A Global Positioning System for the Colon?

Like the ubiquitous global positioning system (GPS), which uses orbiting satellites to determine the precise location of a receiver anywhere on Earth, the magnetic endoscope imager (MEI) is a technically advanced system capable of pinpointing the precise three-dimensional location of a colonoscope within the body. While GPS has revolutionized air, land, and sea travel, it remains to be seen whether MEI will have an equally profound impact on colonoscopic navigation.

Magnetic colonoscopic imaging was developed independently by two groups of clinicians and physicists in the United Kingdom in the early 1990 s [1] [2]. The early prototypes used a specially-designed catheter passed through the accessory channel of standard endoscopes to provide real-time three-dimensional imaging of the colonoscope’s location during colonoscopy. Subsequently, the system has been refined, and it is now commercially available in the form of a colonoscope with a built-in coil-based imaging system. The advantage of the catheter-based system is that potentially any colonoscope or endoscope hanging in one’s closet may be adapted for three-dimensional imaging and localization. The advantage of the prototype magnetic imaging colonoscope is the elimination of the need to pass an imaging catheter separately during the procedure, thereby freeing the accessory channel for suctioning, lavage, and therapeutic maneuvers. The article by Wehrmann and Frühmorgen in the current issue of Endoscopy reports on initial clinical experience with prototypes of this new commercial imaging colonoscope [3].

Although the details of the magnetic imaging system design are proprietary, the physical principles of the device are conceptually straightforward. The position of the tip of the colonoscope, and multiple additional points in the shaft of the instrument, are mapped in three dimensions using sensor coils that are installed in the instrument (or the catheter) by the manufacturer [4] [5]. Sensor coils measure the rate of change of magnetic field flux according to Faraday’s law of induction: the change in magnetic field flux through the coil produces a voltage in the coil circuit that is measured. Therefore, by producing precisely controlled pulsed magnetic fields using magnetic field generators located near the patient (this is conceptually reminiscent of magnetic resonance imaging, in which much stronger precisely controlled magnetic fields are employed) it is possible to measure the location and orientation of each coil quantitatively. This information is collected several times each second, and is used to produce an image of the colonoscope on the viewing screen.

Previous studies using earlier versions of the magnetic imaging system confirmed and refined several notions that clinicians performing colonoscopy have long suspected, either from experience or from using fluoroscopy during procedures [6] [7]. Loops invariably form during colonoscopy. There are loops in the sigmoid and loops in the transverse colon; without imaging, endoscopists know when looping is occurring, but can only guess at the type and location of the loops [8] [9]. With imaging, the formation and reduction of loops is clearly demonstrated. In addition, by placing a mobile sensor coil in the palm of the hand of the endoscopy assistant, it is possible to monitor the effect of abdominal pressure on looping and target the application of pressure for improved effect [8].

The three-dimensional location of the colonoscope tip is also demonstrated with imaging. Although colonoscopy reports typically chronicle the journey through various parts of the colon, the inability of endoscopists to determine the precise location of lesions has long been appreciated, and several endoscopic methods such as tattooing and clipping have been developed to address this problem [10]. In an article by Shah et al. also published in the current issue of Endoscopy [11], the application of magnetic imaging for anatomical localization of the colonoscope tip was studied prospectively. In 29 consecutive patients, the imaging system was used to determine when the colonoscope tip was located at one of four locations: the sigmoid-descending junction (eight patients), mid-transverse colon (four patients), hepatic flexure (11 patients) or splenic flexure (six patients). A random allocation determined which one of the four locations was to be the target in a given patient. Once the location was identified using the imager, two endoscopic clips were placed, contrast was injected into the bowel lumen, and a conventional radiograph was taken to allow precise identification of the clipped site.

