Diagnosis of neuroendocrine tumors - Introduction
 

Holly Atkinson MD

Lowell Anthony MD: Neuroendocrine tumors arise from distinct cells in the body that are essentially ubiquitous. They can arise from the pituitary gland, the thyroid, the adrenals, the pancreas, the lungs. But most of them come from the GI tract. And that allows us to characterize patients in terms of their symptoms and their illness based upon the cell of origin.

Holly Atkinson MD: Which are the hormones that we usually see produced by these tumors?

Lowell Anthony MD: Hormones like insulin, glucagon, gastrin, ACTH. These are all hormones that are involved in normal physiologic processes. Serotonin is one that deserves a lot of recognition because of the carcinoid syndrome. But essentially, almost any hormone can be oversecreted and overproduced.

Holly Atkinson MD: Dick, of course not all the tumors are functional. Talk to us about functional versus nonfunctional tumors.

Lowell Anthony MD

Holly Atkinson MD: Lowell, what about the malignant potential here? How aggressive do these tumors tend to be?

Lowell Anthony MD: Well, neuroendocrine tumors can be quite diverse in terms of their malignant potential, ranging from the totally benign, such as tumors that are common in the pituitary and referred to as adenomas, to the opposite extreme. One of the most aggressive neuroendocrine tumors would mimic carcinoma of the lung.

Holly Atkinson MD: Symptomatology can be quite hard to discern, certainly, in the early stages. Talk to us about the problems that this creates for the clinician.

Richard Warner MD

Thomas O'Dorisio MD: I appreciate that question, because there's one other caveat about the neuroendocrine tumors we've been alluding to: they're not autonomous. They do regulate themselves, and I believe insulin-secreting tumors are among the most deceptive. I think the average time to diagnosis may be from three-and-a-half to five years. I think the type of hormone, how important it is in us, makes a difference to the degree and the detection. Gastrin is so profoundly important in terms of the physiology of gastric acid secretion but it still takes a while. Zollinger used to say it would take up to three years to make the diagnosis of his tumors, you know, because they'd have the bleeding diathesis, and they don't have that now.

Richard Warner MD: Because of the various potent drugs that are available nowadays, the clinical presentation of gastrinoma with a Zollinger-Ellison syndrome is changing. We have widespread use of the proton pump inhibiting drugs so that severe or atypical ulcer disease is not as common with Zollinger-Ellison syndrome or gastrinoma now as it used to be. Diarrhea is probably a more common presenting feature, and about 10% of all gastrinomas which cause a functioning syndrome may not even have diarrhea but present simply with flushing that's very reminiscent of carcinoid. So there's a lot of overlapping features in these tumors.

And one other point that's important is that although a major hormone, or a production of one in particular, is characteristic in most of these cases, many of the tumors have the ability to make multiple hormones, and so we may have overlapping syndromes present.
Holly Atkinson MD: Well, Lowell, let's step back for a moment. What are the most common neuroendocrine tumors? Obviously, they're rare as a category.

Lowell Anthony MD: About half will come from the GI tract and be classified as a well-differentiated neuroendocrine tumor consistent with carcinoid or carcinoid-like features. Another half would arise from the pancreas, and be what we refer to as the islet-cell tumors. So from a clinical perspective, that's where the majority come from.Now, there can be some from the lung. It's a small group. There can be some from the reproductive organs, the ovary, but the majority of them would be-half would pretty much be from the GI tract, and about half would be islet cells.

Thomas O'Dorisio MD: Certainly, the most common pituitary tumor is the prolactin-secreting tumor, then the growth hormone-secreting tumor, then the ACTH-secreting tumor, which Lowell alluded to and which can be part of another multi-hormone secreting tumor from the lung, ACTH. But by and large, when you're talking about neuroendocrine tumors, as we are, those in endocrinology are the pituitary and medullary thyroid cancer.

Richard Warner MD: All carcinoids, which are the commonest of the neuroendocrine tumors, constitute only 0.5% of all malignant growths.
Islet cell neuroendocrine tumors, of which only a small percent were carcinoid, were not very prominent as a source of carcinoid. Rectal carcinoids began to show up as more common, too

Holly Atkinson MD: Dick, what about early diagnosis? If you are able to get a patient early, what's the effect upon quality of life and ultimate outcome? Let's talk about quality of life first.

