An endoscopic Doppler probe: Preliminary clinical evaluation

Ulmsound in Med. & Bml. Vol. LI, No. 2, pp. 347-353, Printed in the U.S.A.

0301-5629/85 $3.00 + .Cil Pergamon Pres Ltd.

1985

aClinica1 Supplement AN ENDOSCOPIC DOPPLER PROBE: PRELIMINARY CLINICAL EVALUATION

F. E. SILVERSTEIN-~,M. DELTENRE$,G.TYTGAT§,R. W. MARTIN", W. LESTERHUIS§, A.BURETTE* and D. A. GILBERT? tDepartment of Medicine, Division of Gastroenterology RG-24, University of Washington, Seattle, WA 98 195, U.S.A. $ De Gastroenterologie, Hopital Universitaire Brugmann, Bruxelles, Belgium QThe Division of Gastroenterology, Academie Medical Center, Amsterdam, The Netherlands II Center for Bioengineering/Department of Anesthesiology RN- 10, University of Washington, Seattle, WA 98 195, U.S.A. A new pulsed Doppler catheter has been developed for use during gastrointestinal fiberendoscopy. Modern gastrointestinal endoscopy allows inspection of the wal1 of the entire esophagus, stomach, duodenum, and colon. Diagnosis is performed by inspecting the surface of the mucosa, biopsy, and cytology. However, it is not possible to determine the characteristics of the blood vessels of the wal1 beneath the mucosa with available techniques. We have developed a Doppler system which is miniature and can be passed down the biopsy Channel of standard fiberendoscopes. This 8 MHz device incorporates range limiting to select the depth of interrogation by Doppler. The length of the probe is 2 m, and the diameter is 1.8 mm. This device has been tested in animal studies for efficacy and safety and is now being tested clinically in patients. The probe has been tested for two indications: the detection of the retroduodenal artery prior to endoscopic papillotomy for retained bile duet stones (90 patients); and the detection of flow in esophageal varices before and after endoscopic sclerotherapy (33 patients). Preliminary results are encouraging and suggest that this device can give information about submucosal blood vessels which cannot be detected using other methods.

1.

therapy for bleeding esophageal varices, and foreign body removal. For some of these techniques, studies have established that the morbidity, mortality, tost, etc., are lower with the endoscopic approach than with the comparable surgical therapy (Bloom et al., 1973). A limitation of the current technology is that the endoscopist has little information about the structures beneath the mucosa. In some instances it may be important to know the location of these submucosal structures. This is especially true when dealing with endoscopic therapy where knowing the location and patency of blood vessels may be important to direct endoscopic treatment and to minimize the chance of complications. We have developed a miniature Doppler probe which can be passed down the biopsy Channel of standard fiberendoscopes to examine blood vessels beneath the mucosa. Details of the device and results of animal studies have been reported elsewhere (Martin et al., 1984). This paper wil1 describe briefly the key features of the device and report on the current status of clinical evaluations.

INTRODUCI'ION

The fiberoptic endoscope has provided access to the lining of the gastrointestinal tract. In most patients the entire esophagus, stomach, duodenum, colon, and rectum can be visualized. These devices can be passed with little or no sedation and can be used for both diagnosis and treatment (Silverstein and Rubin, 1983). For diagnosis the endoscopist is able to visualize the lining of the organ and obtain photographs. Suspicious areas can be sampled with mucosal biopsy or brush cytology to determine if malignancy is present. Dyes can be sprayed on the area to determine if abnormal mucosa is present. The endoscopist can also get a subjective impression of the pliability or distensibility of the wal1 by distending the organ with air. At the present time there are several types of endoscopic therapy which are used routinely in clinical gastroenterology. These include: polypectomy to remove benign growths, papillotomy to open the distal bile duet and release retained stones, injection sclero347

348

Ultrasound in Medicine &

2. CLINICAL

Biology

INDICATIONS

There are several clinical situations in which the Doppler probe might be useful. We evaluated the use of the probe during two procedures: esophageal variceal sclerosis and endoscopic papillotomy. 2.1. Esophageal variceal sclerosis

