In-Vivo Real-Time Magnetic Resonance Monitoring of Endoscopic Laser Applications in the Porcine Gastrointestinal Tract

• Introduction
• Methods
• The MR Endoscope
• Open-Gantry MR Imaging
• Laser
• Animal Procedures
• Results
• Discussion
• Acknowledgments
• References

Background and Study Aims: Endoscopic laser therapy involves a risk of perforation, mainly because the depth of tissue destruction is not visible. Magnetic resonance (MR) imaging is capable of showing temperature changes, and is therefore suitable for monitoring thermal therapies such as laser. This animal study assessed the feasibility of real-time MR monitoring of endoscopic laser applications in the gastrointestinal tract.

Materials and Methods: The procedures were carried out using an MR-compatible endoscope in three live pigs in a 0.5-Tesla interventional MR system. Nd:YAG laser applications were performed in the lower gastrointestinal tract (n = 7) and upper gastrointestinal tract (n = 5), and were monitored using real-time color-coded T1-weighted gradient echo sequences. The postmortem macroscopic tissue coagulation sizes were compared with the lesion diameters seen on real-time MR.

Results: The endoscope did not cause any artifacts during continuous MR imaging. Ten of the twelve laser lesions were visible with temperature-sensitive MR imaging, and their sizes correlated well with the diameters of the postmortem macroscopic coagulation zones (r = 0.76, P = 0.009). Two laser lesions were not visible on MR due to technical limitations inherent with the healthy animal model.

Conclusions: The formation of endoscopic laser lesions in the porcine gastrointestinal tract can be accurately visualized using real-time temperature-sensitive MR imaging. This new technique has the potential to spare healthy tissue while ensuring full treatment coverage of the targeted lesion with fewer therapy sessions.

Introduction

Laser therapy for stenosing tumors of the gastrointestinal tract has become a well-recognized treatment option [1] [2] . One of the drawbacks of the technique is the risk of iatrogenic perforation, since the depth of laser-induced heat-coagulation cannot be assessed endoscopically [3]. Gradual ablation of the tumor using multiple treatment sessions is therefore carried out. If it were possible to visualize the depth of tissue destruction caused by laser energy during the laser application, more radical treatment could be achieved in fewer treatment sessions.

Magnetic resonance (MR) imaging is particularly suitable for monitoring and controlling thermal therapies such as laser, focused ultrasound, cryotherapy, and radiofrequency ablation [4 - 7]. Changes are seen in the MR images after heating, when tissue coagulates, and on freezing, when water protons become immobile within ice crystals [8]. The new generation of open-gantry MR systems with real-time imaging allows physicians to carry out endoscopy, laparoscopy, percutaneous interventions, and surgery during concurrent MR imaging while standing at the side of the patient [9 - 11]. MR-compatible gastrointestinal endoscopes containing an MR receiver coil in the tipp have already been used in the MR environment for diagnostic indications, including local staging of esophageal, gastric, gallbladder, pancreatic, and rectal tumors [12 - 14].

The present animal study assessed the feasibility of therapeutic endoscopic laser application with real-time monitoring in an interventional MR system.

Methods

The MR Endoscope

The MR-compatible endoscope (Olympus XGIF-MR30C, Olympus, Tokyo, Japan) used in the current study is a fiberscope with an outer diameter of 13.6 mm, a working length of 180 cm, and a 5-mm accessory channel. The endoscope is constructed from special plastics, titanium, aluminum, brass, tungsten, copper, and amalgam. These nonferromagnetic materials do not produce any artificial signals on MR imaging, and are not subject to movement forces when placed within the magnetic field. An external circular balloon 4 cm in length (Olympus MD-694, Olympus, Tokyo, Japan), filled with 8 ml dimeglumine gadopentetate (Magnevist, Schering, Berlin, Germany) diluted 1 : 1000 with normal saline was attached to the tip of the endoscope to make it visible on MR. The endoscopic light was delivered by a halogen light source (Olympus CLK-3, Olympus, Hamburg, Germany), and the endoscopic image was provided by a video camera (Olympus OTV-S4, Olympus, Hamburg, Germany).

Open-Gantry MR Imaging

The procedures were performed in an interventional 0.5 T MR system (Signa SP, General Electric Medical, Schenectady, New York, USA) with a 56-cm gap allowing vertical access to the center of the imaging field (Figure [1]). A flexible rectangular surface coil measuring 24 × 28 cm was used for signal transmission and reception. Both the MR images and the endoscopic view were simultaneously visible to the endoscopist on two liquid crystal monitors within the scanner.