Twenty-six of the 29 nine placements were correct, using the radiography as the gold standard. The three incorrect placements had the following causes: 1) the descending colon was mistaken for the sigmoid-descending junction in one of eight patients; 2) the distal transverse colon was mistaken for the mid-transverse in one of four patients; and 3) the distal ascending colon was mistaken for the hepatic flexure in one of 11 patients. The splenic flexure was identified correctly in all six patients. All of the misidentifications were minor, in the sense that the magnetic and radiographic locations were not very distant from each other, and the authors point out that if the magnetic imager had been used as the sole method to assist a surgeon in resecting a malignancy at any of these locations, the resected area would probably have included the target lesion. However, the same authors, in an article published 2 years ago, demonstrated that the endoscopist’s unaided assessment of the scope tip position in the colon was 85 % accurate when compared to the magnetic imager [8]. We now learn that the magnetic imager is 90 % accurate when compared to the radiographic gold standard, and this naturally raises the question of how many of the 15 % of cases in which the unaided endoscopist’s assessment differed from that of the magnetic imager were due to endoscopist error, how many to imager error, and how many to both? In routine practice, endoscopists typically use India ink tattoos or endoscopic clips and radiography to assist surgeons in locating lesions requiring surgery. Both of these methods have been studied extensively and shown to be simple, accurate, and safe [10]. For magnetic localization to be accepted as a viable alternative, we need convincing evidence that the magnetic imager offers at the very least comparable accuracy.

In addition to assisting clinicians in localizing the colonoscope, the magnetic imager has the potential to improve our ability to rapidly reach the cecum. However, attempts to demonstrate a clear objective advantage for the use of magnetic scope imaging in the setting of a clinical trial are hampered by the overwhelming ability of expert endoscopists to rapidly reach the cecum in nearly all patients. It is therefore difficult to demonstrate a higher rate of cecal intubation with the device. In the study by Shah et al. [8], the cecum was reached in all 100 patients in a mean of 12 min without the use of magnetic imaging. In another study by Shah et al. published in the Lancet in the same year [9], the cecum was reached in only 91 % without the imager and 100 % with the imager; however, in the protocol used, the authors considered it unethical to withhold use of the imager if there was no substantial progress within 10 min at any point in the procedure, so that the eight cases that were “unsuccessful” were completed using the imager. It is certainly possible that by allowing more time, the examinations could have been completed without the imager. In the second of these studies [9], cecal intubation by expert endoscopists was minimally faster (8 vs. 9 min) when the imager was used, but this did not take into account the time required to set up the imaging system. These initial results suggest that there may be a slight improvement in both the ability and time required to reach the cecum when the magnetic imager is used. This may be particularly true for physicians in training, when looping of the instrument is responsible for substantially lower success rates of cecal intubation and long procedure durations. That question was partially addressed in the Lancet study [9]: 113 patients were examined by trainees randomly assigned to imager assistance versus no imager assistance. The trainees were relatively advanced, having performed more than 200 previous examinations each. Using the imager, they were able to reach the cecum in 100 % of the cases in an average of 11.8 minutes, while without the imager they reached the cecum in 89 % in an average of 15.3 minutes. The differences were statistically significant. There is certainly potential for the magnetic imager to improve the training and performance of colonoscopy, particularly in the setting of physician training, and we will eagerly await the results of future studies as the system becomes widely available.

The study by Wehrmann and Frühmorgen in this issue describes the authors’ experience with several prototypes of the new magnetic imaging system. In the first part of their study, 69 patients were examined using a probe-based system and 64 patients using an integrated colonoscope system with built-in imaging coils. As expected in a study in which procedures were performed by expert endoscopists, the rate of cecal intubation was well over 90 %, with half of the failures being due to tumor stenosis or poor preparation quality - factors beyond the control of the operator. The cecum was reached on average in 6.8 min. This was similar to the average time of 6.5 min in 100 control patients in the same clinic. The authors note that fluoroscopy was available at all times in their unit, but it is unclear how often it was used to help reach the cecum in the control cases.