Richard Warner MD: Even though these are slow-growing, long-lived tumors in general, nonetheless they do grow and they do very often spread. Once spread, it is impossible to effect a complete cure. The only way for a complete cure is by total surgical resection of the primary, the original tumor before it has spread. Obviously, an early diagnosis is necessary to accomplish that.

Since these tumors are very often without production of symptoms for a long time, the early diagnosis may be very much delayed, so this can be a problem. Diagnosis is not that difficult. It's thinking of the possibility (of NET) that's key to the whole thing. It's got to be in your thoughts. Then you can look for it. And then it's not so hard to find, usually.
Once the tumors have spread, even then an earlier diagnosis is better because debulking and removing a primary tumor, even with spread, has been shown to prolong the survival and improve the quality of life. So the earlier the diagnosis is made, either before the tumor has spread or after, the sooner the patient can undergo treatment which will be of benefit.

Holly Atkinson MD: There's a delay in diagnosis. You're saying early diagnosis obviously affects long-term outcome. It affects quality of life assuredly. How do we close that gap? Lowell, what's the most effective way to get the physician to think about it in his differential diagnosis?

Lowell Anthony MD: It's really going to come down to the person that sees the patient, perhaps for the first time, and has a good relationship with that patient. Symptoms that don't respond to initial types of intervention are the ones that potentially could give rise to more complex problems. This is difficult. You know, irritable bowel syndrome is the most common diagnosis of these patients before the correct diagnosis is made, and that might take from three to seven, maybe ten years.

Holly Atkinson MD: What usually triggers the diagnosis?

Lowell Anthony MD: It would be something more than the diarrhea that may be present, maybe a new symptom, like flushing or some pain. That might do it, or weight change-usually weight loss rather than weight gain.

Thomas O'Dorisio MD: I had an 18-year-old young lady referred to me who had "a history of chronic diarrhea since the age of three," and then developed, over a period of three to four years, a 70-pound weight loss. In fact, this was medullary thyroid cancer, which is a neuroendocrine tumor, one of the few thyroid cancers associated with diarrhea.

Richard Warner MD: One of the two main presenting complaints, at least for carcinoid, would be the occurrence of acute intestinal obstruction, which happens to be the presenting feature in many cases-the tumor is only found at the time of surgery. If one reviews the patient’s history, there will have been previous episodes over a long time that have come and gone, waxed and waned, so this is a common occurrence. The other main presenting complaint is Crohn’s disease, and gastroenterologists see a fair amount of patients with Crohn's. Crohn's disease of the small bowel is not always amenable to tissue confirmation of the diagnosis. The symptoms may be appropriate, with diarrhea, perhaps weight loss, cramps, and also an abnormal x-ray, a small bowel series showing disturbed loop or segments of bowel. But it has been observed in a significant number of patients diagnosed and treated clinically for Crohn's but not responding well, coming to surgery after a number of years, that what they really had was unrecognized carcinoid that took a long time to develop.

Holly Atkinson MD: Once a physician suspects that there may be a neuroendocrine tumor, what are the diagnostic tests?

Richard Warner MD: The workup is divided into several facets. First, the physical examination may discern features of a syndrome, such as flushing or palpation of a mass. These are obvious and would lead to the suspicion. Then we have the laboratory tests, including urine and blood tests for the various hormones that might be suspected. And then we would test for a hormonal marker that might be universal for all of the tumors-the test doesn't distinguish which one, but if it's positive it makes a very strong case for something being present.

Holly Atkinson MD: And what are you looking for?

Richard Warner MD: I would be measuring chromogranin A.

Holly Atkinson MD: Tell us about that.

Richard Warner MD: Chromogranin A is an acidic glycoprotein which is co-secreted from neuroendocrine cells along with whatever other hormones, both active and inactive, that might be secreted. Interestingly enough, it is also often secreted when no active hormones are produced. The test is universally available now. It's a simple blood test in terms of the patient. It just requires serum, and it can be done by all commercial laboratories. That is very helpful and rather stable. It doesn't fluctuate widely.

There are situations in which a false positive, an elevated value, can be obtained, such as in a patient with chronic renal insufficiency or somebody with atrophic gastritis. The absence of gastric acid production ends up causing elevated gastrin, and the elevated gastrin is associated with increased chromogranin A. This will be seen, for example, in patients on proton pump inhibitor drug treatment, where they're made achlorhydric, so you may get a false elevated chromogranin A. You will also see this in the atrophic gastritis and achlorhydria associated with pernicious anemia.