Esophageal varices occur when the submucosal veins of the esophagus dilate. In most cases liver disease causes obstruction of flow in the portal venous system and portal hypertension which results in dilation of the veins of the esophagus. Varices of the esophagus do not cause symptoms until bleeding occurs. In a recent survey of gastrointestinal bleeding, variceal sources were found in 10% of hemorrhagic episodes (Silverstein et al., 1981). Variceal bleeding is associated with considerable morbidity and mortality. The mortality for the first bleed is 33%, and the chance of rebleeding after the first bleed is 65% (Conn, 1974). The treatment for esophageal varices is not optimal. Several approaches have been tried for a period of time and initially felt to be successful. These include: pharmacological therapy using pitressin or propranolol, surgery to shunt the blood from the portal system into the lower pressure systemic system, and other surgical procedures such as transecting the esophagus (Editorial, 1984). However, at the present time, none of these techniques has been demonstrated to prolong survival in patients with bleeding varices. Recently an endoscopic therapeutic procedure has received renewed interest. This is injection sclerotherapy of varices, (Macdougall et al., 1982). Using either a rigid or flexible endoscope, the varix can be visualized in the esophagus and a needle inserted into or next to the varix. A sclerosing agent such as a mixture of sodium tetradecyl sulphate and absolute ethanol is then injected into or next to the varix. Several studies have demonstrated that this therapy wil1 stop the bleeding in 85-90% of patients but have not yet shown conclusively that sclerotherapy wil1 improve survival. 2.2. Endoscopic papillotomy (EPT) This is a new endoscopic therapeutic technique for the treatment of retained common bile duet stones. The incidence of retained bile duet stones afier biliary surgery is approximately 3% (Glenn, 1974). In the United States there are approximately 15,000 cases of retained stones per year. Reoperation to remove stones is associated with an increased morbidity and mortality. In the past decade a new technique has been devised for the endoscopic treat-

March/April 1985, Volume 11,Numher 2

ment of these stones. This involves the endoscopic incision of the papilla of Vater or endoscopic papillotomy (Safmny, 1978). With this procedure a catheter is placed into the papilla of Vater under endoscopic guidance. The catheter has a cutting wire on the tip which, when it is bowed and a radiofrequency current is applied, can incise the roof of the papilla. This incision opens the distal common bile duet. Stones wil1 then exit into the duodenum either spontaneously or with the assistance of special baskets and balloon catheters in 90% of patients. However, there are complications from EPT. One of the major complications is bleeding. The incidence of iatrogenic bleeding from EPT varies from 2.0% to 5.4% (Geenen and Hogan, 1980; Neuhaus and Safrany, 198 1; Classen et al., 1978; Cotton and Vallon, 198 1; Viceconte et al., 198 1; Deltenre et al., 1982). Major bleeding during EPT is probably due to the incision of the retroduodenal artery which may cross the common bile duet within 1 cm of the papilla in approximately 15% of patients. We designed a study to determine if the endoscopic Doppler probe could detect the location of the retroduodenal artery prior to EPT. 3. METHODS 3.1. The endoscopic Doppler probe (EDP)

The probe was designed for use at endoscopy. a. Size. The probe is 1.8 mm in diameter and 2 m in length. This allows the device to be passed down the biopsy Channel of al1 standard fiberoptic endoscopes (Fig. 1). b. Flexibility. The device is flexible with the transducer mounted at the tip of a Teflon endoscopic catheter. c. Frequency. To optimize signals close to the mucosal surface 8 MHz was selected. d. Directionality. At endoscopy it is not always possible to control the portion of the probe which contacts the mucosal surface. The first probe we built had a 360” radiation in a plane perpendicular to the catheter axis, so that any portion of the probe touching the mucosa would generate a Doppler signal (Fig. 2). In other circumstances it may be important to be able to control the direction of radiation from the probe to detect a blood vessel in a particular location. For this application, a directional probe with a 60” sector was built (Fig. 3). e. Ranging. Because the intestinal wal1 and surrounding structures have a rich vascular supply, pulsed Doppler was used to permit range limiting.

An endoscopic Doppler probe 0 F. E.

SILVERSTEINel al.

PARTIAL

Fig. 1. The Doppler probe is seen exiting from a side viewing duodenoscope of the type used for endoscopic

349 SECTOR

Fig. 3. The 60” partial sector probe with range limiting allows one to detect vessel 1 but not vessel 2 similar to the complete sector. However, this probe can also ignore vessel

3 because of its directionality.

papillotomy.

