and beamsplitter system were placed outside the MR-cabin to avoid MR-image artefacts. In the next step, we made a series of multifibre experiments in porcine muscle tissue. The experimental set-up is shown in Fig. 2. All experiments were monitored by MR. Although we gained confidence in MR temperature mapping by the results presented above, fibreoptic
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
9
with an edge length of 15 mm. The fourth applicator was inserted in the centre of this pattern but deeper than the others. Thus, the fibre tips built the corners of a regular tetrahedron. A total power of 44 W was applied to the fibres for 14 min. We repeated this experiment by delivering 4 W for 18 min to each fibre.
Fig. 2. Experimental set-up of LITT multifibre application. The image stack of MR measurements is perpendicular to the applicator axes. The experiments were simultaneously controlled by fibreoptic probes.
measurements were performed simultaneously for reference. We investigated various fibre arrangements and irradiation parameters. The applicators were always guided parallel when inserting them into the tissue.
Triangle The applicators were placed at the corners of an equilateral triangle with an edge length of 15 mm. The input power was 4 W per applicator, the corresponding irradiation time was 14 min. In this experiment, the fourth fibre can be used to control the delivered power with an external power meter.
Rectangle The applicators were placed at the corners of a square with an edge length of 15 mm. We irradiated the tissue for 9.6 min with a total power 45 W until the thermometer placed right in the middle of the rectangle reached a value of 60C. This was done to ensure complete coagulation inside the applicator configuration pattern. We repeated this experiment twice by increasing the edge length to 25 mm, first with the same input power but an irradiation time of 20 min and finally with a total power of 46.5 W and an irradiation time of 12 min. The long irradiation time in the first case is due to the condition for the probe in the middle to reach the limiting temperature of irreversible tissue damage.
Tetrahedron In the third experiment we tested a more complex configuration. Three applicators were placed at the corners of an equilateral triangle
Polygon By placing the applicators on an irregular polygon with varying distances, we aspired to induce a very asymmetrical lesion. Moreover, di#erent powers were delivered to the applicators.
Trapezoid The applicators were placed at the corners of a trapezoid. The irradiation parameters were 4 W for 8 min at each channel. Due to larger induced lesions, the number of image slices was increased to nine. MR images were recorded every 30 s during LITT, whereas after the irradiation the periods of acquisitions were enlarged due to the slight temperature variations. After the experiments, the tissue slabs were sliced perpendicular to the applicator axes for evaluating the results.
RESULTS Validation of MRI-thermometry As a representative example, Fig. 3 shows the result of an LITT treatment in muscle tissue with an exposure time of 6.4 min. The decrease of signal intensity is due to a rising temperature during the procedure. After the laser had been switched o# a clearing can be observed. Thus, increase and propagation of heat around the applicator is clearly visible as well as cooling of the tissue after treatment, respectively. From this result we obtain the quantitative temperature development by means of the reference images and calibration measurements. The computed temperature maps were superimposed as colour-coded plots. Since calculation works nearly in real time, an on-line monitoring of the LITT treatment was possible. The temperature measured by the fibreoptic probes and the corresponding MR values as a function of time are also included in
10
M. Klingenberg et al.
Fig. 3. SRTF images of porcine muscle tissue during single-fibre application (a) before irradiation, (b) 2 min, (c) 5 min after starting the LITT treatment. The positions of the applicator and the fibreoptic probes are shown in the reference image. The quantitative comparisons of the MR-measurement with the fibreoptic measurement are shown in (d) and (e) for probe no. 2 and no. 3, respectively.
Fig. 3. These fibre probes were placed at 5 and 7.5 mm from the applicator, respectively. As can be seen, the values of the thermometers during the irradiation are approximately 5C higher as those of MR measurements. This is due to a falsification of the fibreoptic values because of the intrinsic light absorption of the probes [32]. Thus, after the laser turns o#, the MR values agree very well with fibreoptic measurements. Since MR-thermometry proved to be reliable from room temperature at 20C up to 60C (see Fig. 3d), the assumption of a linear and reversible T1–temperature relationship was justified for these applications even when the tissue was irreversibly damaged. The statistical error of the MR values was about 1.5C.
