I have defined the Barrett’s pattern somewhat to that end, but here is my current definition of Barrett’s: A long-segment Barrett’s would be more than 3 cm of columnar epithelium in the esophagus with intestinal metaplasia. A short-segment Barrett’s would be when you can definitely see a segment less than 3 cm of red mucosa in the lower esophagus and again find intestinal metaplasia. The problem is the finding of intestinal metaplasia of the cardia. If you take biopsies on consecutive patients having endoscopy, as was shown by Speckler first, you will find that about 20% of everyone has intestinal metaplasia, even when you don’t see any Barrett’s endoscopically. If you take biopsies from there that show intestinal metaplasia and there is no visible Barrett’s, I would prefer to say that patient has intestinal metaplasia. If this affects one-fifth of the older population, the cancer risk for the individual has to be very low, and I am very doubtful that any surveillance is required or going to actually be helpful for a person with that situation.

Cancers can arise in short-segment Barrett’s. This was an example where, if you look carefully, you will see three red tongues – here, here and here; the middle one is easier to see. There is a 1.5-cm length of intestinal metaplasia and it is difficult to see, but there is a small adenocarcinoma here, so this was adenocarcinoma of the cardia, arising in a short-segment Barrett’s. So, what recommendations can be made?

Surveillance – these are current suggestions; they are arbitrary, but they are what we are doing, based on the best review of the literature and personal experience. If there is a long or a visible short Barrett’s which is seen on endoscopy and there is intestinal metaplasia on biopsy, then if there is no dysplasia, we have asked the patients to come back after one year, just in case we missed something the first time. We were unhappy about telling them to go away for a long time. If there is still no dysplasia after one year, we ask them to come back for further endoscopy and biopsy every two or three years. If your HMO is strapped for cash, realistically, if that was five years, it would probably be a reasonable cost benefit ratio.

If there is low-grade dysplasia, which we have found in about a third of our Barrett’s patients, we ask them to come back at six months and the if it is still low grade, come back every year.

If there is high-grade dysplasia we recommend surgical resection for the healthy patients who can stand it. If there is no surgery, endoscopic mucosal ablation is done, or we follow every three to six months.

Surveillance is not generally recommended if you have a biopsy showing intestinal metaplasia but you didn’t see any Barrett’s, or you thought there was a Barrett’s on endoscopy and you take enough biopsies, but the pathologist says there isn’t any intestinal metaplasia; it probably isn’t a Barrett’s and the patient would be unfit for surgery. We seem to see patients getting on our list sometimes who are 80 years old, have Alzheimer’s or congestive failure, and really, I don’t think there is any need to continue surveillance on those, because you are not going to do a major operation, whatever you find.

This is data from Brien, Reed and Levine, et al, in Seattle. There is relatively little data, but they did follow high-grade dysplasia for about two to three years. What happened was that patients with high-grade dysplasia initially, 26%, about one quarter went on to get a cancer within two or three years. About half, 47%, still had high-grade dysplasia on the last biopsy done and surprisingly, about a quarter again, 27%, had high-grade dysplasia when first diagnosed, but subsequent biopsies failed to show high-grade dysplasia. They had not done any active treatment; they had done many biopsies and may have removed some of the mucosa, but it is in no way inevitable that patients with high-grade dysplasia will go get cancer, at least in the next two or three years.

Our usual recommendation for treatment of high-grade dysplasia in patients who are otherwise healthy is to do an esophagectomy. This is the resection of the shaded area with the pulling up of the stomach to anastomose to the high thoracic esophagus with a pyloromyotomy, because you’ve cut the vagus nerve.

What do we see if we look microscopically at the resected specimen? Yellow is low grade and this one had four small areas of high-grade in red and a tiny cancer here. This patient had a 2.5-cm long Barrett’s with one tiny area of high-grade dysplasia here. This one had a little low grade, high grade and a very early cancer there. Now, some of the patients had larger areas of high-grade dysplasia. This is to show that you really need to take systematic biopsy samples of the whole esophagus when you are doing endoscopy if you are going to find that and even then, it was perhaps only luck that this was found.

Cancer treatment

What do you do if you find a Barrett’s? What I like to do is to look very carefully for any suspicious nodularity or stricture at endoscopy and to first biopsy any nodule or stricture. Benign ulcers usually are benign. Then, sample the whole of the Barrett’s esophagus by taking four-quadrant biopsies, as shown with these little x’s, all the way down the Barrett’s esophagus to try to find anything that might be hidden there.

