on the ASGE SuTab tip of the week. Today's tip is about looking for NICE 3 changes in lateral spreading lesions and how to use that information. But first we'll worry about the NICE classification, the MBI, International Colorectal Endoscopic Classification, which is developed for use with Olympus MBI, but the principles here can be applied to really any type of scope and other systems also have analogous findings. For example, KUDO 5 is basically the same as type 3. So the NICE classification type 1 and type 2, we hear a lot about in the resect and discard controversy. Type 1 are the sessile serrated lesions, also called sessile serrated adenomas or sessile serrated polyps and the hyperplastic polyps. So it's the entire serrated class of lesions except for traditional serrated adenomas. The NICE classification makes no attempt to subdivide hyperplastic versus SSL. There is a classification, the WASP classification that does that and it works pretty well for larger serrated class lesions. Type 2 lesions are conventional adenomas and they differ from type 1 in that they have thick brown vessels in them and then they have white pits that are often variable in shape, oval, tubular, or branched. And there is another classification which subdivides type 2 into 2A and 2B. That's the JNIC classification. Type 2B has some irregularity in it and is associated with hygrid dysplasia and superficial submucosal cancer. Superficial means that there's less than 1,000 microns of invasion. Remember that the clinical definition of colorectal cancer is invasion into the submucosa. And then there's type 3, which is basically where there is disruption of the vascular pattern seen in type 2 and that indicates deep submucosal invasive cancer with high specificity, not perfect sensitivity. You can't always find those changes in patients or lesions that have deep submucosal invasive cancer. But the value of identifying type 3 is that when it's present, it has traditionally been considered a contraindication to any form of endoscopic resection because deep submucosal invasion has been considered an independent predictor of lymph node metastasis. So surgery is the best approach. Now, I will say there's some controversy about that now as to whether if that's the only adverse histologic feature that's present, whether there really is a high risk of lymph node metastasis but that has been the traditional approach. And I will say that I still generally practice that way. If I find deep submucosal invasion, I'm going to send the patient to surgery. And that's the importance of this because I will say that the place that I see this is typically in referred polyps that have been biopsied by the referring physician and no cancer is evident. And the issue is that the wrong part of a lesion has been biopsied. So bottom line, what's the importance of NICE 3? It's that NICE 3 predicts deep submucosal invasive cancer. And that is a contraindication to endoscopic therapy traditionally. So when we encounter a lateral spreading lesion, we want to look over the entire surface of it looking for NICE 3 changes because if they're present, we shouldn't remove the lesion endoscopically. And secondly, to demonstrate cancer so that we're comfortable sending the patient to a surgeon and the surgeon's comfortable in performing a surgical resection because they know cancer is present, we have to take the biopsy specifically from that location where the NICE 3 changes are seen. Here's the classic NICE 2 pattern. In the KUDO system, this would be 3L and 4. The brown lines are blood vessels, the white squiggly tubular lines, sometimes branching, those are presumed to be the pits of the KUDO system. So this is the conventional adenoma. And when we're looking for NICE 3, we're basically looking for disruption of this vascular and surface pit pattern. The first point I want to make is about looking for these changes in adenomas versus serrated lesions. Now, we all know that when you look at so-called interval cancers that arise after someone's had colonoscopy, that they are more likely to have passed through the serrated pathway. And that's often interpreted as meaning that serrated lesions are more dangerous than conventional adenomas. And on an individual lesion basis, I don't think there's any evidence that that's true. And in fact, just the opposite. If you look at the risk of cancer in a conventional adenoma compared to a serrated lesion of similar size, it's more than seven times higher in the conventional adenomas. And talking to experts, myself included, it's very hard to find videos of serrated lesions where you can still see some serrated lesion and see cancer in it. Whereas everybody has tons of adenomas that have a focus of cancer in them endoscopically. And serrated lesions in general, it's very easy to inspect the surface of them because they're relatively flat. Adenomas tend to have sometimes a much more irregular surface. The size of them may be very large. They may have much more of a bulky component. And so they're harder to inspect. But when we are looking at the residual adenoma and saying that there's disruption of the vascular pattern, that applies mostly to adenomas because adenomas have nice too. They have a vascular pattern that can look disrupted. Serrated lesions don't have that kind of vascularity to look for that transformation from vascular pattern intact to vascular pattern disrupted. So going back to the issue of over-representation of serrated lesions in interval cancers, why is that? Well, it's not because the individual lesions are more dangerous, it's because they're more likely to be missed. Here's an example of a small cancer arising in an SSL. And I mentioned I only have a handful of these videos, have many videos of cancers and NIS3 arising in conventional adenomas, but these are quite uncommon. We all see a lot of cancers arising through the serrated pathway that are recognized for molecular features like BRAF mutation or epigenetic inactivation of MLH1. But seeing a small cancer where they're still intact SSL is uncommon. Now within the area of cancer, yes, you will see NIS3 changes there. You can see that there's no real vascular pattern. I'm just saying that when you look at the benign part of the lesion, you're not gonna see a transformation of intact vascular