Holly Atkinson MD: Lowell, what are some other markers to work with, to unscramble this?

Lowell Anthony MD: It's helpful, when the patient's being diagnosed or seen initially, to be thinking about the chromogranin A. But you should also see if the neuron-specific enolase or pancreatic polypeptide is elevated, particularly in some of the more aggressive tumors. These hormonal markers may not be specific for any one particular tumor, like the serotonin would be for the carcinoid. 5-hydroxyindoleacetic acid, the product of end-metabolism of serotonin, is the foundation marker for carcinoid diagnosis. But you can still have carcinoid and not necessarily have that elevated marker. So this is where all these hormones that we've discussed here play a role. You've got to be thinking broadly, the chromogranin A being the most broad; then serotonin, if there's flushing or diarrhea present, potentially; and then some of these more sophisticated hormonal markers, if there are clinical syndromes present, particularly hypoglycemia for insulin, hyperglycemia for glucagon, etc.

Thomas O'Dorisio MD: For multiple endocrine neoplasia 1, the 95 percent or greater of these folks who are affected will present with parathyroid hyperplasia, hypercalcemia. One of the things that has helped me the most has been a history, a family history of kidney stones. That's a tip-off. Or if they're learned or they're aware enough of their family, a pituitary tumor will sometimes be a tip-off.
And then, those tumors of the pancreas would be the history of recurrent ulcer disease or episodes that define a symptomatic hypoglycemic, as Lowell alluded to.

For multiple endocrine neoplasia type 2, medullary thyroid cancer, it's a bit more difficult. The thing that's helped me has been the history of high blood pressure in a very young relative or sudden death, from natural causes, supposedly, in a young person, because that would be the pheochromocytoma part of it.

Richard Warner MD: You did mention neuron-specific enolase. Others that are sometimes helpful are, for example, substance P and neurokinin K. These are not always positive, but when they are they help to confirm the presence of a neuroendocrine tumor, usually carcinoid. There's a wide number of different markers that may be increased. My custom, when I have a patient, is to do a panel of all of these markers initially, not only to help confirm the diagnosis, but also to clarify what kind of tumor it is, what they're producing and what might be expected clinically down the line.

Holly Atkinson MD: Lowell, you mentioned serotonin and carcinoid. Let's talk about that in a little more detail, because obviously serotonin is broken down very quickly. How do you go after that?

Lowell Anthony MD: Serotonin is an amine hormone that gets secreted by the carcinoid tumor cell, and we can measure it in the blood. But it's compartmentalized, meaning that it's going to exist in high concentrations in the platelets. So if we're going to look at total serotonin production, we're going to have to measure it in all the compartments within the blood. The corollary is simply to look at total serotonin production over a 24-hour period and look at the metabolite of 5-HIAA. Now, there's a small group of patients who lack the final enzyme to make 5-HIAA. We may have to back up one particular metabolite and look at the urine for other metabolites of serotonin besides the 5-HIAA. We've got to remember that 5-HIAA, even though it's the marker that we have the greatest experience with, dates back nearly 50 years.

Richard Warner MD: The 5-HIAA is most usually elevated with functioning midgut carcinoids. Hindgut carcinoids very rarely produce it. They also very rarely cause a carcinoid syndrome, and foregut carcinoids-that is, those arising in the lung or the stomach, the thymus or the pancreas, for example-may make serotonin, but they less often metabolize it to 5-HIAA. Some of these tumors follow different routes because of their chemical or their metabolic aberrations, so that 5-HIAA, if positive, is very helpful, but if negative does not exclude the diagnosis. Unfortunately, because it's the oldest of these known markers, 5-HIAA is often the only one that's relied upon. But you can't; you must look at these other indicators if you suspect one of these tumors.

Thomas O'Dorisio MD: It's very clear that unless you have a fairly high degree of tumor mass in the liver, your 5-HIAA may not rise. It's very short-lived. The half-life is less than a minute and a half. Our experience is that it's very sawtooth, and unless it's sustained-that is, unless you really have it produced and sustained long enough-you may not catch it. So that to put all your stock in that one test, to define whether it might be there or not, is dangerous.

Richard Warner MD: The chromogranin A is much more stable and uniform, with less fluctuation, and if I had to choose only one to measure, across the board for carcinoid and all the others, I would choose chromogranin A.

