ERCP and EUS Procedures: The Obstructed and Indeterminate Biliary Tree

0:00Welcome to Board Pearls. This is episode three of seven of the ERCP and EUS Procedures chapter, in the Endoscopic Procedures module. In this episode we cover the obstructed and indeterminate biliary tree: getting a diagnosis from a stricture, choosing and placing the right stent for a malignant obstruction, and resecting an ampullary lesion rather than stenting around it.

0:22Start with the stricture you cannot name. A patient has painless jaundice, a tight narrowing in the bile duct, and imaging that says cancer without proving it. The whole episode turns on two related problems. How do you get a diagnosis from a lesion you can barely sample, and once you have it, how do you achieve drainage that lasts. And running underneath both is a third question that keeps flipping the answer. When is the right move not a stent at all, but resection, or a referral for transplant.

0:50Take tissue acquisition first, because almost everything else depends on it. The standard first move at ERCP is brush cytology. You pass a brush across the stricture, drag it back and forth, and send the cells. It is easy, it adds little time, and its sensitivity for malignancy is low, missing most cancers. That failure is the whole teaching, so understand where it comes from. A brush samples superficial cellular shedding. It scrapes the surface and collects whatever has sloughed into the lumen. But cholangiocarcinoma is characteristically submucosal and infiltrative. It grows in the wall, beneath the lining, and it sheds sparingly into the duct. So the brush is sampling the one place the tumor is least likely to be. A negative brush in a stricture that looks malignant tells you almost nothing.

1:38Once you see that the problem is sampling depth, the entire diagnostic cascade becomes a single idea. Every step that raises sensitivity does so by getting closer to the tumor or by adding a signal the brush cannot capture. Add intraductal forceps biopsy and you capture deeper tissue, with architecture rather than loose cells. Add FISH, fluorescence in situ hybridization, looking for polysomy. Now you detect aneuploidy, cells with multiple chromosomal gains that mark malignancy even when the cytology looks bland. Stack those onto the brush and the yield climbs substantially. You have added architectural and molecular signal to a sample that was only cellular.

2:18Then go further still. Cholangioscopy with SpyGlass-directed biopsy pushes the yield higher again, and the reason is depth and aim. Instead of blind sampling, you thread a small scope into the duct and look directly at the stricture. You see the mucosa. You can target the biopsy at the part that looks wrong, an irregular surface, tortuous neovascular tumor vessels, friable tissue that bleeds when touched. Those visual features raise your confidence before any histology returns. There are formal visual classification systems for these cholangioscopic patterns, the Mendoza classification among them. The point for boards is the principle, not the categories. Direct vision plus a targeted bite beats a blind brush every time.

3:05The same logic explains the best comparison in this whole section. For a pancreatic head mass, EUS with a fine-needle biopsy is far more sensitive than ERCP brushings of the same patient's duct. Why so much better. The brush samples the duct surface, while EUS samples the mass itself, seeing the lesion directly through the duodenal wall and reaching into the parenchyma where the tumor actually lives. So for a pancreatic head mass that needs tissue before neoadjuvant chemotherapy, the answer is EUS-guided biopsy of the mass, not brushings. And you favor fine-needle biopsy over fine-needle aspiration when you need core architecture and immunohistochemistry, which follows directly from getting closer to the lesion.

3:50Now the exception that sits on top of the entire hierarchy, because this is where careful reasoning leads a good clinician to the wrong answer. A patient has a hilar mass, bilateral intrahepatic dilation, and is being worked up for liver transplant under the Mayo protocol. You have just internalized that EUS gets you the best tissue, so the reflex is to put a needle into that hilar lesion under EUS. Do not. In a transplant candidate, EUS fine needle aspiration of the primary hilar tumor is contraindicated. The needle tract crosses the peritoneal cavity, and it can drag tumor cells along that path and seed the peritoneum. The Mayo protocol requires absence of extrahepatic disease. Peritoneal seeding is extrahepatic disease, so the needle you used to diagnose the cancer is the needle that disqualifies the patient from the cure. The same reasoning bars transabdominal percutaneous biopsy of the primary. You do not poke the primary hilar lesion from outside the bile duct in someone who might be transplanted.

4:51So how do you confirm the diagnosis without ending the transplant. You stay inside the biliary tree. Brush cytology and intraductal biopsy at ERCP sample the primary without ever entering the peritoneum, which is exactly why those lower-sensitivity tools are the preferred approach here. The hierarchy inverts because the goal changed. And there is one more move that feels paradoxical until you see the logic. Suspicious regional lymph nodes ARE biopsied, even by EUS. Positive nodes mean the patient is already extrahepatic, already off the transplant list regardless of how you found out. Confirming nodal disease does not create disease that was not there. So the node biopsy is not really a diagnostic test for the tumor. It is a transplant-eligibility test, and that is how you frame it to the patient. Sample the node to rule the patient in or out. Never sample the primary, because that bite is the one that can iatrogenically create the very disease that disqualifies them.