What are we looking for? Well, besides cancer, we are looking for dysplasia, which is a precancerous condition. This is taken from a resection specimen, to get a nicer picture, but this is a Barrett’s, and if you look at this side of this villous-like structure, you will see the obvious goblet cells here. Here, the nuclei are at the base of the cells in order and that is non-dysplastic here. Now, when we look at the other side of this villous structure here, there are more nuclei. The nuclei are above other nuclei; that is, they are stratified. This is low-grade dysplasia; that is non-dysplastic Barrett’s, low-grade dysplasia. The same is true here; this is non-dysplastic here and this is low-grade again. There are too many nuclei here and the nuclei are stratified. This is common and we have found low-grade dysplasia in about a third of our patients and high-grade dysplasia in considerably less than 5%. Here is high-grade dysplasia. On this side you will see lots of nuclei; large, densely-staining nuclei, not much cytoplasm and the nuclei extend up to the lumen surface of the cells. On this side, it is even worse; there is very high-grade dysplasia with malignant-looking cells. These really are cancer cells to look at, but my pathologist colleague tells me that because they actually haven’t broken through the basal lamina, which is there and here, this is by definition high-grade dysplasia and not invasive adenocarcinoma; it is sometimes difficult to tell the difference on a biopsy.

If you were to operate on someone with high-grade dysplasia in different series, anywhere from about 10 to 50% of patients in whom the endoscopist has found Barrett’s high-grade dysplasia and not seen a cancer, cancer is found in 10% to 50% in the resected specimen, depending on how carefully it has been looked for. Some of these patients have a cancer right now; others will get a cancer later.

The same patient, approximately one year later, had another endoscopy and he had developed a Barrett’s esophagus in the meantime. Here on the right is the typical epithelium found in a Barrett’s esophagus – glandular; note the goblet cells. You can see the little sort of oval holes in the cells here. This is intestinal metaplasia; this is not the normal type of epithelium found at the upper end of the stomach . This is the hallmark of Barrett’s esophagus and the type of epithelium that most adenocarcinomas of the esophagus are found to occur in.

Here are some endoscopic pictures; we are looking down in about the mid esophagus here at an untreated patient with Barrett’s esophagus. You will see these erosions, these triangular, pale-centered, bright-red margined lesions in the squamous mucosa. This is typical reflux esophagitis. If you look further down here, however, you will see red mucosa which covers a long segment of the lower esophagus and that is the Barrett’s mucosa. Here, on the retroflexed view, looking upward, is the hiatal hernia that this patient had.

Supposing we treat a patient like this effectively to prevent reflux, either with proton pump inhibitors, such as omeprazole, or with a laparoscopic Nissen. What will happen? What will probably happen is that you will control their heartburn and acid reflux and their esophagitis will resolve if you endoscope them again, but the Barrett’s is almost certain to remain. It may develop squamous islands. These are squamous islands partially occupying some of the area of columnar mucosa. Doing an anti-reflux operation, or controlling with Prilosec, has not been shown to reduce the cancer risk in the remaining Barrett’s esophagus, so you still have to keep on with surveillance.
Breast cancer
What is the risk of adenocarcinoma in patients with Barrett’s? There are many series; these are just three of the larger, more recent series, number of patients followed, number of follow up years median and the number of cancers per patient years. Take the top series of 166 patients, followed for an average of 9.3 years, one cancer per 180 follow up years, or out of 180 patients, one would get cancer each year. That looks bad, but it isn’t quite as bad as you might think. Some patients we see with Barrett’s really feel sure that they are going to die of cancer, but that is just not the case.

Here are four series, one of which was a paper I had in the New England Journal in 1985; these are longer follow ups. These are patients followed up 3-cm Barrett’s or longer, the number of follow up years and if you follow people for long enough, a proportion will die of a cause of some sort. If you follow everyone in this room for 100 years, you would get 100% mortality. So if you follow people with Barrett’s long enough, people will die of something. These are the numbers of those who died and these are the numbers of those who died of cancer of the esophagus; it was between 2.5 and 6%, roughly about 5%. Therefore, in these series, with the present methods of management, only about 5% of patients with a long-segment Barrett’s actually died of esophageal cancer. We sometimes find early cancers in these people, and at least four other patients in these different series developed a cancer but did not die of the cancer, they died of something else. That is quite different from a cancer causing an obstructive mass in the esophagus. So, 95% of patients with Barrett’s will not die of esophageal cancer.

