27
ERCP Management of Malignancy: Tissue Sampling, Metal Stent Placement, and Ampullectomy

Douglas A. Howell^1,2

¹ Tufts University School of Medicine, Boston, MA, USA

² Pancreaticobiliary Center, Maine Medical Center, Portland, ME, USA

Introduction

Endoscopic biliary stent placement to palliate jaundice has become the treatment of choice with self‐expanding metal stents (SEMS) carrying significant patency and cost advantages [1]. Prior to metal stent placement, a secure pathologic diagnosis is almost always warranted.

Increasingly, endoscopic ultrasound (EUS)‐guided fine needle aspiration (FNA) can safely and reliably succeed in tissue sampling. A disadvantage may be the need for a second procedure, an added small risk, problems with availability, and added cost. Access to the hepatic hilum may also be difficult or unsuccessful. EUS techniques are discussed elsewhere in this volume.

Tissue sampling at ERCP has a long history, with one of the very first sphincterotomies done to introduce a biopsy forceps for bifurcation stricture biopsy in 1973.

Comprehensive tissue sampling to achieve a high yield comparable to EUS requires a special skill set, including training, thorough knowledge of available equipment, and experience [2].

Following tissue diagnosis, biliary stent placement is now routinely appropriate in most obstructive jaundice patients with both resectable and nonresectable malignancies. The choice of an SEMS compared with a temporary plastic stent, the training, equipment choices, and techniques are addressed later in this chapter.

Finally, the special challenge of managing ampullary neoplasms is covered in the last section of the chapter.

Tissue sampling at ERCP

c27i001 To begin, specimens can be obtained from strictures of the biliary ducts and, rarely, from the pancreatic duct (PD) by aspirating bile or pancreatic juice or performing brush cytology, ERCP‐performed FNA, and forceps biopsy (Video 27.1). Tumors directly involving the duodenal wall, usually just above the papilla in the case of pancreatic cancer, can be similarly sampled.

We will discuss these techniques, the equipment, and details of training individually.

Fluid aspiration

In the simplest but lowest yield technique, bile or pancreatic juice can be collected after deep cannulation. This older technique has declined in use and is now rarely reported due to higher yield techniques.

An important observation is that traumatizing the stricture prior to aspiration increases yield. Hard dilators or balloon dilation will disrupt the surface epithelium and permit malignant cells to enter the ducts. These procedures probably should not be employed to exclusively increase yield since better techniques are now more effective.

Brush cytology

Brush cytology remains the most popular technique at ERCP for tissue sampling. The technique and equipment are straightforward.

First, a guide wire is maneuvered through the stricture, usually of the CBD or biliary bifurcation. After an exchange to leave the guide wire in place, a device is placed for brushing.

Several devices are marketed to permit guide wire‐assisted sampling:

Geenen brush (Cook Endoscopy): A guiding catheter is placed over the guide wire and maneuvered above the stricture. A sphincterotomy is not necessary. The guide wire must be removed, and the Geenen brush is then advanced through the guiding catheter, advanced out through the end, and the catheter is pulled below the stricture. The spring tip of the brush will maintain the position of the brush within the stricture during brushing. Both brush and catheter then are removed after the brush is pulled back into the catheter to prevent the loss of cells. The brush tip is subsequently cut off with wire cutters and placed immediately into transport media. The major disadvantage of this older technique is the loss of guide wire access requiring the replacement of the guide wire by recannulation.
Monorail brush: The above described equipment can be altered to maintain access by punching a small hole with a sharp 21‐G needle about 1–2 cm inside the guiding catheter tip. The guide wire then enters the tip and exits the punched hole. The Geenen brush can then be preloaded prior to advancement over the in‐place guide wire. The major difference is that the guide wire must be carefully pulled back to free the guide wire from the guiding catheter after it has been positioned well above the top of the stricture. This has been termed an “intraductal exchange.” At no time should the guide wire tip be pulled back below the stricture. After this maneuver, the brush is extended and the stricture is sampled by brushing to and fro. The tip is then pulled back into the catheter, and the device is removed leaving the original guide wire in place.