5:48Hold that idea, because it returns in stenting. Now move to the second half of the diagnosis problem, the ampullary lesion, where the whole point is to resect rather than stent around it.

6:00An ampullary adenoma sits at the major papilla. It can be sporadic and found incidentally, or it can be part of duodenal polyposis in familial adenomatous polyposis, where the ampulla is the single most common site of duodenal adenoma. Either way it follows the adenoma-to-carcinoma sequence, so it has to come out. The surgical answer for an ampullary lesion is a Whipple, a pancreaticoduodenectomy, which is major morbidity. The entire case for endoscopic ampullectomy is one trade. You accept a procedure that causes pancreatitis in roughly ten to fifteen percent of cases, mostly mild and manageable, in order to spare the patient a Whipple. Every technical choice in ampullectomy is in service of protecting that trade.

6:45That trade only holds if the lesion is actually resectable through the scope, which means no intraductal extension and no invasive cancer. So before you resect, you image with EUS or intraductal ultrasound. The question EUS answers is whether the adenoma has crept up inside the bile duct or pancreatic duct beyond the papillary orifice. This matters because the snare can only grab what sits at the surface. It cannot reach tissue that has tracked up into a duct. If you snare a lesion with intraductal extension, you leave disease behind and recurrence follows. You have given the patient an endoscopic procedure that failed to do the one thing surgery would have done. Intraductal extension is therefore a contraindication to endoscopic ampullectomy and sends the patient to surgery. Cross-sectional imaging rounds out staging by excluding nodes and metastases.

7:37Pathology drives the same trade in a subtle way. You biopsy the lesion first, but biopsy understages roughly thirty percent of the time. Carcinoma turns up in the resected specimen that the pre-procedure biopsy called benign. So you cannot fully trust the biopsy, and that is the argument for en bloc resection. Capture the lesion in one piece, send the whole thing, and let the pathologist stage the entire specimen with intact margins. Piecemeal resection fragments the specimen, so it cannot stage reliably, and its outcomes are worse. The practical rule is en bloc snare resection for lesions up to about two to three centimeters, with piecemeal reserved for larger lesions when you simply cannot capture them whole.

8:19Two technical points read counterintuitively until you reason them out. First, you generally do not inject submucosal fluid at the ampulla, even though lifting a lesion is standard everywhere else in the gut. The duodenal submucosa is thin at the papilla, and injecting fluid pushes the lesion up and away from the snare, making it harder to capture rather than easier. So you skip the lift, drop the snare around the lesion, and tighten directly. Second, at the end of the case you place a small prophylactic pancreatic duct stent, typically a five French, across the pancreatic orifice. Resection traumatizes that orifice and edema can choke off pancreatic drainage. The stent holds the channel open through the swelling, which is the same mechanism as routine PD stenting for pancreatitis prophylaxis. This is the single most strongly supported step for preventing post-papillectomy pancreatitis, the dominant complication of the procedure. A short biliary stent is often placed alongside it to prevent obstruction from edema and to ease later access.

9:23After resection the trade has to be maintained over time, because local recurrence runs commonly around twenty to thirty percent overall, higher in FAP and higher after piecemeal resection. So you survey. Endoscopy at three and six months inspects the resection bed. Recurrence found early is usually amenable to repeat endoscopic resection, which preserves the whole point. The patient never gets a Whipple.

9:44FAP changes the surveillance picture, and it leans on one tool that is easy to get wrong on a stem. To see the ampulla en face you need a side-viewing duodenoscope. A standard forward-viewing scope looks straight ahead and slides right past the papilla on the side wall. So FAP duodenal surveillance, ampullary inspection in particular, uses the side-viewing scope. It runs every six to twelve months in patients with established duodenal polyposis or a prior ampullary lesion. You stage the overall duodenal burden with the Spigelman classification, which scores polyp size, number, histology, and dysplasia. When that reaches Spigelman stage four, with high-grade dysplasia or extensive ampullary involvement, endoscopy can no longer keep up with diffuse polyposis. That is the point where the answer becomes surgery, a pancreas-sparing duodenectomy or a pancreaticoduodenectomy, because the field disease is no longer durably controlled through the scope.

10:46Now to the third strand, durable drainage of a malignant obstruction, where the right stent depends on where the obstruction is. The organizing rule is simple to state and the exceptions are where the teaching lives. Stent type follows location and side-branch anatomy.