Imaging data sets were obtained using a T1-weighted gradient-echo (GRE) sequence with the following parameters: flip angle 60°; relaxation time 18.9 ms; echo time 9.2 ms; field of view 30 × 30 cm; matrix 256 × 160; section thickness 10 mm; bandwidth 12.5 kHz; and image updating every 3 s in a single axial, sagittal, or coronal plane. Signal intensity changes derived from the GRE sequence during laser application were used to assess temperature alterations. A color-coded subtraction method was applied to further enhance visualization of the temperature changes [10]. A reference magnitude signal intensity map was acquired before treatment within an operator-defined region of interest (Figure [2 b]). These values then were subtracted from continuously updated magnitude maps obtained during laser application. The signal differences demonstrating temperature changes were color-coded in increasing order as blue (no temperature increase), green, yellow, red, orange, and white (hottest area). Based on previous similar experiments assessing the qualitative correlation between color changes and the degree of tissue destruction [7], a lesion was defined as an area containing a red, orange, or white signal. The largest diameter of the lesion on the color-coded MR scan obtained at the end of each laser application was measured on the imaging monitor using a reference scale that was incorporated into the image.

Laser

The neodymium-yttrium aluminum garnet (Nd:YAG) laser source (VersaPulse Select, Coherent, Palo Alto, California, USA) emitting continuous output at a wavelength of 1064 nm, was positioned outside the scanner room to avoid safety hazards and corruption of the MR signal. A 10-m long laser fiber bridged the distance between the generator and a connector box installed within the scanner. A 365-μm bare fiber (SlimLine 356, Coherent, Palo Alto, California, USA) was inserted through the accessory channel of the endoscope. The energy loss due to the connecting box and cable amounted to 10 % [10]. Laser energy was delivered continuously at a power of 5 W.

Animal Procedures

The study was conducted according to a protocol approved by the local animal experimental committee. The procedures were performed on three live domestic pigs under general anesthesia. Arterial blood pressure, electrocardiography, and peripheral oxygen saturation were monitored throughout the procedure using an MR compatible unit (MagLife, Odam Bruker, Wissembourg, France). The animals were placed in a supine or lateral position in the opening within the magnet, perpendicular to the long axis of the scanner (Figure [1]).

The endoscope was inserted into the rectum and colon (two animals) or the upper gastrointestinal tract (one animal) under continuous MR monitoring. Laser applications were carried out in the rectum (n = 5), transverse colon (n = 2), gastric cardia (n = 1), gastric body (n = 1) and antrum (n = 1), pylorus (n = 1), and in the duodenal bulb (n = 1).

The intended laser application sites were marked using a submucosal injection of 1 - 2 ml dimeglumine gadopentetate (Magnevist, Schering, Berlin, Germany), diluted 1 : 1000 with normal saline, through a modified Teflon sclerotherapy catheter with a 22-gauge titanium alloy tipp taken from an MR-compatible biopsy needle (Lufkin, E-Z-EM, Westbury, New York, USA) (Figure [2 a]). This was necessary to allow localization of the appropriate imaging plane for the temperature-sensitive MR scanning, as the animals did not have any spontaneous pretreatment mass lesions.

The laser fiber was then inserted through the accessory channel of the endoscope and positioned with the tip at a distance of 5 mm from the mucosal surface. Laser application was performed until the lesion seen on color-coded temperature-sensitive MR imaging did not demonstrate any further growth.

After completion of the experiments, the animals were sacrificed by injecting potassium chloride, and the treated stomach and colon segments were excised. The largest diameters of the macroscopic coagulation zones were measured and compared with the corresponding lesion sizes obtained using MR imaging.

Pearson correlation analysis was performed for the lesions that were visible on MR, to determine the correlation between the measurements made on gross inspection and on MR images. A P value of < 0.05 was considered significant.

Results

Esophagogastroduodenoscopy and colonoscopy were safely performed using the interventional MR system. The endoscope, any accessories used (including the titanium needle), and the laser fiber did not cause any artifacts on continuous MR imaging. The attachment of a balloon filled with gadolinium diluted 1 : 1000 with normal saline allowed real-time MR localization of the tip of the endoscope in all experiments (Figure [2 a]).

At autopsy, all 12 laser lesions caused macroscopically visible coagulation zones, with the largest diameters ranging between 5 mm and 13 mm (mean 9.4 mm) (Figure [2 d]). Ten of the 12 lesions were successfully monitored using color-coded real-time MR imaging. Two 9-mm rectal lesions in the first animal experiment were not visualized on MR due to initial difficulties with accurate definition of the appropriate imaging plane. Color changes appeared after 20 - 50 s of laser treatment. Temperature spread during laser application was visualized by a growing extension of the color focus. The maximum extension of the lesions was reached after laser delivery for 3 - 6.5 min (mean 4 min), with energy outputs ranging between 900 J and 1950 J (mean 1200 J). Termination of the laser application resulted in a decrease in the color changes within 60 - 120 s.