Several newer devices come preloaded and prepunctured for speed and ease (Fusion Cytology Brush, Cook Endoscopy; RX Cytology Brush, Boston Scientific Corp.). These devices, like the monorail device, can be used with a short (260‐cm) guide wire.
Double‐lumen technique: Two over‐the‐guide wire techniques using double‐lumen devices are in use. Preloaded 6‐ and 8‐Fr catheters have a guide wire lumen and a preloaded brush (Cytomax, Cook Endoscopy; RX Cytology Brush, Boston Scientific Corp.). In their use, a long guide wire remains in place, and the device is simply inserted, and the stricture is brushed. The smaller 6‐Fr device requires a 0.021 guide wire to be in place, and the 8 Fr is compatible with the standard 0.035 guide wire. This larger device is rather stiff and certainly not optimal for PD sampling.

Several points about brushing need to be made at this point:

Do not pull the brush out of any device as most of the cells will be lost. Always “park” the brush just inside the tip of the device; then remove the entire device being used.
You must stent the pancreas after transtricture brush sampling of a pancreatic stricture. This explains the rare use of this technique as the management of that stent is then problematic. Therefore, we only use this technique when a pancreatic stent is warranted such as treating a pseudocyst or fistula above a suspected malignant stricture.
Brush cytology has a low yield despite its popularity. Expect only an 8–30% true positive yield. Other techniques have a higher yield, and when these are employed, brushing adds little. In our large series of forceps biopsy followed by ERCP FNA, brushing added no additional positives [3]. Our center no longer uses brush cytology as we will discuss later.

Fine needle aspiration

In an attempt to increase positive results compared to the disappointing ERCP brush cytology yield, a Chiba‐type 22‐G needle for FNA in the bile duct was introduced several years ago (HBAN 22, Cook) [4]. This device increased yield by sampling deeper than a brush. The technique of ERCP FNA using this device requires sphincterotomy and then freehand placement. Technically, placement is moderately difficult due to needle length.

The device has a smooth atraumatic ball tip and is loaded with a retractable 22‐G 7‐mm‐long needle. The needle‐containing catheter is precurved to assist in cannulation.

Placement requires placing the ball tip on the lower edge of the sphincterotomy and sliding the device upward while slightly advancing the endoscope tip. Placement can almost always be accomplished but does demand some hands‐on training and experience.

The needle tip is then advanced into the undersurface of the stricture and forcefully thrust in deeply. After removal of the stylet, gentle aspiration with 10‐cc dry syringe is performed, the needle is then withdrawn, and the content of the aspiration is expressed into transport media [5].

Forceps biopsy and cytology

In multiple studies, biopsy forceps have the highest yield in tissue sampling at ERCP [6]. Multiple forceps, which can be relatively easily introduced, are available, both reusable (ERCP forceps, Olympus) and disposable (Cook and Boston Scientific Corp). In general, pediatric forceps are necessary as gastroscopy standard‐size forceps are excessively stiff when advanced over the elevator.

Biopsies can be obtained from proximal strictures or distal biopsies (Figures 27.1 and 27.2).

Introduction requires sphincterotomy. Multiple biopsies are advised with a significant increase in yield by performing six biopsies rather than the standard three [7].

In summary, brush, FNA, and forceps can be employed individually at ERCP, but true positive yields of each technique vary and are generally about 30–40%.

Photo depicts freehand 6-Fr pediatric forceps biopsy of a low stricture at ERCP. — **Figure 27.1** Freehand 6‐Fr pediatric forceps biopsy of a low stricture at ERCP.

Photo depicts forceps biopsy of a Bismuth IIIA tumor of the bifurcation using the Howell biliary introducer. — **Figure 27.2** Forceps biopsy of a Bismuth IIIA tumor of the bifurcation using the Howell biliary introducer.

Triple sampling

To address these low yields of each of these techniques, centers have reported combining sampling procedures in a single setting. When using all three techniques, called “triple sampling,” significantly higher yields are produced [6].

Triple sampling requires considerable effort, training, and experience. At present, reports have been confined to only a few dedicated centers. Even so, true positive yields of tissue sampling have only approached 70–80% for pancreatic cancer.

