Associate Professor of Medicine at the University of Chicago Medicine, Dr. Uzma Siddiqui discusses the clinical history and diagnoses of incidental lesions.
[MUSIC PLAYING] UZMA D. SIDDIQUI, M.D.: So I usually go through a checklist in my mind what I'm looking at these lesions and trying to decide how I'm going to manage them. And the first thing is, what's the clinical history in this patient? Was this just an incidental lesion that was found on an endoscopy for some other reason, or is this patient having symptoms such as GI bleeding or obstructive symptoms? Is the patient 100 years old, multiple comorbidities, and does it really matter if they have a duplication cyst in the esophagus? So you definitely need to consider your patient when you're evaluating the lesion. What did the EGD show? What did the mucosa show? Obviously, if it's subepithelial, the mucosa should look normal overlying it. What is the location of the lesion, because that's also going to guide your differential. Once you decide to do the EUS, then a couple things. Is it actually inside the wall, or is it extramural, outside the wall, and is some other organ or tumor compressing the GI tract? If it is intramural, then what wall layer does it involve? And then what is the architecture and appearance of the lesion? And then we'll talk a little bit about different ways to sample the tissue, and then finally I'll touch a little bit briefly on management. And that's based on how much do we need to know what the diagnosis is and to confirm it, and then what is the potential for malignancy and having complications down the road. So when you think about subepithelial lesions, you can divide them, just in the beginning, as extramural or intramural lesions. And then extramural, obviously, outside the GI tract wall, can be normal adjacent organs, adjacent tumors. Sometimes then the liver can compress into the stomach. Pseudocysts sometimes do that as well. And then once you're talking about lesions inside the wall, you have numerous benign lesions that, again, if you can determine with some level of accuracy that it's a benign duplication cyst or lipoma, you don't have to do anything about those lesions. Nor do you have to suggest any surveillance or followup. And then obviously there are those that have the possibility of malignancy. Carcinoids, granular cell tumors, and GIST. And those are the ones that you ideally would like to try to get a firm diagnosis on, because then depending on what you find or diagnose, you're going to have to decide what the management of the patient's going to be, whether it's surveillance and endoscopic resection or surgery. So this is just a couple examples of how adjacent organs-- in this case you have a bulge in the stomach, it's just normal liver. Bulge in the proximal stomach is just the normal heart. So sometimes these can appear as subepithelial lesions. And I'm sure you've already heard this multiple times throughout the day, but when you have decided the lesion is intramural and you want to look at which wall layer it arises from, you have to just consider in general these four layers, which is simplified from a more detailed view of the wall. But you have a hyperechoic or bright superficial mucosa, a dark muscle layer or muscularis mucosa, and then another bright submucosa layer, hyperechoic. And then the last muscle layer is dark, hypoechoic, in the muscularis propria. So again, once you decide what layer the lesion comes off of, then that's going to help you decide on your differential. The majority of these lesions do arise from the submucosal layer. Some of them can also involve some of the more superficial layers. And then stromal tumors, or muscle cell tumors, by definition have to arise from the muscle layers. The muscularis mucosa rarely, and more commonly from the fourth layer of muscularis propria. But again, knowing what layer the lesion comes off of is going to help you determine what lesion you're dealing with. And then the echogenicity-- again, these terms correspond to their comparison to those wall layers I just showed you. If something's hyperechoic, it's going to be bright white, essentially, like fat in a lipoma, typically. If it's hypoechoic, this usually represents solid tumors. This is an example of a stromal tumor. If something's anechoic, where you see essentially a black structure, those are usually fluid-filled structures or cysts or vessels, the gallbladder. And so again, this is going to help you with the differential. And then sometimes lesions are a little bit of a mixed pattern, where they appear a little bit bright and dark, so those would be heterogeneous. So what is the diagnostic yield of subepithelial lesions? Well, depending on the study you look at, it can be horrendous or it can be fantastic. And I think the problem with a lot of the studies is that you're dealing with a wide variety of lesions that may be solid, may be cystic. And that can definitely give you a huge range of yield. Obviously these lesions are