This allows the endoscopist to examine structures from 1-11 mm deep into the wall. Without this feature, signals from major vessels adjacent to the wal1 can confound the Doppler signal from the area studied. f. Safety. The probe has been designed to be safe for use inside the gastrointestinal tract. This was accomplished by incorporating an electrical isolation circuit to protect the subject from electrical leakage. 3.2. Animal studies Prior to clinical studies of this probe a series of animal experiments was performed in dogs. The purposes of these studies were to test the probe for efficacy and to determine the safety of the device. The results of these experiments have been described elsewhere (Martin et al., 1984). 3.3 Clinical studies There are several clinical circumstances in which this type of endoscopic Doppler probe would be COMPLETE

SECTOR

Depth Limited (D)

Fig. 2. This illustration of the 360” radiation probe also shows how range or depth limiting (distance D) can detect signals from vessel 1 and not vessel 2. Any surface of the probe is active, therefore there is no need for probe rotation.

useful. We have examined the use of the EDP during two procedures: esophageal variceal sclerosis and endoscopic papillotomy. a. Esophageal variceal sclerosis. A significant problem with both clinical and research evaluation of sclerotherapy has been the inability to determine whether flow is present in the varix before and after injection with the sclerosing agent. To address this question we designed an initial clinical trial where the endoscopic Doppler probe would be placed against the varix to answer several questions about sclerotherapy. Is the structure really a varix? Is there flow in the varix? After injection is there residual flow? And finally, does the vein need to be reinjected? To answer these questions, a preliminary clinical trial of the EDP device was planned in patients about to undergo sclerotherapy. Patients presenting to the University of Amsterdam Hospita1 in Amsterdam, Netherlands, were evaluated. The probe was tested in 33 patients with grade 3 or 4 (moderate to severe) varices. The 360” probe was used. The probe was placed on the varix, 2 cm proximal to the esophagogastric junction. The probe was either placed parallel to the varix, or at a slight angle. Depth of penetration was set at 4 mm to minimize signals from vessels adjacent to the esophagus. In 5 patients, the EDP was used to evaluate the patients’ esophageal varices just before and immediately after injection sclerotherapy. In 10 patients, the EDP was used to study the varices before and 5-21 days after sclerotherapy to monitor longer term effects. b. Endoscopic papillotomy (EPT). This study was performed at Brugmann University Hospita1 in Brussels, Belgium. In 90 patients about to undergo endoscopic retrograde cholangiopancreatography or endoscopic papillotomy, the EDP was tested. The EDP was either passed into the papilla and up the

350

Ultrasound in Medicine & Biology

common bile duet or placed on the duodenal wal1 just above the papilla. This is the region which is incised during papillotomy and where the retroduodenal artery can be inadvertently cut. For these studies a new device was added to the system incorporating an analog computer to attempt to quantitate the flow in the vessel being studied. The Doppler signals between 200 Hz-6 kHz were processed to obtain a voltage representing the time average blood velocity across the vessel cross section and a signal representing the square of the average amplitude of the Doppler signal present. These two signals were multipled and further averaged to give a signal related to blood flow. This approximate relationship reflects the fact that the square of the signal is related to the amount of ultrasonic power scattered which is in turn related to the number of blood cells present in the radiation pattern. The velocity is related to how fast the cells are moving. Therefore, the product reflects volume blood flow in an approximate manner. 4. RESULTS 4.1. Esophageal variceal sclerosis

NO complications occurred in patients with varices from studies with the EDP. Specifically, use of the EDP on varices with surface erosions did not result in bleeding. In al1 33 patients studied, a pulsatile arterial signal could be detected: 29 patients had strong arterial signals; 4 had weak arterial signals. These signals were assumed to be from the thoracic aorta or heart adjacent to the distal esophagus. Venous signals were often more difficult to hear because of motion (peristalsis, respiration, and retching) which made keeping the probe in one position on the varix difficult. In 11 cases a strong, low-pitched, constant venous signal was heard; in 15 cases a weaker venous signal was heard; in 7 cases no venous signal was

