Sampling at ERCP may be performed at the level of the papilla or of the biliopancreatic ducts. Samples collected at the level of the biliopancreatic ducts allow for diagnosing malignancy with a specificity close to 100% but present a moderate sensitivity in most studies. In this article, the different aspects of sampling at ERCP are discussed, and a special focus is placed on the means that are routinely available to the endoscopist for obtaining a high sensitivity for the diagnosis of malignancy.
- •
Samples collected at the level of the biliopancreatic ducts allow for diagnosing malignancy with a specificity close to 100% but present a moderate sensitivity in most studies.
- •
A learning curve has been demonstrated for the team endoscopist-cytopathologist to reach a satisfactory sensitivity for the diagnosis of malignancy.
- •
Strategies to increase sensitivity for cancer diagnosis include the following:
- ○
Improved sampling method, using a grasping basket for cytopathological examination or combining brushing with biopsy sampling;
- ○
On-site microscopic examination of samples collected for cytopathological as well as histopathological examination;
- ○
The use of liquid-based cytology methods, combined or not with direct smears;
- ○
The use of ancillary diagnostic techniques to analyze cytopathological samples.
- ○
Indications
Sampling at endoscopic retrograde cholangio-pancreatography (ERCP) may be performed at the level of the duodenal papilla, of the biliary tree (including the right and left intrahepatic ducts), of the Wirsung duct and of the Santorini duct. In all of these locations except the papilla, sampling for cytopathological as well as histopathological examinations may be valuable.
- •
Papillary biopsy specimens may help in diagnosing autoimmune pancreatitis, immunoglobulin (Ig)G4-related sclerosing cholangitis, and ampullary neoplasm.
- •
In the biliary tree, sampling is recommended every time that a stricture with no established or evident diagnosis (eg, anastomotic stricture following liver transplantation) is demonstrated. Dominant strictures in primary sclerosing cholangitis (PSC) pose a particular challenge.
- •
In the main pancreatic duct, sampling is usually recommended in case of dilation of unknown etiology without stricture (suspicion of intraductal papillary mucinous neoplasm) or in case of stricture. In patients who undergo endoscopic drainage of the main pancreatic duct for painful chronic pancreatitis, stricture sampling is usually performed at the time of first stent insertion because of the increased risk of pancreatic cancer in these patients.
Sample procurement and preparation
Sampling for Cytopathological Examination
Material
Brushes are the most commonly used devices to sample the biliopancreatic ducts; they provide material for cytopathological examination. Models with a double-lumen catheter should be preferred for an easy, over-the-wire, manipulation ( Fig. 1 ). The flagella at the tip of some brush models is useful in angulated ducts but it may hinder brush movements inside the intrahepatic ducts. Single-lumen cytology brushes are rarely used nowadays; they should not be used over the wire because retrieving the brush through the whole length of the catheter (to replace it with a guide wire) causes a substantial loss of cellular material on the inner side of the catheter.
Another device, the grasping basket, has been designed for sampling the biliary tree only ( Figs. 2 and 3 ). Compared with brushes, it provides more material for cytopathological examination. It is available in some countries only.
The Soehendra stent retriever is sometimes used to pass very tight strictures; in some cases material adequate for cytopathological examination may be retrieved from the serrations of the retriever. In the same vein, Devereaux and colleagues have placed 155 plastic stents into a fixative after retrieval from the biliary and pancreatic ducts. Cellular debris collected from the stent by a cytotechnician in the cytopathology laboratory allowed for diagnosing cancer with a sensitivity and specificity of 11% and 100%, respectively.
Methods
Sampling of biliopancreatic strictures for cytopathological examination is simple but meticulous execution is crucial to obtain decent results.
- •
First, all the material necessary for adequate sample preservation should be prepared before beginning the sampling procedure ( Fig. 4 ). This is crucial because desiccation of monocellular layers obtained after smearing is extremely rapid.
Fig. 4
Example of bench prepared before performing biliary brushing at ERCP. From left to right, glass slides labeled with patient identification; container filled with alcohol for preservation of smeared samples, wire cutters to separate the cytology brush from its supportive wire, and vial of PreservCyt solution (Cytic Corp., Boxborough, MA, USA). Bench preparation beforehand is critical to limit preservation artifacts.