11:01Start distal. Pancreatic head adenocarcinoma, distal cholangiocarcinoma, or ampullary cancer obstructs the lower bile duct, and the patient is often heading for neoadjuvant chemotherapy before a Whipple. The right setup is a fully covered self-expanding metal stent. Reason from what threatens patency. An uncovered metal stent is a bare mesh, and tumor grows straight through the gaps and reoccludes it. A covered membrane seals the lumen so the tumor cannot grow in, which buys roughly four to six months of patency, enough to carry the patient through chemotherapy. And because it is covered, it is removable at the Whipple. A plastic stent, by contrast, gives only about three months and needs repeat ERCP to swap, every exchange delaying surgery and adding risk. So fully covered metal is the durable choice here precisely because the covered design defeats ingrowth.

11:54Two related points cluster here. Because you need tissue before chemotherapy and drainage for the cholestasis, you do both in one sitting. Same-session EUS fine needle biopsy of the mass plus ERCP with the stent, one anesthesia event, shortest time to chemotherapy. And there is a restraint worth knowing. In a clearly resectable tumor without cholangitis and without severe jaundice, you do not routinely drain before surgery just because the bilirubin is up. A randomized trial settled this: routine preoperative biliary drainage in resectable disease added complications without benefit. So you drain the patient with cholangitis, the patient whose jaundice is severe enough to interfere with chemotherapy, and the patient going to neoadjuvant therapy. You do not reflexively stent every resectable head mass just to see the bilirubin fall.

12:50Now the hilum, and here one of those design choices reverses, which is exactly the kind of flip a stem will probe. Unresectable hilar cholangiocarcinoma, perihilar tumors, or gallbladder cancer obstruct the confluence where the right and left ducts meet, producing bilateral intrahepatic dilation. Your distal instinct says reach for a covered stent to block ingrowth. At the hilum that instinct is wrong. The covered membrane that protected you distally now physically blocks the orifices of the contralateral side branches sitting right at the confluence. The very feature that prevents tumor ingrowth would occlude the ducts you are trying to drain. So at the hilum you use an uncovered metal stent, whose open mesh lets bile from the side branches flow through into the stent. Distal, covered to block ingrowth. Hilar, uncovered to preserve side-branch flow. Same disease family, opposite membrane, because the anatomy changed.

13:46The hilum carries a second rule that pairs with the stent choice, and it is a real threshold worth teaching with its mechanism. The goal is to drain at least fifty percent of viable liver. Below that, enough cholestatic liver remains undrained that the bilirubin will not normalize and jaundice persists despite a technically successful stent. Above it, jaundice resolves. So drainage is not about how many ducts you can cannulate. It is about how much functioning liver you actually decompress. That is why you map the anatomy with MRCP before the procedure. The MRCP shows which territories are viable and which are atrophic, and atrophic territories must not be stented. Puncturing a shrunken, non-functioning segment introduces bacteria and infection without draining anything that contributes to clearing the bilirubin. So you plan unilateral versus bilateral stenting around which lobes hold the viable mass. Drain both when both lobes carry similar viable volume. Drain one when the other is atrophic and the contralateral lobe holds the majority of working hepatocytes.

14:47One last calibration. Plastic stents in malignant obstruction are reserved for a life expectancy under three months. Their patency is too short to justify the cost and procedure when someone will live longer, but their lower deployment cost fits when survival is brief. So plastic is not the cheap default. It is the choice for the patient whose prognosis is measured in weeks.

15:10Step back and the whole episode is one decision tree run twice. First, get the diagnosis by getting closer to the lesion, brush to intraductal biopsy and FISH to cholangioscopy to EUS. The exception is the transplant candidate. There you protect candidacy by staying inside the duct and biopsying nodes rather than the hilar primary.

15:33Then achieve durable drainage by matching the stent to the obstruction. Covered metal distally to block ingrowth, uncovered metal at the hilum to preserve side branches, and always enough viable liver drained to clear the bilirubin. And throughout, the question that keeps rewriting the answer is whether a stent is even the goal. For the ampullary adenoma the answer is resect it, not stent around it. For the transplant-eligible hilar tumor the answer is do not touch the primary with a needle and protect the path to a cure. The thread tying it together is that location, resectability, and transplant eligibility decide the move long before the device does.

16:10Next episode picks up where standard ERCP fails. When you cannot cannulate the papilla or the surgical anatomy has rerouted the gut, the question becomes how to reach the bile duct another way, through EUS-guided rescue and through altered anatomy.

16:26For the full chapter, the practice vignettes, and the topic-tagged question bank, head to board pearls dot com. You'll find the rest of the series on Apple Podcasts, Spotify, or wherever you listen to podcasts. That brings us to the end of episode three of seven of chapter twenty nine, and I'll see you in the next one.