below the mucosa, so using standard forceps typically gives you a pretty poor yield, less than 40%. You can try to increase that a little bit by using larger-capacity forceps, biopsies, but still not great at around 60%. If you try to use EUS-FNA, again, the yield is all over the place. And typically, you're not going to do that well with small lesions, but if you have a larger lesions that are over two to three centimeters, solid lesions, and have on-site cytopathology, then that's where you're going to start seeing the higher yield. And you can't underestimate the assistance you get from having an on-site cytopathologist. But again, in these lesions, if it's a small one we usually never call them. A lot of studies have looked at trying to essentially take off that top mucosal layer or unroof the lesion, and then biopsy the lesion once it's exposed. So you can use a forceps to unroof the tissue. And this study in 2011 compared the diagnostic yield of EUS-FNA to using jumbo forceps to unroof tissue, and they actually showed they had over 90% yield using that technique with the jumbo forceps to unroof, compared only 67% with FNA. So in that case, you've got more tissue with the forceps, but only after you unroof the lesion. And then more recently this year Ken Binmoeller out of California described a technique where you suction a lesion into a cap. You place an endoloop to ligate it, and then you unroof the lesion by cutting through it with a needle knife. And he had 24 subepithelial lesions that were less than two centimeters, and using this technique-- they called it the suck-ligate-unroof technique-- and they had 100% diagnostic yield. And then they found, when they went to scope these patients a few months later, that in 13 out of the 24 lesions, they had actually already fallen off. So they were advocating this technique not only as gathering tissue and getting a diagnosis, but also as a treatment strategy for these lesions. But then that brings up the question, should you just, if possible, if it's a mucosal or submucosal lesion that's essentially I would say just slightly into the submucosa, should you just try to snare the lesion or perform endoscopic mucosal resection? And if you look at the data, definitely the more tissue you get, the more likely you are going to be able to get a diagnosis. And then just to touch on the management and diagnosis aspect a little bit, the question then is, if your diagnosis is in doubt, should you just EMR the lesion out? I didn't mean it to rhyme, but you know, it works. And the question becomes, if you do have these small lesions, sometimes it's difficult to tell what it is on EUS-FNA. EUS-FNA is probably useless. Should you just bite the bullet and just take the lesion out right then and there, so that not only will you have a diagnosis, but you'll have essentially treated the lesion. But that's not set in stone. There's no definitive guidelines that say you should do that. And obviously I would only recommend-- I tend to do this, but with smaller lesions, less than two centimeters. Because once you start dealing with larger lesions, you're definitely increasing your risk for perforation, bleeding. And you don't want to have those complications for something that essentially probably was benign and wouldn't have caused the patient any consequences. So if we talk about benign submucosal subepithelial lesions, duplication cysts are something we probably see quite often in our referral practice. They're just benign anatomic anomalies arising during embryonic development. Typically they occur in the distal esophagus. Here you see there's normal overlying mucosa. And again, being in the distal esophagus, that's a tipoff. Sometimes these are found incidentally on CAT scans that are done and can often appear as though it's a solid lesion on a CAT scan, but when you do the EUS, you can tell that it's anechoic and you have the differential of a cyst at that point. It doesn't have Doppler flow, so again, it's not a vessel. And in terms of wall layer, a duplication cyst can either have their own completely extramural wall layer structure, or if they are intramural, then they are arising from the submucosal layer. So again your location and the appearance of a cystic lesion in the distal esophagus are the tipoff that it's a duplication cyst. I typically do not FNA those lesions, mainly because there is a high risk of infection, and if you are going to do that then the patient definitely should receive some antibiotics. Lipomas, like everybody knows, are just benign fatty deposits. In fact, on endoscopy you usually can pretty definitively diagnose them if they're large enough. They're very soft when you palpate them with biopsy forceps. They have a pillow sign with a little indentation. Most commonly we see them in the distal stomach, the duodenum. They arise in the submucosal layer, and they are bright white or hyperechoic. So this is pretty classic for a lipoma. And you could try to biopsy it, unroof it and sample the