Various Fibre Arrangements Various applicator arrangements and treatment parameters were investigated in this study. In all cases MR measurements again proved to be reliable. Figure 4 shows the results of multifibre applications after slicing the treated muscle tissue. The coagulated area is clearly visible as a whitish colour.
Triangle As can be seen from Fig. 4(a), the coagulated zone in the plane perpendicular to the applicators is of triangular shape with rounded corners. Initially, MR-monitoring provided a temperature increase as in the single-fibre experiments. After a while it was observed that the temperature profiles of the individual fibres run into one another, building isotherms of concave shape. Finally an extended homogeneous heating of the tissue was obtained.
Rectangle As expected, the coagulated area nearly forms a rectangle with rounded corners (Fig. 4b). The edge length of this rectangle is about 33 mm for a 15 mm applicator distance. For the greater distance of 25 mm the corresponding values are 40 mm (26 W, 12 min) and 45 mm (20 W, 20 min). Figure 5 shows the temporal evolution of the temperature for the latter experiment. The initial temperature of the tissue was set at 16C. The contour plots show the temperature increments in central applicator plane induced after 5 min, 10 min, 15 min and 20 min, respectively. The location of each fibre is marked as a filled black circle. The
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
Fig. 4.
11
Photographed lesions of multifibre applications. (a) Triangle; (b) rectangle; (c) polygon; (d) trapezoid.
concave structures could be observed. Afterwards a homogeneous heating of the tissue was obtained again. The final produced lesions were of a convex shape. Due to a longer irradiation time in the second experiment, the extension of the final lesion was something larger than in the first case.
Polygon The result is shown in Fig. 4(c). From left to right the applied powers were 4 W, 3 W, 4 W and 5 W, and the treatment time was 11.2 min. The coagulated area shows convex as well as concave structures.
Fig. 5. Contour plot of the induced temperature increase for the rectangular fibre arrangement (a) 5 min, (b) 10 min, (c) 15 min and (d) 20 min after starting the LITT treatment (central applicator plane). The initial temperature of the tissue was set to 16°C.
isotherms are characterised by a rectangular shape with little deviations only. By measuring the coagulated area after slicing these specimen (central applicator plane) and including the results of the temperature maps we estimated the coagulated volume to be 30–40 cm3.
Tetrahedron In both cases the induced isotherms were copying the tetrahedral arrangement of the applicators. Until the laser had been turned o#,
Trapezoid As can be seen from Fig. 4(d), there is a concave dent of the coagulated area. Figure 6 shows the temporal evolution of the temperature for this experiment. According to the corresponding temperature maps in this figure, the dent of the isotherms in this area is not as large as expected from the photographed lesion. The initial temperature of the tissue was set to 20C. The contour plots show the temperature increments induced after 1 min, 3.5 min, 6 min and 8 min, respectively (central applicator plane). The isotherms exhibit a concave shape at the base line of the trapezoid where the distance of two fibres was stretched. The delimitation of the lesions was quite di$cult, since the change of the visible damage to the undamaged state of tissue is marked by a di#use border. Consequently, the lesion
12
Fig. 6. Contour plot of the induced temperature increase for the trapezoidal fibre arrangement (a) 1 min, (b) 3.5 min, (c) 6 min and (d) 8 min after starting the LITT treatment (central applicator plane). The initial temperature of the tissue was set to 20°C.
marks were determined with an accuracy of only 1 mm. Comparing these results of visible inspection with the MR measurements we found a correlation between the greatest enlargement of the 50C isotherm and the corresponding visible lesion margin for all fibre arrangements.
DISCUSSION Our primary results of multifibre application demonstrated the possibility of inducing lesions of rather complex shape. By suitably spacing several fibres, a huge variety with respect to the coagulation patterns is obtained. In contrast, the single-fibre applications only yield an ellipsoidal necrosis. We first investigated symmetrical fibre arrangements like triangle and rectangle because they could be more easily reproduced and validated by treatment simulations. By selecting one fibre configuration, it is possible to generate coagulation patterns of corresponding shape, but with bulged or dented borders depending on the distance of the applicators and the treatment parameters, respectively. Apart from this, the simultaneous application of multiple fibres is significantly more e$cient as a sequential irradiation of the target volume with one fibre due to the constructive superposition of temperature increase.