pattern, like you will in a conventional adenoma where you go from NIS2 to NIS3 because the intact part of the polyp without cancer is NIS1, as you can see right in front of us on the left side of the lesion. So here's the left side of the same lesion with a yellow line drawn around the intact SSL. And you can see the uniform size dark pits, but there are no blood vessels there. This is the NIS type one pattern. Now, if we look within the ulcer to the right where the cancer is, yes, there are NIS3 changes in the sense that there is a completely amorphous, irregular vascular pattern, but you're not gonna see the transition from intact NIS2 pattern to NIS3 in this particular lesion because the part of the lesion that is still benign is NIS1, it doesn't have NIS2. So it's a little bit different in that regard. And then I think the other feature again is that seeing cancer in lesions where you can still see some residual benign SSL is extremely uncommon. So although we're gonna look for NIS3 changes, morphologic changes, and completely disrupted pit pattern and blood vessel pattern in every lesion, for the most part, we're gonna find it in conventional adenomas. So here is a lesion that's been referred for endoscopic resection. It's been tattooed, it's been biopsied. The biopsy showed high-grade dysplasia. It's about 50% of the circumference. It looks like a non-granular lateral spreading lesion, and there's a clear morphologic change suggestive of cancer. It's basically ulcerated over about 50% of the lesion. We've got the near focus on now, and in a moment, we're gonna go to NBI. But you can tell that there is intact vasculature and pit pattern around the edges of the lesion as well as the flat area off to the right. We're not looking at that area right now, but down here in the base of the lesion, the vascular pattern is amorphous and disrupted, and that is NIS3. Now you can also see that in NBI, the color of that area is different compared to the surrounding area. It is whiter, it's lighter in color, and that also is a typical change of NIS3. So this lesion, which I would say just morphologically is an overt cancer, also has NIS3 changes in the base of it, and that's the area where we need to focus our biopsies. All the other rules that we've talked about apply to it, should be the last thing that we do in the procedure. I like to take multiple biopsies right from the base of that area where the ulceration is. I usually will put a trap on the scope so that anything that comes off and in the channel can be sucked back into the trap, and we can add that to the specimen container. We wanna get enough tissue here to give the pathologist a good chance to make the diagnosis, but we wanna get that tissue right out of that area where the NIS3 changes are present. Here's another lesion, this one's in the rectum. You can see in NBI that the blood vessel pattern is mostly NIS2, it's intact. Until we get over to this proximal edge, and now there's an area that is whiter in color, and the blood vessels are more irregular. The pattern is amorphous. In some small areas, there are very few blood vessels. That's what gives it the whitish color compared to the rest of the lesion, so this is NIS3. This is another lesion that has been biopsied, just showed benign tubular adenoma. But again, we've got to inspect the entire surface of it because these features are predictive of deep submucosal invasion. We've gotta target biopsies to that area. Here's this area of NIS3 in near focus. This is where biopsies will show cancer. This is a granular lesion in the rectum, and I mentioned that the more distal you are in the colon, the greater the risk of cancer in individual lesions. This is a granular mixed lesion because it has a large nodule in it, and we see this nodule off to the right of the image, and the color of the nodular part is whiter in NBI than the rest of the lesion, the flatter part of the lesion. Now we're looking at that flat part of the lesion where the NIS2 pattern is still intact, but as we look at this discolored area in the nodule, we can see that there is disruption of the vascular pattern. There you can see the white color again, the very amorphous fewer vessels, and the vessels are quite irregular. This is deep submucosal invasion in the nodule of a granular mixed lesion, a large lesion in the rectum. That's typically, if there's cancer present, it'll be present in that nodule. General, the sensitivity of NIS3 changes for cancer is about 50%, but it has very high specificity. When you see it, there is a very high probability of cancer. The morphology of the lesion does affect the sensitivity, so you do much better predicting cancer using these features in flat lesions. Bulky lesions are much harder to predict. The surface features may not predict what's going on deeper in the lesion. Also important to know for predicting the risk of cancer that location of the colon matters. It's higher for individual lesions as we move distally, especially in the rectosigmoid, and then morphology matters. It's highest in non-granular lesions, especially if they have pseudodepression, and it's lowest in granular lesions, particularly those that are homogeneous. It's higher if a nodule is present. In the absence of NIS3 features, there is the possibility of endoscopic resection, but we still should consider features such as location in the colon and morphology because they could affect the approach to endoscopic resection. This gets at the issue of ESD versus EMR or on-block resection versus piecemeal resection because on-block resection allows us to get an accurate measurement of the depth of submucosal invasion if cancer is present, and if it's superficial, less than 1,000 microns, no other adverse histologic features, we can avoid a surgery, and that's particularly advantageous in the rectum because the morbidity of surgical procedures in the rectum is higher. So we're not only considering whether or not NIS3 features are present in endoscopic versus surgical resection, but are there other features once we've seen that NIS3 is absent that could affect the approach to endoscopic resection? Next week, more about NIS3 on the ASGE SuTab tip of the week.

NICE classification

deep submucosal invasive cancer

lateral spreading lesions

sessile serrated lesions

conventional adenomas