It has been shown that postoperative radiation and chemotherapy do not increase survival in adenocarcinoma. Preoperative chemotherapy alone is not helpful, although there is a recent paper in the New England Journal of a large multi-center study confirming that again. Presently, preoperative, so-called neoadjuvant, chemotherapy plus radiotherapy looks somewhat more promising,. with radiation being given to improve control over local disease, and sometimes by the time the operation is done, about 20 to 25% of patients do not have any tumor left, and with chemotherapy to control micrometastases, which are present in so many of these patients, but of course can’t be seen but show up as distant metastases at a later date. Thinking was changed by this paper from Walsh and his colleagues in Dublin, Ireland, which appeared in the New England Journal in 1966. They took about 113 patients and randomized them to either have surgery alone or to have surgery preceded by multimodal therapy, which was 5-FU plus cisplatin plus radiation to the primary tumor. As you will see, by three years the survival was much better in the patients given the neoadjuvant preoperative therapy compared to proceeding immediately to surgery.

While it is fairly difficult to do radiation and chemotherapy, it has its own problems and complications and this paper, it was decided at our institution, was not enough to change the entire thinking and to proceed on this one paper alone. There was a smaller one from the University of Michigan with a similar trend to give everyone preoperative therapy. So the present situation is that we surgeons, oncologists and radiation therapists are collaborating with multiple other North American centers in the CALGB-97 study, which is attempting to reproduce this study with preop chemo and radiation followed by surgery versus surgery alone. We are randomizing patients into that study, but it will be several years before the results are finally determined.

I’m afraid that for many patients, possibly half of all patients with adenocarcinoma of the esophagus presenting with symptoms, an operation is not possible because of distant metastases, because of age or other medical conditions, and sometimes all that can be done is the endoscopic placement of a metal spring stent to enable them to be swallow, as is being done here.

Let’s go back to Barrett’s esophagus, the place where most adenocarcinomas of the esophagus begin. Normally, all of your esophagus is lined with pale pink squamous mucosa and the Z line, the junction, should be at the lower end of the esophagus. In Barrett’s esophagus, a variable length of lower esophagus is lined with red columnar epithelium, easily seen at endoscopy. Barrett’s esophagus is an acquired condition and is associated with severe gastroesophageal reflux. Most patients will have a weak lower esophageal sphincter and over 90% of patients will also have a sliding hiatal hernia. These are at least two of the factors that probably cause reflux to occur, reflux then damaging the normal squamous lining here, which becomes replaced later by columnar. This is an acquired condition and this is an interesting and unusual patient, but we are showing a typical finding, a patient with reflux symptoms. On the left here, biopsy from the lower esophagus shows squamous epithelium with changes of esophagitis, elongated papillae and a thickened basal cell here.

The first slide to give you a typical history is a patient I saw about a year ago, a man of 61, who came with a complaint of three months of increasing dysphagia for solid foods, which were sticking at the chest level. When asked, he also had a 30-year history of heartburn and sour reflux, for which he had never sought other treatment except taking antacids, so this had never been investigated. In the population, there are many people with heartburn who have not had it investigated. A barium x-ray was done at his home hospital and this showed an esophageal ulcer with what was called edema. However, they recognized that it wasn’t just edema and did an endoscopy at his home hospital, which showed an adenocarcinoma and a Barrett’s esophagus which had not been recognized previously.

The results of treatment of patients of this type are unfortunately not very good. This shows the results of surgical resection in seven fairly large series. Here are the numbers of patients who had resection for adenocarcinoma of the esophagus and here is the mortality rate. While these series may have been written a few years ago, the median mortality here was about 8%. Here is the complication rate. The median complication rate here was about 40%, with such problems as anastomotic leaks, pneumonia, postoperative strictures and empyemas. This is difficult surgery. For all that the patients went through, and the surgeons, the five-year survival was not very great. The median five-year survival for patients presenting with a mass in the esophagus which was an adenocarcinoma in these series was about 15%. How could we do better? This is a very brief review of a lot of literature in the oncology and radiation area.