For the ten lesions that were visible on MR, Pearson correlation analysis of the corresponding lesion diameters derived from MR imaging and from the macroscopic inspection demonstrated a correlation coefficient of r = + 0.76 at P = 0.009 (Figure [3]). The MR lesions and the corresponding macroscopic coagulation zones demonstrated size differences of ≤ 2 mm in eight experiments (7 mm vs. 8 mm in the rectum, 11 mm vs. 13 mm in the rectum, 5 mm vs. 5 mm in the transverse colon, 7 mm vs. 8 mm in the transverse colon, 7 mm vs. 9 mm in the cardia, 10 mm vs. 9 mm in the corpus, 13 mm vs. 12 mm in the antrum, and 10 mm vs. 11 mm in the pylorus, respectively) (Figure [2 c, d]). The lesion sizes were underestimated by more than 2 mm by MR in one rectal laser application (6 mm vs. 10 mm) and in the lesion in the duodenal bulb (5 mm vs. 9 mm).

Discussion

MR imaging makes it possible to provide temperature maps of the affected tissues during thermal therapy - information that has not previously been available using ultrasound or computed tomography. MR-guided thermal therapy has been successfully applied in solid organs, including the brain, liver, muscle, and vertebral disks using percutaneous and laparoscopic approaches [7] [10] [11] [15] . This is the first application of MR monitoring of laser therapy delivered via a flexible gastrointestinal endoscope. We found that the formation of laser lesions in the porcine gastrointestinal tract can be accurately visualized using real-time MR imaging.

The lesion sizes on the temperature-sensitive MR scans correlated well with the diameters of the postmortem macroscopic tissue coagulation areas. Previous experiments on temperature-sensitive imaging during laser and radiofrequency applications in the paraspinal muscles, vertebral disks, and liver [7] [16] [17] yielded even better correlations, for various reasons. In these organs, motion artifacts are less prevalent than during flexible endoscopy in the bowel. Some of these experiments were carried out ex vivo on porcine and human cadavers [16] [17] . In addition, rigid transcutaneous laser and radiofrequency applicators allow the use of a stereotactic imaging guidance system, which automatically adapts the orientation of the MR imaging plane to the needle's orientation and position [17]. Finally, different thermosensitive imaging sequences were used in some of the experiments [7] .

In the present study, temperature sensitivity was based on the T1-relaxation time of a fast T1-weighted GRE sequence. T1 increases as the temperature increases, leading to a reduction in the signal intensity on the MR image. The T1 technique only offers qualitative visualization of laser effects, and does not allow quantification of temperature changes. The results of the present study confirm findings previously published in the literature showing that this qualitative method provides a good measure of the extent of tissue destruction [16] [17] . In addition, the T1 method is a fast technique that is readily available on virtually all MR scanners and with moderate sensitivity to motion artifacts. By contrast, quantitative temperature-sensitive MR methods, such as the proton frequency shift technique, are more complex to implement and require the use of longer image update times. The proton frequency shift method shows much greater sensitivity to motion artifacts, resulting in image distortions that can even be caused simply by the presence of an endoscopist in the scanner room.

The lack of a visible pretreatment lesion inherent with the healthy animal model in the present study made accurate MR localization of the intended treatment sites more complicated. This led to the failed and underestimated MR delineation of some laser lesions. In addition, since the various tissues surrounding the bowel wall usually result in different signal intensities on T1-weighted images, similar temperature changes may result in different signal intensity alterations, suggesting different temperatures. These limitations would be greatly reduced when treating real gastrointestinal mass lesions. Experience with transcutaneous MR-guided laser-induced thermotherapy in patients with liver or brain neoplasms demonstrated accurate destruction of the tumors, followed by encouraging long-term results [4] [15] .

This experimental study provides a technical basis for monitoring endoscopic thermal therapy in the gastrointestinal tract using MR. In this setting, MR imaging provides the endoscopist with visualization of the third dimension, outside the gut wall. The use of endoscopic laser treatment in an interventional MR environment demonstrated the potential to spare healthy tissues while ensuring full treatment coverage of the targeted lesion during fewer therapy sessions.

Acknowledgments

Daniel Külling, M.D., was supported by the 1999 Ludwig Demling Research Prize awarded by the Olympus Europe “Science for Life” Foundation.