8

March/April 1985, Volume 11,Number 2

heard despite plating the probe on what appeared to be variceal structures. The venous signal was steady and not pulsatile. In one of the seven patients with endoscopic varices but without venous flow detected, a needle was passed into what appeared endoscopically to be a varix. When the needle was removed, no bleeding was noted. The investigators felt that this was a spontaneously thrombosed varix. We were clearly learning as we began using the EDP, so some of our fitst results are less specific than later. In more recent studies we could almost always find a venous signal in patent varices. In one additional patient with severe upper gastrointestinal bleeding, heavy folds were encountered in the stomach with overlying erosions. The endoscopist was unable to determine if the source of bleeding was the erosions or whether the folds were, in fact, varices. Placement of the EDP probe revealed a clear venous signal and a correct diagnosis of bleeding gastric varices was made. In the 5 patients studied just before and immediately after injection sclerotherapy, al1 had venous flow before injection. In 2 patients there was no flow immediately after injection (40%) (Fig. 4). In the remaining three patients (60%) the venous signal did not change despite the injection of adequate amounts of sclerosing material. In the 10 patients studied between 5 and 2 1 days after sclerotherapy, al1 had a clear (strong or weak) venous signal prior to sclerotherapy. At followup between 5 and 21 days after sclerotherapy, 3 patients (30%) had no venous signal detected. Endoscopically these patients’ varices had a pale appearance and a somewhat firmer consistency. In one of these three patients, later endoscopy demonstrated their varices to disappear. In the other two patients the size of the remaining varices was clearly diminished. In the remaining 7 patients (70%) the venous signal was stil1 detectable and therefore these patients underwent further sclerotherapy.

1TlMEtSEC) 1 BEFORE

SCLERDTHERAPY

AFTER SCLEROTHERAPY

G

Fig. 4. Spectrogram in a patient with esophageal varices treated with sclerotherapy. Prior to treatment a strong venous signal is present in the varix. Immediately after treatment no further venous signal is detected even at increased gain.

An endoscopic Doppler probe 0 F. E. SILVERSTEINa al.

351

-Time

i

6mm in the Common Bile Duet

Fig. 5. Spectrogram showing weak arterial signals with the probe inserted 6 mm into the common bile duet.

4.2. Endoscopic papillotomy NO complications were encountered in testing the endoscopic Doppler probe in 90 patients during ERCP. Both the complete and 60” partial sector probes were used in these studies. Vascular signals were obtained in 77 of the 90 patients (85%). The probe was successfully inserted into the common bile duet (at least 15 mm) in 78 of the 90 patients (87%). A weak arterial signal was detected in the lower common duet in 66 of these 78 patients (85%) (Fig. 5). In the 12 patients in whom the EDP could not be inserted into the common bile duet, the arterial signal was searched for on the duodenal wal1 above the papilla and was found in 11 (92%). When deep insertion into the common bile duet was obtained, strong vascular signals were audible at the junction of the cystic duet and the common bile duet (Fig. 6). Endoscopic papillotomy was then successfully performed in 46 patients after testing with the EDP (Table 1). Slight oozing of blood was observed in 13 of the 46 patients (28%) and seemed to be unrelated

to the presence of a weak arterial signal. NO patient had to be operated upon or transfused. In one case a strong arterial signal was detected within 15 mm of the papillary orifice. After a 1-2 cm papillotomy in this patient a slight oozing of approximately 5 cc was noted. The mean time added to the procedure for using the EDP was 5 min with a range of 2-10 min (Deltenre et al., 1984). In the studies using the analog computer to evaluate blood flow and check the systems reproducibility, a clearcut differente existed between the weak arterial signal in the peri-Vaterian region and the strong arterial signal in the cystic duet area (Table 2). 5. DISCUSSION

The endoscopic Doppler probe represents the application of established technology in the field of vascular studies to fiberoptic endoscopy. When one performs therapeutic procedures on blood vessels of

-Time

k------In

the Common!%! Duet Proximal to Cystic Duet -

Withdrawn ~~~~ ---+i Bik Ottct

Fig. 6. Spectrogram showing strong arterial signal at the leve1 of the cystic duet. As the probe is pulled rapidly out of the bile duet the signal disappears.