- •
Standard brushing consists of introducing the brush/sheath as a unit over a guide wire through the stricture, withdrawing the unit brush/sheath immediately below the stricture, and then moving the brush back and forth precisely within the stricture at least 10 times. All these procedures should be performed under continuous fluoroscopic control with a sufficient magnification and amount of contrast medium in the biliary tree to ensure that the brush is correctly positioned within the stricture. Pulsed mode is recommended to limit radiation doses. The brush is then pulled into the sheath still located immediately below the structure and the unit brush/sheath is finally removed.
- •
If the grasping basket is used, the stricture is first dilated using a 6-mm balloon, then the basket (compressed in its sheath) is introduced upstream from the stricture, deployed and withdrawn through the stricture; the basket is finally retracted into its sheath still located immediately below the structure.
- •
After pancreatic brushing, a plastic pancreatic stent is usually inserted through the stricture to prevent post-ERCP pancreatitis, although the supporting evidence is weak in the absence of pancreatic sphincterotomy.
Diagnostic accuracy
In 18 studies ( Table 1 ), biliary brushing yielded a mean sensitivity for cancer diagnosis of 48%, with results extremely discordant between studies (from 9% to 93%); specificity was close to 100%. When averaging all studies listed in Table 1 , biliary brushing was significantly less sensitive for the diagnosis of pancreatic as compared with biliary cancer (41% vs 59%, respectively; P <.0001). This lower sensitivity is generally attributed to the fact that biliary strictures caused by pancreatic cancer may be related to a compression, as opposed to an invasion, of the biliary wall and to a possible low cellularity and desmoplastic reaction in the case of pancreatic cancer.
| First Author, Year | Sensitivity | Specificity | ||
|---|---|---|---|---|
| Overall | Pancreatic Cancer | Bile duct Cancer | ||
| Venu et al, 1990 | 77% (23/30) | 60% (3/5) | 80% (20/25) | 100% (88/88) |
| Rupp et al, 1990 | 93% (27/29) | 91% (21/23) | 100% (6/6) | 88% (7/8) |
| Foutch et al, 1991 | 45% (5/11) | 0% (0/6) | 100% (5/5) | 100% (3/3) |
| Ryan, 1991 | 34% (10/29) | 30% (6/20) | 44% (4/9) | 100% (17/17) |
| Howell et al, 1992 | 9% (2/22) | 0% (0/18) | 50% (2/4) | 100% (5/5) |
| Kurzawinski et al, 1993 | 64% (21/33) | 65% (15/23) | 60% (6/10) | 100% (7/7) |
| Ferrari et al, 1994 | 46% (18/39) | 66% (16/29) | 20% (2/10) | 100% (22/22) |
| Ponchon et al, 1995 | 33% (15/45) | 15% (3/20) | 44% (12/25) | 97% (64/66) |
| Sugiyama et al, 1996 | 48% (15/31) | 36% (5/14) | 59% (10/17) | 100% (12/12) |
| Mansfield et al, 1997 | 45% (20/44) | 38% (10/28) | 63% (10/16) | 100% (2/2) |
| Vandervoort et al, 1999 | 14% (8/56) | 11% (5/46) | 30% (3/10) | 100% (37/37) |
| Glasbrenner et al, 1999 | 53% (27/51) | 35% (11/31) | 80% (16/20) | 90% (19/21) |
| Jailwala et al, 2000 | 24% (18/76) | 24% (11/46) | 23% (7/30) | 100% (29/29) |
| Farrell et al, 2001 | 71% (20/28) | 78% (14/18) | 60% (6/10) | 83% (10/12) |
| Fogel et al, 2006 | 33% (42/126) | 36% (32/88) | 26% (10/38) | 100% (8/8) |
| Kitajima et al, 2007 | 67% (21/36) | 60% (9/15) | 71% (15/21) | 100% (7/7) |
| Dumonceau et al, 2008 | 43% (20/42) | 38% (11/29) | 69% (9/13) | 100% (6/6) |
| Kawada et al, 2011 | 84% (72/86) | 71% (35/49) | 100% (37/37) | 100% (8/8) |
| Total | 48% (387/814) | 41% (207/508) | 59% (180/306) | 98% (351/358) |
Various methods have been tested to increase the sensitivity of sampling methods for cancer diagnosis at cytopathological examination:
- •
Repeat brushing: performing 2 brushings, 1 before and 1 after stricture dilation using 2 different brushes during a single ERCP, increased the sensitivity for cancer diagnosis in 2 studies (from 34% to 44% [ P = .001] and from 40% to 45% [ P >.05]).