The MR-endoscope was provided by Olympus Japan.

The authors do not have any financial or other interest in the manufacture or distribution of any device mentioned in the manuscript.

The authors are grateful to the members of the international MR-Guided Endotherapy Working Group (Simon Bar-Meir, M.D., Tel Aviv, Israel; Robert H. Hawes, M.D., Charleston, South Carolina, USA; and Thomas Rösch, M.D., Munich, Germany) for reviewing the manuscript.

References

• 1 Spinelli P, Fante M D al, Mancini A. Endoscopic palliation of malignancies of the upper gastrointestinal tract using Nd:YAG laser: results and survival in 308 treated patients.  Lasers Surg Med. 1991;  11 550-555
  PubMedSearch in Google Scholar
• 2 Mathus-Vliegen E M, Tytgat G N. Analysis of failures and complications of neodymium:YAG laser photocoagulation in gastrointestinal tract tumors: a retrospective survey of 18 years experience.  Endoscopy. 1990;  22 17-23
  PubMedSearch in Google Scholar
• 3 Schulze S, Fischerman K. Palliation of oesophagogastric neoplasms with Nd:YAG laser treatment.  Scand J Gastroenterol. 1990;  25 1024-1027
  PubMedSearch in Google Scholar
• 4 Vogl T J, Mack M G, Straub R, et al. Magnetic resonance imaging-guided abdominal interventional radiology: laser-induced thermotherapy of liver metastases.  Endoscopy. 1997;  29 577-583
  PubMedSearch in Google Scholar
• 5 Cline H E, Hynynen K, Watkins R D, et al. Focused US system for MR imaging-guided tumor ablation.  Radiology. 1995;  194 731-777
  PubMedSearch in Google Scholar
• 6 Klotz H P, Flury R, Schonenberger A, et al. Experimental cryosurgery of the liver under magnetic resonance guidance.  Comput Aided Surg. 1997;  2 340-345
  CrossrefPubMedSearch in Google Scholar
• 7 Steiner P, Botnar R, Dubno B, et al. Radiofrequency-induced thermoablation: monitoring with T1-weighted and proton-frequency-shift MR imaging in an interventional 0.5-T environment.  Radiology. 1998;  206 803-810
  PubMedSearch in Google Scholar
• 8 Matsumoto R, Oshio K, Jolesz F A. Monitoring of laser and freezing-induced ablation in the liver with T1-weighted imaging.  J Magn Reson Imaging. 1993;  3 770-776
  CrossrefPubMedSearch in Google Scholar
• 9 Lufkin R B. Interventional MRI.  St. Louis, MO; Mosby, 1999
  Search in Google Scholar
• 10 Steiner P, Zweifel K, Botnar R, et al. MR guidance of laser disc decompression: preliminary in vivo experience.  Eur Radiol. 1998;  8 592-597
  CrossrefPubMedSearch in Google Scholar
• 11 Klotz H P, Flury R, Erhart P, et al. Magnetic resonance-guided laparoscopic interstitial laser therapy of the liver.  Am J Surg. 1997;  174 448-451
  CrossrefPubMedSearch in Google Scholar
• 12 Kulling D, Feldman D R, Kay C L, et al. Local staging of esophageal cancer using endoscopic magnetic resonance imaging: prospective comparison with endoscopic ultrasound.  Endoscopy. 1998;  3 745-749
  PubMedSearch in Google Scholar
• 13 Kulling D, Feldman D R, Kay C L, et al. Local staging of anal and distal colorectal tumors with the magnetic resonance endoscope.  Gastrointest Endosc. 1998;  47 172-178
  PubMedSearch in Google Scholar
• 14 Inui K, Nakazawa S, Yoshino J, Ukai H. Endoscopic MRI.  Pankreas. 1998;  16 413-417
  PubMedSearch in Google Scholar
• 15 Anzai Y, Lufkin R, DeSalles A, et al. Preliminary experience with MR-guided thermal ablation of brain tumors.  Am J Neuroradiol. 1995;  16 39-48
  PubMedSearch in Google Scholar
• 16 Steiner P, Botnar R, Goldberg N, et al. Monitoring of radiofrequency tissue ablation in an interventional magnetic resonance environment: preliminary ex vivo and in vivo results.  Invest Radiol. 1997;  32 671-678
  CrossrefPubMedSearch in Google Scholar
• 17 Schoenenberger A W, Steiner P, Debatin J F, et al. Real-time monitoring of laser diskectomies with a superconducting, open-configuration MR system.  AJR Am J Roentgenol. 1997;  169 863-867
  PubMedSearch in Google Scholar

Daniel Külling, M.D.