Ultrasound in Medicine & Biology

352

Table 1. Bleeding from EPT and lower common

bile duet

arterial signals Ooze

Melena

40

1 9

0 0

None

5

3

0

Total

46

13

0

Signal Strong arterial Weak arterial

n 1

the extremities, one usually directs therapy based upon results of angiography, ultrasound/Doppler, etc. We and others (Beckley et al., 1982; and McCormack et al., 1983a, 1983b) have now demonstrated that a Doppler catheter can be constructed which is smal1 enough to be passed via the biopsy Channel of any standard gastrointestinal endoscope. The design aspects which we have incorporated are beneficial in accomplishing our goals for such a unit. The complete sector probe is easiest to use in situations where it is not important which surface of the probe touches the target. The partial sector probe is more helpful when directionality is important to detect the location of a specific blood vessel. Range limiting appears to be essential because of the multiple vessels in and adjacent to the intestinal wal1 which can confuse the signal. At the present time there is no other method available for studying vessels of the intestinal wal1 during endoscopy other than endoscopic Doppler probes. Although angiography can be used, it has several disadvantages when compared to the endoscopic method. Angiography is invasive and has significant associated morbidity; the procedure cannot be performed during fiberoptic endoscopy; angiography cannot see the precise anatomy of an area such as the papilla of Vater except with special subselective catheterization. Even if the retroduodenal artery can be seen by angiography, its relationship to the papilla of Vater cannot be determined. Finally, angiography is not a good method to study esophageal varices as it either requires a delayed film after a visceral mesenteric injection or the more invasive techniques such as splenoportography. Table 2. Computerized evaluation detected in the common n

of the arterial bile duet

Mean flow

flows

Range

Papilla region (up to 20 mm into CBD)

Junction cystic dwt and CBD

45

167 f

55

35

502 +r 270

60-450

80-2000

March/April 1985, Volume 11, Number 2

The preliminary results of the clinical studies of the EPT probe in esophageal sclerotherapy have led to several conclusions: (1) that it is possible to safely detect flow in esophageal varices with the EDP, and that although movement makes detection of venous signals more difficult than arterial signals, the detection rate increases with experience; (2) that EDP can monitor the effects of sclerotherapy; (3) blood flow in varices seems to stop before the varices disappear endoscopically; (4) further clinical studies are necessary before definite conclusions about the usefulness of the EDP in esophageal sclerotherapy can be drawn. The preliminary clinical results from the use of the EDP to study the vasculature about the papilla of Vater have demonstrated that this device can evaluate the vascular pattern of the peri-Vaterian region. Ultimately the EDP may be able to prevent iatrogenic bleeding during EDP by locating the retroduodenal artery, and if this artery is in an anomalous position, too close to the papilla of Vater, alert the endoscopist to either cancel the papillotomy or make a shorter incision. Further studies are necessary to verify whether a correlation exists between the amplitude of the detected signal obtained with the analog computer and the risk of bleeding from EPT. Two groups have described the development of a smal1 Doppler catheter which can also be used at endoscopy (Beckley et al., 1982; McCormack et al., 1983a; McCormack et al., 1983b). These devices use continuous wave Doppler. The latter group have used their probe in the esophagus to examine esophageal varices. These studies indicated that in some instances the variceal flow is bidirectional or caudad rather than being entirely cephalad. The advantage of the continuous wave Doppler is that in sorne ways the circuitry is simpler than the pulse Doppler (Atkinson and Woodcock, 1982) and therefore may be less expensive. However, the continuous wave probe requires two transducer elements (instead of one) (McCormack et al., 1983a), or one transducer with a special balancing circuit (Beckley et al., 1982). The disadvantage of continuous wave Doppler is that it is not possible to range gate the signal, and therefore there is a potential problem with strong signals from deep arterial structures adjacent to the esophageal wall, particularly with the one transducer method. Only with more experience and perhaps comparisons of the two devices wil1 we be able to determine which technique is better for applications at fiberoptic endoscopy. The probe used in these studies fulfills some of the criteria necessary for successful use at endoscopy. It is smal1 enough to be passed via a standard endoscope, relatively inexpensive, and safe. The in-

An endoscopic Doppler probe 0 F. E. SILVERSTEINet al.

terpretation of signals appears to be fairly straightforward. This is important because most gastrointestinal endoscopists are not familiar with sophisticated Doppler techniques. In the future increasing applications of endoscopy for diagnosis and treatment are expected. Endoscopy has the advantages of being less expensive, less dangerous, and less complicated than the alternative surgical procedures. The type of Doppler device reported in this paper may become essential to guide the endoscopist in these treatmentsto maximize the effectiveness and safety of endoscopic therapy. Acknowledgmenf-This research was supported by a grant from Advanced Technology Laboratories, Squibb Medical Systems, Bellevue, Washington.