- •
Rapid on-site cytopathological examination (ROSE): if ROSE is unexpectedly negative for malignancy, some investigators repeat brushing.
- •
Grasping basket: in a randomized controlled trial (RCT) with blind examination of biliary samples collected using either a brush or the grasping basket, the grasping basket yielded a higher sensitivity for cancer diagnosis than brushing (80% vs 48%, respectively; P = .018). Complications at 30 days were standard in terms of incidence (7%), severity, and type. In another study, the grasping basket was also found to yield more cellular material than brushing. Another alternate sampling device, a long brush with stiff bristles, did not provide a higher sensitivity for cancer diagnosis compared with a standard brush despite a higher cellular yield.
- •
Stricture dilation: stricture dilation aims at exposing deep tissue layers, which may be useful for example in case of biliary compression by a pancreatic cancer. Four studies were identified that compared the diagnostic yield of biliary brushing preceded or not by stricture dilation ( Table 2 ). No significant difference was found in terms of sensitivity for cancer diagnosis but a single study systematically used balloons (hence larger dilation) for stricture dilation so that this technique likely deserves further controlled trials.
Table 2
Influence of stricture dilation before biliary brushing
First Author, Year
Design for Stricture Dilation
Dilation Diameter, mm
Sensitivity for Cancer Diagnosis
No Dilation
Dilation
Farrell et al, 2001
Different patients, not randomized
3.3
85% (17/20)
57% (8/14)
de Bellis et al, 2003
Brushing repeated in all patients following dilation
3.3–8.0
34% (40/116)
31% (36/116)
Ornellas et al, 2006
Brushing repeated in all patients following dilation
3.3
40% (16/40)
28% (11/40)
Dumonceau et al, 2008
Different patients, not randomized
6.0
35% (8/23)
59% (16/27)
Total
41% (81/199)
34% (71/207)
Biopsy Sampling
Material
Biopsy forceps of varying sizes and shapes may be used for sampling the biliary or pancreatic ducts under fluoroscopic control using a standard duodenoscope; forceps specifically designed for ERCP sampling are available ( Fig. 5 ). If no sphincterotomy is performed, a biopsy forceps with a slide slit accepting a guide wire may be useful (available in some countries only).
Methods
The simplest technique consists of introducing a standard biopsy forceps transpapillary after endoscopic sphincterotomy. Because of the stiffness of standard biopsy forceps, sampling is facilitated by the insertion of a guide wire into the desired duct (for biliary sampling, the endoscope is positioned well below the papilla to lean the biopsy forceps against the duodenal wall). Biopsies are usually obtained from the inferior aspect of the stricture but they may also be obtained from the middle part of the stricture. Biopsy samples are placed into preservative as standard biopsies, except if ROSE is used (see later in this article). As pancreatic stenting for the prevention of post-ERCP pancreatitis is underused, it is important to recall that, in case of pancreatic sphincterotomy and biopsy sampling, the insertion of a plastic pancreatic stent is recommended.
Peroral cholangiopancreatoscopy is usually reserved for patients with an incomplete diagnosis after a workup using standard techniques because, compared with standard ERCP, it presents a series of limitations. These include a limited availability, more requirement for sphincterotomy, increased incidence of complications, increased procedure duration and costs. Furthermore, as far as sampling is considered, the effectiveness of peroral cholangiopancreatoscopy seems limited: in a prospective clinical cohort study that involved 15 endoscopy referral centers, sensitivity for cancer diagnosis using single-operator cholangioscopy-directed biopsy sampling was 49%. The sensitivity calculated following the intention-to-diagnose principle is likely lower as procedural success of biopsy sampling was reported in approximately 90% of cases. Visual assessment of the bile duct wall may be more informative than sampling, in particular if ultrathin gastroscopes are used because they provide better images. This technique, named direct peroral cholangioscopy, allowed in a recent feasibility study for sampling some indeterminate biliary strictures; however, it is feasible in selected cases only and it still requires much improvement.