The most interesting part of the Wehrmann and Frühmorgen study is their careful comparison of fluoroscopy and magnetic imaging. The authors were able to demonstrate that, in comparison with fluoroscopy, there were substantial inaccuracies in the magnetic imager. The catheter-based system was particularly inaccurate, incorrectly assessing 21 % of loops, while the early prototype of the integrated colonoscope system was incorrect in 5 % of the loops. In addition, there were substantial inaccuracies in the assessment of the scope tip: up to 7 cm (“little discrepancy”) in half of the patients and complete lack of correlation with fluoroscopy in 19 % of the patients with the catheter system. The integrated colonoscope had “little discrepancy” in 63 % and a complete lack of correlation with fluoroscopy in 5 % of the patients. These results are particularly interesting, as previous studies of magnetic imaging have not presented this type of accuracy data; although earlier studies did not use the same prototype, we must consider that there may have been similar issues in those catheter-based systems as well. In the second part of the report by Wehrmann and Frühmorgen, an improved prototype of the integrated colonoscope system was tested in 25 patients. The newer system performed better - detecting 100 % of the loops and precisely localizing the tip in 80 %, with “little discrepancy” in the remaining 20 %. These data indicate that the new integrated system delivers what it promises: relatively accurate measurement of the colonoscope location.

Magnetic imaging colonoscopy has arrived at our endoscopy centers in the form of a commercially available instrument with integrated imaging coils. The question remains, “Do we need such a device?” Perhaps a good analogy is the automobile navigation system. It is certainly possible to drive all your life without one, but if the car comes already equipped with it or if you can get it without paying too much extra, then it may be a nice addition to have a clear picture of where you are! All over the world, physicians are performing colonoscopy very successfully without relying on extracorporeal magnetic imaging. The studies reported in this issue of Endoscopy do not yet convince us that this device will have a major impact on the practice of expert endoscopists or busy community gastroenterologists. The most important role for such a novel and compelling device may lie in its ability to demonstrate to our trainees the shape and location of the loops that they form when performing colonoscopy. Indeed, pioneers in colonoscopy performed their procedures under fluoroscopic guidance, a practice long since abandoned. The lessons learned performing procedures under fluoroscopy have been lost on subsequent generations of endoscopists. Perhaps magnetic endoscope imaging will give the next generation of trainees an appreciation for loop formation and reduction and thereby shorten the learning curve. This would certainly be welcomed both by mentors of endoscopy and by countless numbers of patients.

References

• 1 Bladen J S, Anderson A P, Bell G D. et al . Non-radiological technique for three-dimensional imaging of endoscopes.  Lancet. 1993;  341 719-722
• 2 Williams C, Guy C, Gillies D, Saunders B. Electronic three-dimensional imaging of intestinal endoscopy.  Lancet. 1993;  341 724-725
• 3 Wehrmann K, Frühmorgen P. Colonoscopic use of a new 3-D magnetic imaging system.  Endoscopy. 2002;  34 905-908
• 4 Guy C, Gillies D F. Probe system. United States Patent #5,429,132. 1995
• 5 Williams C B, Saunders B P, Bell G D. et al . Real-time magnetic three-dimensional imaging of flexible endoscopy.  Gastrointest Endosc Clin N Am. 1997;  7 469-475
• 6 Saunders B, Bell G D, Williams C B. et al . First clinical results with a real time, electronic imager as an aid to colonoscopy.  Gut. 1995;  36 913-917
• 7 Waye J D. Imaging of the colonoscope: magnetic, fluoroscopic, or neither?.  Gastrointest Endosc. 2000;  52 131-133
• 8 Shah S G, Saunders B P, Brooker J C, Williams C B. Magnetic imaging of colonoscopy: an audit of looping, accuracy and ancillary maneuvers.  Gastrointest Endosc. 2000;  52 1-8
• 9 Shah S G, Brooker J C, Williams C B. et al . Effect of magnetic endoscope imaging on colonoscopy performance: a randomised controlled trial.  Lancet. 2000;  356 1718-1722
• 10 Ellis K K, Fennerty B. Marking and identifying colon lesions: tattoos, clips and radiology in imaging the colon.  Gastrointest Endosc Clin N Am. 1997;  7 401-410
• 11 Shah S G, Pearson H J, Moss S. et al . Magnetic endoscope imaging: a new technique for localization of colonic lesions.  Endoscopy. 2002;  34 900-904