REFERENCES Atkinson P. and Woodcock J. P. (1982) Doppler Ultrasound and its Use in Clinical Measurement, pp. 22-26. Academie Press, New York. Beckley D. E., Casebow M. P. and Pettengell K. E. (1982) The use of a Doppler ultrasound probe for localizing blood flow during upper gastrointestinal endoscopy. Endosc. 14, 146-147. Bloom B. S., Goldhaber S. Z., Sugarbaker P. H., O’Connor N. E. (1973) Fiberoptics: Morbidity and Cost. N. Eng. J. Med. 288, 368.

Classen M., Wurbs D. and Hagenmuller F. (1978) Complications of endoscopic papillotomy, in Endoscopic Sphincterotomy of the Papilla of Vater (Edited by L. Demling and M. Classen), pp. 59-62. G. Thieme. Conn H. 0. (1974) Therapeutic portocaval anastomisis: To shunt or not to shunt. Gastroenterology 76, 1065-7 1. Cotton P. B. and Vallon A. G. (1981) British experience with duodenoscopic sphincterotomy for removal of bile duet stones. Br. J. Surg. 68, 373-375.

353

Deltenre M., Hermanus A., De Reuck M., Van Gossum M. and Rajan A. (1982) La Sphincterotome endoscopique chez le patient cholecystectomise. Rev. Med. Bruxelles 3, 263-268. Deltenre M., De Reuck M., Silverstein F. E., Martin R. W., Burette A. and Gilbert D. A. (1984) Systeme Doppler adapte a I’endoscopie digestive: etude pilote. Acts Gastroenterologica Belgica 41, 3-10.

Editorial (1984) Bleeding Esophageal Varices. Lancet 1, 139- 141. Geenen J. and Hogan W. J. (1980) Endoscopic access to the papilla of Vater. Review.s of the Fourth European Congress ESGE (Edited by M. Classen, H. Henning, E. Seifert), pp. 47-56. G. Thieme, Stuttgart. Glenn F. (1974)Retained calculi within the biliary ductal system. Ann. Surg. 179, 528-539.

Martin R. W., Gilbert D. A., Silverstein F. E., Deltenre M., Tytgat G., Gange R. and Myers J. (1984) An endoscopic Doppler probe for assessing intestinal vasculature. Ultrusound Med. Biel. in press. McCormack T., Smallwood R. H., Walton L., Martin T., Robinson P. and Johnson A. G. (1983a) Doppler ultrasound probe for assessment of blood flow in oesophageal varices. Lancet 1, 677678. McCormack T. T., Rose J. D., Smith P. M. and Johnson A. G. (1983b) Perforating veins and blood flow in oesophageal varices. Lancet 2, 1442-1444.

Neuhaus B. and Safrany L. (1981) Complications of endoscopic sphincterotomy and their treatment. Endosc. 3, 97-99. Safrany L. (1978) Endoscopic treatment of biliary tract disease: An intemational study. Lancet 2, 983-985. Silverstein F. E. and Rubin C. E. (1983) GastrointestinaI endoscopy. In Harrison’s Principles of Internal Medicine (Edited by R. G. Petersdorf et al.), 10th Edn, pp. 1683-1689. McGraw-Hill, New York. Silverstein F. E., Gilbert D. A., Tedesco F. J., Buenger N. K., Persing J. and 277 members of the American Society for Gastrointestinal Endoscopy (1981) The national ASGE survey on upper gastrointestinal bleeding 1: Study design and baseline data. Gastrointest Endosc. 27, 73-79. Viceconte G., Viceconte G. W., Tietropaolo V. and Montori A. ( 1981) Endoscopic sphincterotomy: indications and results. Br. J. Surg. 68, 376-380.