Diagnostic accuracy
In 6 studies ( Table 3 ), biopsy sampling at ERCP yielded a mean sensitivity for cancer diagnosis of 55%, with large variations between studies (from 32% to 82%); specificity was close to 100%. Contrary to a common belief, the diagnostic yield of biliary biopsy sampling is thus not much higher than that of biliary brushing (55% vs 48%, respectively; Table 3 and Table 1 ). When averaging all studies listed in Table 3 , biliary biopsy sampling was significantly less sensitive for the diagnosis of pancreatic as compared with biliary cancer (46% vs 63%, respectively; P = .02). This lower sensitivity is attributed to the frequent extraluminal location of pancreatic cancers.
| First Author, Year | Sensitivity | Specificity | ||
|---|---|---|---|---|
| Overall | Pancreatic Cancer | Bile Duct Cancer | ||
| Kubota et al, 1993 | 82% (18/22) | 50% (2/4) | 89% (16/18) | 100% (5/5) |
| Ponchon et al, 1995 | 45% (13/29) | 46% (6/13) | 44% (7/16) | 97% (35/36) |
| Sugiyama et al, 1996 | 81% (25/31) | 71% (10/14) | 88% (15/17) | 100% (12/12) |
| Jailwala et al, 2000 | 32% (24/76) | 33% (15/46) | 30% (9/30) | 100% (10/10) |
| Tamada et al, 2002 | 73% (27/37) | 50% (6/12) | 84% (21/25) | 100% (18/18) |
| Kitajima et al, 2007 | 58% (21/36) | 60% (9/15) | 57% (12/21) | 100% (7/7) |
| Total | 55% (128/231) | 46% (48/104) | 63% (80/127) | 99% (87/88) |
Various methods have been tested to increase the sensitivity of biopsy sampling for cancer diagnosis:
- •
Number of biopsy samples, size of biopsy forceps: in a RCT, 31 patients with a suspected hilar malignant biliary stricture were randomized to have 3 biopsies performed using either a standard or a jumbo biopsy forceps (7F or 9F, respectively); sampling was then repeated using the alternate biopsy forceps (3 biopsies again). Overall, the sensitivity for cancer diagnosis based on 6 biopsies was 58.6% without a difference between standard and jumbo biopsies. The sensitivity increased with the number of biopsies up to a maximum reached with 4 biopsies (37.9%, 48.3%, 55.2%, and 58.6% with 1, 2, 3, and 4 biopsies, respectively).
- •
Site of biopsy sampling: findings at percutaneous cholangiography in patients with a cholangiocarcinoma suggest that the number of biopsies required to reach a high sensitivity is related to the morphologic type of the tumor, with 3 biopsies from the tip of the lesion being sufficient with lesions of the papillary type, whereas more biopsies from the margins and within the stenosis would be required for lesions of the nodular or infiltrating type.
- •
ROSE: biopsy samples may be examined on-site following a particular preparation (see later) ; biopsy sampling may thus be repeated until ROSE yields a result positive for malignancy.
Combined Techniques
Table 4 shows that combining brushing and biopsy sampling increases the sensitivity of biliary sampling for cancer diagnosis. The difference, in each particular study, between the less sensitive technique and the combination of both techniques ranged between 8% and 34%. This suggests that the increase in sensitivity is generally worth the hassle of combining both sampling procedures. Kitajima and colleagues have measured that the mean time required for brushing and for biopsy sampling was 2.4 and 8.5 minutes, respectively (this corresponded to 20% of the total procedure duration, and it is short compared with the additional time that would be required for endoscopic ultrasonography [EUS]-guided sampling).