M. Klingenberg et al.
This study has shown the suitability of the presented MR method for temperature monitoring of LITT. The temperature maps can be computed nearly in real time, since the acquisition time of the SRTF sequence takes only 3 s per slice and the calculation of quantitative values can be performed within seconds. In porcine muscle tissue, the statistical error of MR temperature measurement (ROI evaluation) was 1.5C. The spatial resolution of 1.51.57.0 mm3 was su$cient to describe the laser-induced temperature changes, although a better resolution would be desirable to reduce averaging e#ects in regions with a steep temperature gradient. By comparison with fibreoptic probes, it has been demonstrated, that the MR results were very reliable. Thus, the temporal evolution of the threedimensional temperature distribution of the whole treated volume was provided by stacks of 4–9 temperature maps which were updated every 20–30 s. As an alternative to the temperature measurement employing T1 the temperature dependence of the proton resonance frequency (PRF) is often used [34]. Harth et al. [35] and Stollberger et al. [36] reported very good results of temperature mapping during the treatment of brain tissue by laser. The advantages of the PRF-method lie mainly in the high and, in di#erent tissue types, uniform temperature sensitivity of the PRF which simplifies the calibration very much. However, in tissues containing fat this approach is disturbed or not applicable due to temperature-induced susceptibility changes [37]. Moreover, the method is very sensitive to motion. Our results demonstrate that by the T1-based method temperature distributions of similar good quality with comparable temporal and spatial resolution could be archived. Therefore, we regard our method equivalent to PRF-methods for ex vivo studies as presented here. In clinical applications it might be an alternative, e.g. in the breast with its high portion of adipose tissue or in body regions a#ected by respiratory motion as in the liver. The temperature contour plots shown in Figs 5 and 6 provide an excellent insight into various motivations for multifibre applications. Although the rectangular fibre arrangement enables a rather homogeneous tissue coagulation, the trapezoidal fibre arrangement may result in sparing out specific tissue sections. Hence, symmetrical fibre arrangements like the rectangle will be best suited for those
Multifibre Application in Laser-Induced Interstitial Thermotherapy under On-Line MR Control
applications where large tissue volumes need to be coagulated homogeneously. However, in all other applications that require specific irradiation planning (e.g. whenever there exist critical structures or tissues that shall not be coagulated), asymmetrical fibre arrangements like the trapezoid are the preferred method of choice. In general, most tumour geometries will be of a rather complex shape that cannot be fitted by symmetrical fibre arrangements. It is thus the task of thorough irradiation planning to find the best-suited fibre arrangement and to calculate appropriate laser parameters. Our study was restricted to the use of four fibres. Fitting the coagulated volume to a given tumour geometry is certainly improved with higher numbers of fibres applied. However, the use of any additional fibre will make the LITT procedure less minimalinvasive. Hence, a suitable compromise between invasivity and performance must be found for each specific application. There has been many approaches by other groups for enlarging the induced necrosis by multifibre LITT [8,12,28,29]. By applying 45 W in 12.3 min Albrecht et al. [28] reported on significantly greater individual lesion volumes using four fibres simultaneously. They determined the superposed necrosis in porcine liver to be 50 cm3. Ivarsson et al. [29] presented a mathematical model for the multifibre LITT and performed an experimental study that indicate the suitability of this application for the treatment of large tumours. However, all these publications focus on the generation of large lesions and not on the fitting to the irregular shape of the tumour volume itself. Only Wyman [38] investigated the optimisation of source location in multifibre treatment by plane-cut fibres. To our knowledge, the present work is the first to investigate the application of multiple fibres for the conformal irradiation of tumours including the employment of di#using light sources and MR-thermometry.
CONCLUSIONS We investigated the applicability of multifibre LITT for fitting the physically irradiated volume to the clinically planned target volume. For a successful realisation within a clinical set-up three-dimensional and non-invasive temperature monitoring is necessary. The MR technique presented in this paper gives the
13
basis for fitting individual tumour geometries: the temperature maps were calculated quantitatively and nearly in real time. Multifibre LITT under quantitative on-line temperature monitoring o#ers promising prospects for clinical LITT applications. Our approach is based on the simultaneous use of four fibres. Compared to a sequential singlefibre application, this is less time consuming and more e$cient with respect to the extension of the necrosis. Yet, the optimal tumour volume fitting has to be supported by a simulation of temperature development in advance. Thus, the development of a treatment planning programme for multiple fibre treatment is necessary to enable individual tumour fitting. We conclude from our experiments, that multifibre LITT enables the treatment of complex-shaped tumour geometries while adjacent critical structures can be preserved.