Department of Internal Medicine University Hospital of Zürich

Rämistrasse 100 8091 Zürich Switzerland

Fax: Fax:+ 41-1-255-4503

Email: E-mail:kulling@gmx.ch

References

• 1 Spinelli P, Fante M D al, Mancini A. Endoscopic palliation of malignancies of the upper gastrointestinal tract using Nd:YAG laser: results and survival in 308 treated patients.  Lasers Surg Med. 1991;  11 550-555
  PubMedSearch in Google Scholar
• 2 Mathus-Vliegen E M, Tytgat G N. Analysis of failures and complications of neodymium:YAG laser photocoagulation in gastrointestinal tract tumors: a retrospective survey of 18 years experience.  Endoscopy. 1990;  22 17-23
  PubMedSearch in Google Scholar
• 3 Schulze S, Fischerman K. Palliation of oesophagogastric neoplasms with Nd:YAG laser treatment.  Scand J Gastroenterol. 1990;  25 1024-1027
  PubMedSearch in Google Scholar
• 4 Vogl T J, Mack M G, Straub R, et al. Magnetic resonance imaging-guided abdominal interventional radiology: laser-induced thermotherapy of liver metastases.  Endoscopy. 1997;  29 577-583
  PubMedSearch in Google Scholar
• 5 Cline H E, Hynynen K, Watkins R D, et al. Focused US system for MR imaging-guided tumor ablation.  Radiology. 1995;  194 731-777
  PubMedSearch in Google Scholar
• 6 Klotz H P, Flury R, Schonenberger A, et al. Experimental cryosurgery of the liver under magnetic resonance guidance.  Comput Aided Surg. 1997;  2 340-345
  CrossrefPubMedSearch in Google Scholar
• 7 Steiner P, Botnar R, Dubno B, et al. Radiofrequency-induced thermoablation: monitoring with T1-weighted and proton-frequency-shift MR imaging in an interventional 0.5-T environment.  Radiology. 1998;  206 803-810
  PubMedSearch in Google Scholar
• 8 Matsumoto R, Oshio K, Jolesz F A. Monitoring of laser and freezing-induced ablation in the liver with T1-weighted imaging.  J Magn Reson Imaging. 1993;  3 770-776
  CrossrefPubMedSearch in Google Scholar
• 9 Lufkin R B. Interventional MRI.  St. Louis, MO; Mosby, 1999
  Search in Google Scholar
• 10 Steiner P, Zweifel K, Botnar R, et al. MR guidance of laser disc decompression: preliminary in vivo experience.  Eur Radiol. 1998;  8 592-597
  CrossrefPubMedSearch in Google Scholar
• 11 Klotz H P, Flury R, Erhart P, et al. Magnetic resonance-guided laparoscopic interstitial laser therapy of the liver.  Am J Surg. 1997;  174 448-451
  CrossrefPubMedSearch in Google Scholar
• 12 Kulling D, Feldman D R, Kay C L, et al. Local staging of esophageal cancer using endoscopic magnetic resonance imaging: prospective comparison with endoscopic ultrasound.  Endoscopy. 1998;  3 745-749
  PubMedSearch in Google Scholar
• 13 Kulling D, Feldman D R, Kay C L, et al. Local staging of anal and distal colorectal tumors with the magnetic resonance endoscope.  Gastrointest Endosc. 1998;  47 172-178
  PubMedSearch in Google Scholar
• 14 Inui K, Nakazawa S, Yoshino J, Ukai H. Endoscopic MRI.  Pankreas. 1998;  16 413-417
  PubMedSearch in Google Scholar
• 15 Anzai Y, Lufkin R, DeSalles A, et al. Preliminary experience with MR-guided thermal ablation of brain tumors.  Am J Neuroradiol. 1995;  16 39-48
  PubMedSearch in Google Scholar
• 16 Steiner P, Botnar R, Goldberg N, et al. Monitoring of radiofrequency tissue ablation in an interventional magnetic resonance environment: preliminary ex vivo and in vivo results.  Invest Radiol. 1997;  32 671-678
  CrossrefPubMedSearch in Google Scholar
• 17 Schoenenberger A W, Steiner P, Debatin J F, et al. Real-time monitoring of laser diskectomies with a superconducting, open-configuration MR system.  AJR Am J Roentgenol. 1997;  169 863-867
  PubMedSearch in Google Scholar

Daniel Külling, M.D.

Department of Internal Medicine University Hospital of Zürich

Rämistrasse 100 8091 Zürich Switzerland

Fax: Fax:+ 41-1-255-4503

Email: E-mail:kulling@gmx.ch