| First Author, Year | n a | Brushing | Biopsy | Combination |
|---|---|---|---|---|
| Ponchon et al, 1995 | 204 | 33% | 26% | 35% |
| Pugliese et al, 1995 | 52 | 53% | 53% | 61% |
| Howell et al, 1996 | 28 | 58% | 31% | 65% |
| Sugiyama et al, 1996 | 43 | 48% | 81% | 81% |
| Jailwala et al, 2000 | 133 | 26% | 37% | 48% |
| Kitajima et al, 2007 | 60 | 72% | 65% | 74% |
| Weber et al, 2008 | 58 | 41% | 53% | 60% |
a Patients who had attempted biliary sampling with both methods were included; sensitivity was calculated following “intention-to-diagnose” principles if details sufficient for calculation were provided in the original articles.
Of note, the yield of biopsy sampling was rarely reported following the intention-to-diagnose principle in the original studies. Therefore, the sensitivity of biopsy sampling may have been overestimated compared with that of brushing because biopsy sampling is less often successful than brushing. For example, Ponchon and colleagues reported a sensitivity of biopsy sampling for cancer diagnosis of 43% (35 of 82 patients with a malignant diagnosis); this corresponds to a sensitivity of 26% in intention-to-diagnose because biopsy sampling was attempted in 134 patients with a malignant diagnosis. In the same study, brushing was successful in most attempted cases. Harewood and colleagues reported successful biliary biopsy sampling in 39% of patients in whom biliary brushing had been obtained. On the other hand, Sugiyama and colleagues reported successful biliary biopsy sampling and brushing in 87% and 88% of 52 patients, respectively. In some studies, the inclusion of patients with ampullary cancer also favored biopsy sampling in the comparisons listed in Table 4 .
Three of the studies listed in Table 4 used the Howell biliary introducer for at least one of the samplings ( Fig. 6 ). Because this device may be difficult to introduce into the desired duct due to its large diameter and rigidity, some investigators use it mostly when the insertion of standard biopsy forceps is challenging.
Bile Collection
Bile or pancreatic juice may be collected for cytopathological examination during ERCP (usually by aspiration upstream from a stricture) or via a nasobiliary or a nasopancreatic drain. A retrospective study compared the diagnostic yields of these techniques with other sampling methods in 214 patients who had a final diagnosis of malignancy. The sensitivity for cancer diagnosis was 48% for brush cytology, 41% for biopsy forceps, 30% for aspirated bile cytology, and 24% for nasobiliary drain cytology (the latter figure dropped to 14% if the result of the first nasobiliary drain cytology only was taken into account, as bile sampling was repeated up to 4 times). In patients who had both brushing and biopsy samplings, bile collected through the nasobiliary drain yielded additional information in very few cases. Therefore, the investigators recommended this sampling technique only in patients with a suspected cholangiocarcinoma and neither brushing nor biopsy sampling at the time of ERCP. Nevertheless, bile is an important fluid for the discovery of new markers of biliopancreatic cancer.
Sample procurement and preparation
Sampling for Cytopathological Examination
Material
Brushes are the most commonly used devices to sample the biliopancreatic ducts; they provide material for cytopathological examination. Models with a double-lumen catheter should be preferred for an easy, over-the-wire, manipulation ( Fig. 1 ). The flagella at the tip of some brush models is useful in angulated ducts but it may hinder brush movements inside the intrahepatic ducts. Single-lumen cytology brushes are rarely used nowadays; they should not be used over the wire because retrieving the brush through the whole length of the catheter (to replace it with a guide wire) causes a substantial loss of cellular material on the inner side of the catheter.
Related posts:
Therapeutic and Advanced ERCP Is Rapidly Progressing
Endoscopic Retrograde Cholangiopancreatography for Stone Burden in the Bile and Pancreatic Ducts
Endoscopic Retrograde Cholangiopancreatography for Distal Malignant Biliary Stricture
Endoscopic Ultrasonography-Guided Endoscopic Retrograde Cholangiopancreatography
Complications of Endoscopic Retrograde Cholangiopancreatography
Endoscopic Retrograde Cholangiopancreatography
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