ACKNOWLEDGEMENTS The authors are especially grateful to the DFGGraduiertenkolleg ‘Tumordiagnostik und-Therapie’ at the German Cancer Research Centre for supporting this work. We are also grateful for the grant by the fortune programme of the Medical Faculty Tu¨ bingen.
REFERENCES 1. Niemz MH. Laser–Tissue Interactions. Fundamentals and Applications. Berlin, Heidelberg, New York: Springer-Verlag, 1996. 2. Niemz MH, Klancnick EG, Bille JF. Plasma-mediated ablation of corneal tissue at 1053 nm using a Nd:YLF oscillator/regenerative amplifier laser. Lasers Surg Med 1991;11:426–31. 3. Roggan A, Mu¨ ller G. Laser-induced interstitial thermotherapy. Bellingham, Washington: SPIE – The International Society for Optical Engineering, 1995. 4. Bown SG. Phototherapy of tumours. World J Surg 1983;7:700–9. 5. Muschter R, Hofstetter A, Hessel S et al. Laser induced thermotherapy of benign prostatic hyperplasia – fundamentals and clinical experiences. Minimal Invasive Medizin 1994;2:51–4. 6. Mu¨ ller-Lisse UG, Heuck AF, Thoma M et al. Predictability of the size of laser-induced lesions in T1-weighted MR images obtained during interstitial laser-induced thermotherapy of benign prostatic hyperplasia. JMRI 1998;8:31–9. 7. Sohn C, Wallwiener D, Kurek R et al. Treatment of the twin-twin transfusion syndrome: initial experience using laser-induced interstitial thermotherapy. Fetal Diagn Ther 1996;11:390–7. 8. Chapman R. New therapeutic technique for treatment of uterine leiomyomas using laser-induced interstitial
14
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
M. Klingenberg et al. thermotherapy (LITT) by a minimal invasive method. Lasers Surg Med 1998;22:171–8. Vogl T, Mack MG, Straub R et al. Percutaneous MRI-guided laser-induced thermotherapy for hepatic metastases for colorectal cancer. Lancet 1997;350:29. Germer CT, Albrecht D, Roggan A et al. Experimental study of laparoscopic laser induced thermotherapy for liver tumours. Br J Surg 1997;84:317–20. Kahn D, Harth T, Kiwit J et al. In vivo MRI thermometry using a phase-sensitive sequence: Preliminary experience during MRI-guided laser-induced interstitial thermotherapy of brain tumors. JMRI 1998;8: 160–4. Vogl TJ, Mu¨ ller PK, Mack MG. MRT-gesteuerte laserinduzierte Thermotherapie (LITT) in der Onkologie. Medizin Bild 1997;2:31–9. Wallwiener D, Kurek R, Pollmann D et al. Palliative therapy of gynecological malignancies by laserinduced interstitial thermotherapy. Lasermedizin 1994;10:44–51. Wyman DR, Schatz SW, Maguire J. Comparison of 810 nm and 1064 nm wavelengths for interstitial laser photocoagulation in rabbit brain. Lasers Surg Med 1997;21:50–8. Germer CT, Roggan A, Ritz JP et al. Optical properties of native and coagulated human liver tissue and liver metastases in the near infrared range. Lasers Surg Med 1998;23:194–203. Sturesson C. Interstitial laser-induced thermotherapy: Influence of carbonization on lesion size. Lasers Surg Med 1998;22:51–7. Schwarzmaier HJ, Kaufmann R, Kahn T, Ulrich F. Applicators for the laser-induced thermotherapy – basic considerations and new developments. In: Mu¨ ller G, Roggan A (eds) Laser-induced Interstitial Thermotherapy. Bellingham, Washington: SPIE – The International Society for Optical Engineering, 1995:249–62. Roggan A, Albrecht D, Berlien HP et al. Development of an application-set for intraoperative and percutaneous Laser-induced interstitial thermotherapy (LITT). Proc SPIE 1995;2327:253–60. Wallwiener D, Kurek R, Sohn C et al. Study of the on-line monitoring by ultrasonography of the spreading of tissue necrosis in heterogenous tissue induced by interstitial thermotherapy. In: Mu¨ ller G, Roggan A (eds) Laser-induced Interstitial Thermotherapy. Bellingham, Washington: SPIE – The International Society for Optical Engineering. 1995:363–74. Anzai Y, Lufkin RB, Hirschowitz S et al. MR-imaginghistopathologic correlation of thermal injuries induced with interstitial Nd:YAG laser irradiation in the chronic model. JMRI 1992;2:671–8. Tracz RA, Wyman DR, Little PB et al. Comparison of magnetic resonance images and the histopathological findings of lesions induced by interstitial laser photocoagulation in the brain. Lasers Surg Med 1993;13: 45–54. Schwarzmaier HJ, Yaroslavsky IV, Yarovslavsky AN et al. Treatment Planning for MRI-guided laserinduced interstitial thermotherapy of brain tumors – the role of blood perfusion. JMRI 1998;8:121–7.