tissue, and sometimes you can even extrude the fat. But in this one, it has such a classic appearance, most people hopefully would be able to trust your endosonographic diagnosis. And again, you don't have to do anything about it. Pancreatic rest, ectopic pancreas tissue, usually arise in the antrum, have a very classic endoscopic appearance. Again, you have this central umbilication. And on EUS, again in the submucosa, and they have more of a heterogeneous appearance. Sometimes you can actually make out ductile structures within them, but more than likely, this endoscopic appearance should probably be a tipoff to let you know what it is. And again, you don't have to do anything about these. Now, once we move on to lesions that are malignant or have potential for malignancy, that's when we start talking about possible surveillance or resection. Granular cell tumors are neural or Schwann cell origin. Typically, again, these occur in the esophagus. They're very firm white nodules. On EUS, if you're lucky, you can see them as a small submucosal lesion. Typically they're hypoechoic, and they can also be a little bit more superficial into the mucosal layers. And often you will get a diagnosis on biopsy, and then the diagnosis is clenched with immunostaining S100. The majority of these lesions are going to be benign, but if they become quite sizable, over four centimeters, then there have been reports of malignancy and sometimes perhaps GI bleeding or luminal obstruction. So in terms of management for these lesions, you could survey the lesion. And I had one patient, this one, who had an EGD and an EUS probably every year for five, six years. Or you could just resect the lesion, have your diagnosis, and now it's done, so the patient doesn't need to be surveyed. So typically I will EMR these lesions if they're less than two centimeters. Once they get bigger, you could consider doing ESD if you wanted to remove it. Carcinoid tumors, they're endocrine-cell origin with malignant potential. The most common occurrence for them is in the GI tract, and then once they're in the GI tract 20% of the time they occur in the rectum and then after that more commonly in the stomach. Here's an example of a rectal carcinoid. On EUS it involves-- oftentimes the mucosal layer is extending into the submucosal layer and appears as a hypoechoic lesion. And again, when they're smaller lesions, I tend to just do an EMR to essentially cure the patient. In the stomach, here's an example of another similarly hypoechoic lesion in the submucosa. Now in the stomach, I just want to bring up the fact that when you're talking about carcinoids, there's definitely three types of carcinoids. And oftentimes I think this is overlooked. And when the patient's referred to me with supposedly this one gastric nodule, they were told they have cancer, and they're very high-strung when they come to see you. And then when you scope them, you find out that actually in fact it looks like they have atrophic mucosa consistent with possibly atrophic gastritis. They have multiple small nodules throughout their stomach. And if you were to check agastrin, it would be quite high. And these patients have what's called type one gastric carcinonoids. They're the most common. They're multifocal, and they have a very indolent course. So typically, for these patients, if the nodules become a little bit more sizable, at around one to two centimeters, we'll try to EMR them out. I don't really bother with the smaller than one centimeter lesions. And then I do tend to survey them, probably once a year, just to ensure I remove any larger lesions. Type two gastric carcinoids are rare, but they're associated with ZE syndrome. But again, similar indolent course. And then the one you have to really worry about is type three. Those patients do not have atrophic gastritis, they have normal gastrin levels, and those patients tend to have solitary lesions. And those are treated essentially like a gastric cancer because they have a very aggressive course. The majority of patients tend to have local or liver mets, and those require a consultation with your surgeon for possible gastrectomy with lymph node sampling. So I think it's really important to know what type of carcinoid you're dealing with. And again, the vast majority are going to be type one. So then moving on to stromal lesions, the two layers that are involved always are your muscle cell layers and most commonly the muscularis propria. Leiomyomas are benign stromal tumors. You don't have to necessarily follow those up, if you can get a definitive diagnosis. And most commonly they occur in the esophagus. This happened to be a patient I scoped last month who had one in the cardia, though. Again, it's a hypoechoic mass lesion that arises off the muscularis or dark muscularis propria layer. So we always-- I tend to do an FNA. I call the on-site cytopathologist. They can't, obviously, do staining in the room, but they'll tell me they have spindle cells, and then they can run