Intraprocedural ERCP tissue diagnosis

A major disadvantage of all techniques of ERCP tissue sampling has been the necessary delay in tissue processing before a definite diagnosis can be made. Although a clinical diagnosis can at times still permit permanent stent placement, only a temporary plastic or a potentially removable coated SEMS should generally be used without a tissue diagnosis.

To address this problem, we have developed a technique of intraprocedural diagnosis during a single ERCP, which permits appropriate stent placement. This forceps technique was patterned after the neurosurgical technique of intraoperative brain squash prep. In brief, the resection margins are biopsied and simply spread onto a glass slide for Papanicolaou staining during the operation. Rapid reading by a cytologist then guides further resection.

c27i001 c27i001 For sampling malignant appearing biliary strictures, a 5‐Fr or 6‐Fr forceps specimen is obtained from the lower aspect of the stricture and then forcefully smashed into a monolayer on a dry glass slide (Figures 27.3 and 27.4). An in‐suite cytotech then immediately stains sequential specimens, which are examined by the cytopathologist. More specimens follow until a diagnosis is definitely made (Videos 27.2 and 27.3).

Photo depicts small 6-Fr forceps biopsy specimen on saline-moistened pad—appropriate for immediate SMASH prep. — **Figure 27.3** Small 6‐Fr forceps biopsy specimen on saline‐moistened pad—appropriate for immediate SMASH prep.

Photo depicts vigorous smashing of the forceps specimen between two dry glass slides to form a monolayer before rapid Papanicolaou staining. — **Figure 27.4** Vigorous smashing of the forceps specimen between two dry glass slides to form a monolayer before rapid Papanicolaou staining.

Using this technique, cholangiocarcinomas can almost always be adequately sampled, often in just a few biopsies, but pancreatic cancer strictures are more difficult requiring 10 or 12 specimens at times. Overall, the intraprocedural diagnosis of pancreatic cancer causing a biliary stricture is about 75%. We then employ a comprehensive approach if 10–12 SMASH preps are negative by sending additional biopsies for histology and performing at least one ERCP FNA. This raises the yield to nearly 90% for pancreaticobiliary cancers [8].

Beware that metastatic disease producing biliary obstruction is much deeper and less infiltrative so the yield of ERCP tissue sampling, even by this comprehensive approach, is only 50%. EUS is clearly the choice in this setting.

Metal stent placement

After a definite diagnosis of malignancy has been made, most centers choose to place temporary plastic or permanent metal expandable stents. This permits clearance of jaundice, referral and further workup, and appropriate surgical or oncology consultation. A major recent development has been the practice of placing SEMS in both resectable and nonresectable patients [9, 10]. This strategy can also facilitate neoadjuvant chemoradiation to downstage borderline resectable cases to then permit successful Whipple’s resection, a strategy requiring some months of successful stenting [11].

Having a successful stent in place with preoperative clearance of jaundice also tends to simplify the surgery if, unfortunately, unresectability is discovered—a not infrequent event. Operating on an undrained patient places considerable pressure on the surgeon to perform an otherwise unnecessary and technically challenging biliary bypass.

Stent placement requires considerable knowledge, training, and experience as the consequences of stent misplacement can be very serious.

The first question facing the endoscopist is stent selection: plastic vs. metal. If metal is indicated, should the SEMS be coated or uncoated? What should be the overall length and diameter? Should the design be a foreshortening or nonforeshortening type? These decisions are greatly influenced by tissue diagnosis, as discussed above, and stricture location. The major types of SEMS are shown in Figure 27.5.

There are several scenarios for distal malignant obstruction that we can outline.

Tissue‐proven cancer, not resectable

Decision about stent type is based on life expectancy. A poor Karnofsky score with a low functional status suggests a short life span. For instance, metastatic disease to the liver of greater than 10% predicts short life expectancy. A 10‐Fr or an 11.5‐Fr plastic stent with a median patency of 90–120 days supports plastic stent selection in this setting. Be aware the 11.5‐Fr diameter is more difficult to place and cannot be removed through the endoscope but probably does add some days or weeks of patency.