23. Vogl TJ, Weinhold N, Mu¨ ller PK et al. Early clinical experience with pre-operative MR-guided laser-induced thermotherapy (LITT) of liver metastases. Fortschr Ro¨ ntgenstr 1996;164:413–21. 24. Cholewa D, Wacker F, Roggan A et al. Magnetic resonance imaging: Controlled interstitial laser therapy in children with vascular malformations. Lasers Surg Med 1998;23:250–7. 25. Schulze CP, Kahn T, Harth T et al. Correlation of neuropathologic findings and phase-based MRI temperature maps in experimental laser-induced interstitial thermotherapy. JMRI 1998;8:115–20. 26. Sturesson C, Andersson-Engels S. Tissue temperature control using a water-cooled applicator: implications for transurethral laser-induced thermotherapy of benign prostatic hyperplasia. Med Phys 1997;24: 461–70. 27. Orth K, Russ D, Du¨ rr J et al. Thermo-controlled device for inducing deep coagulation in the liver with the Nd:YAG laser. Lasers Surg Med 1997;20:149–56. 28. Albrecht D, Germer CT, Isbert C et al. Interstitial laser coagulation: Evaluation of the e#ect of normal liver blood perfusion and the application mode on lesion size. Lasers Surg Med 1998;23:40–7. 29. Ivarsson K, Olsrud J, Sturesson C et al. Feedback interstitial diode laser (805 nm) thermotherapy system: ex vivo evaluation and mathematical modeling with one and four-fibres. Lasers Surg Med 1998;22:86–96. 30. Stepanow B, Brix G, Bader R, Lorenz WJ. A novel method for fast temperature mapping. Proceedings of the twelfth annual meeting of the Society of Magnetic Resonance in Medicine, 1993:742. 31. Bohris C, Klingenberg M, Schreiber WG et al. Quantitative MR temperature monitoring during laser-induced interstitial thermotherapy (LITT). Proceedings of the sixth meeting of the International Society of Magnetic Resonance in Medicine, 1998:2001. 32. Manns F, Milne PJ, Gonzales-Cirre X et al. In situ temperature measurements with thermocouple probes during laser interstitial thermotherapy (LITT): Quantification and correlation of a measurement artefact. Lasers Surg Med 1998;23:94–103. 33. Bohris C, Schreiber WG, Jenne J et al. Quantitative MR temperature monitoring of high intensity focused ultrasound therapy. MRI 1999;17:603–10. 34. Ishihara Y, Calderon A, Watanabe H et al. A precise and fast temperature mapping using water proton chemical shift. Magn Reson Med 1995;34:814–23. 35. Harth T, Kahn T, Rassek M et al. Determination of laser-induced temperature distributions using echo-shifted turbo-FLASH. Magn Reson Med 1997;38: 238–45. 36. Stollberger R, Ascher PW, Huber D et al. Temperature monitoring of interstitial thermal tissue coagulation using MR phase images. JMRI 1997;8:188–96. 37. De Poorter J. Noninvasive MRI thermometry with the proton resonance frequency method: study of susceptibility e#ects. Magn Reson Med 1995;34:359–67. 38. Wyman DR. Selecting source locations in multifibre interstitial laser photocoagulation. Lasers Surg Med 1993;13:656–63.
Paper received 7 April 1999; accepted after revision 5 July 1999.