the immunostaining back in the lab. And for leiomyomas, those tend to stain positive for alpha smooth muscle actin and desmin. Now GIST are the ones we always worry about. Those are malignant stromal tumors. Again, they look very similar, in fact identical almost, to leiomyomas. Some people claim they can tell the difference between the two on EUS, looking for things like halos or differences perhaps in echogenicity. But I can usually never tell a difference, and I just wait for the pathologist to stain it and let me know which one it is. But GIST occur in the stomach most commonly. Again, a hypoechoic lesion from the muscularis propria. And very rarely you'll get it off the muscularis mucosa, in which case you could possibly resect it. And then again, on-site cytology will give you spindle cells. But the diagnosis is made with the immunohistochemical staining. GIST typically are positive for c-KIT or CD117 stains. More recently, there's been this DOG1, or essentially it's called discovered-on-GIST stain, that was found to be expressed in a small proportion of c-KIT negative GIST. So my pathologists, again, when we send them a possible stromal tumor, they send all these markers at once because sometimes there's a little bit of overlap and it's harder to send one stain later. So they'll do c-KIT, the CD117, DOG1, anti-smooth-muscle actin, and the desmin, and then come up with the diagnosis from there. So all GIST do have malignant potential. But there are some high-risk factors that we look for on EUS to predict a more aggressive course. A larger size, over three centimeters, and a regular border if it's heterogeneous, has echogenic foci, cystic spaces, or tumor necrosis. So these are all what we call high-risk features. More recently, the National Comprehensive Cancer Network in 2010, they came up with guidelines on GIST management. Any GIST that's over two centimeters, they recommend that they be resected. And then for lesions that are less than that and have no high-risk features, they suggest, but again they're not definitive, that you should do an EUS every six to 12 months. So that's typically what we end up doing if they're not resected. Probably once a year we'll do an EGD-EUS. Now is there a role for endoscopic therapy? If the lesion arises from the second layer of the muscularis mucosa, then definitely you could do an EMR and get rid of the lesion. If the lesion's off the muscularis propria, obviously you can't remove that endoscopically without causing a perforation. But there is a technique called loop and let go that you can try, on lesions less than two centimeters, where essentially-- in this case this was a small GIST in the stomach-- you strangle the lesion with an endoloop and then come back in a few weeks and the lesion has dropped off. The main risk is that you get significant ulceration. And I will say that very few lesions actually are amenable to this technique, because they have to be somewhat raised and able to be captured with that endoloop, and you have to apply so much pressure that you start strangling it. So it's something you can think of, but it doesn't work in all cases. So just in conclusion, when you're evaluating subepithelial lesions in the GI tract, just a few considerations. Is it actually inside the wall or outside of it? Are you going to be dealing with a benign or a potentially malignant lesion? EUS is definitely, I think, very helpful just to determine what wall layer it's coming off of, what does it look like sonographically. It helps you come up with the differential. If it's a larger lesion, you have on-site cytopathology, then perhaps FNA would be useful. Your diagnostic yield is definitely going to depend on the lesion type and what technique you use. If you can or feel comfortable EMRing a lesion that's more superficial, perhaps that's going to give you the most yield and also help treat the lesion. And then management obviously depends on the etiology. And then, obviously, once you start looking at these lesions for a while and your EUS experience grows, I think you come up with your own algorithm on how you treat them. And then just speaking of EUS experiences, I just wanted to share one with you guys. Back in 2006, when I was a brand-new first-year attending, I had the pleasure of spending the summer with the world-famous and talented endoscopists at the Academic Medical Center. They probably don't even remember this. And two of them are here today, Drs. Marco Bruno and Paul [INAUDIBLE]. And you know, when you're a new attending, you start thinking how you're going to plan your career. And a couple things stood out to me that summer. And two goals that I came up with were, one, I wanted to host a really nice EUS course. And the second one was, if I worked really hard, one day I would hopefully get promoted to full professor like Dr. [INAUDIBLE] did that summer, and I could get to have a '70s-themed party to celebrate in Amsterdam. So that's the goal. All right, with that I'm going to stop. Thank you.