Thomas O'Dorisio MD: The commercial assays are available for chromogranin A on both the West Coast and certainly in the Midwest (of the USA). Not to name different laboratories, but there are some that provide chromogranin A in the plasma and-

Richard Warner MD: And also pancreastatin, and indeed I usually do both.

Holly Atkinson MD: Let's move on to imaging. Lowell, give us a starting place here. What will various imaging modalities do in the workup?

Lowell Anthony MD: It's helpful to start off with common imaging modalities and work up to the more sophisticated kinds.
What is frequently done is an ultrasound of the abdomen or a CT scan. Then you would eventually go on to an OcreoScan if a neuroendocrine disease is suspected.

Thomas O'Dorisio MD: The history of the OcreoScan and how it evolved from the octreotide acetate in 1989 is fascinating. It is probably the single most helpful test to rule in or rule out a neuroendocrine tumor-I mean, a gastrin-secreting tumor.
I thought it was very powerful-and it was cost-efficient, or cost-effective. It was Krenning's original work in the Rotterdam group that really pioneered the OcreoScan or the somatostatin-labeled indium-111, congener-labeled indium-111. If you look at that, about 9.2 out of 10 neuroendocrine tumors will be positive. That's a huge percent. OcreoScan is clearly the alpha pup, if you will, in terms of this type of technology, where octreotide binds with the highest affinity to receptor subtype 2.

Richard Warner MD: I agree with you. OcreoScan is the gold standard for imaging these tumors. It cannot always localize them very precisely.
If you do an OcreoScan in search of a suspected gastrinoma and you get a hotspot in the middle of the epigastrium, that's pretty good confirmation that that's a possible diagnosis. But it doesn't tell you if it's in the pancreas or in the duodenum or exactly where it is. You have to follow up with something else to localize. My general practice has been endoscopy, with endoscopic ultrasound, if necessary, in order to localize and even with needle biopsy to confirm the lesion which had been suggested by the positive OcreoScan.

Holly Atkinson MD: What's on the horizon?

Thomas O'Dorisio MD: Substance P scintigraphy for lymphoma is one that's being developed, and vasoactive intestinal peptide, or VIP, scintigraphy. This is very appealing for many of the brain tumors, specifically, because there's a lot of VIP receptors in brain tumors. To have a scintigraphy there, or to use indium or even technetium-99M-labeled, would be marvelous.

Richard Warner MD: VIP scintigraphy has been shown to be more effective than OcreoScan for imaging medullary thyroid carcinoma, and now there have been reports from Europe of very effective utilization of isotopes of gallium and of lutecium for imaging or for somatostatin receptor scintigraphy, with the sensitivity being fourfold greater than that of the OcreoScan, the indium-111.

In addition to that, for specific tumors, such as carcinoid, the use of carbon-11 attached to dopamine has been very effective in indicating tumors which did not show or didn't show well with the OcreoScan.

Thomas O'Dorisio MD: In development in Europe, to my knowledge, is another somatostatin congener, or analog, if you will, a congener that will see more than one receptor subtype for somatostatin. That's going to have broad implications for something like colon cancer localization, since somatostatin receptor subtype 1 is much more predominant than subtype 2.

Lowell Anthony MD: There's another somatostatin analog on the market for scanning called Neotech. This is depreotide. This is an analog of somatostatin that's similar to OcreoScan, but it doesn't see exactly the same type of information that we see with the OcreoScan. So in my practice, if the OcreoScan's negative and I really am convinced that patient has a neuroendocrine tumor, then I'll consider ordering a Neotech scan, or sometimes an FDG-PET scan if I think there's quite a bit of mitotic cells there. It's on the market for the evaluation of solitary pulmonary nodules, and indicates more likely that a solitary pulmonary nodule may be malignant rather than necessarily a neuroendocrine-type tumor.

Richard Warner MD: The Neotech scan utilizes technetium-99, and is particularly attracted to somatostatin receptor type 3, which is not the main one that attracts the OcreoScan. And that's why it will show up some tumors which do not image on the OcreoScan.

Thomas O'Dorisio MD: I think the Neotech is extremely valuable in the lung. It's very noisy in the abdomen, but I think Lowell made a very keen point there that I hadn't thought about or realized or known. The Neotech almost tells you, too, about the aggressivity. As Richard says, it's a subtype 3 receptor, but it has a higher affinity for it, rather than the receptor subtype 2. Receptor subtype 2 is about 90% or so loaded on the neuroendocrine tumors, and in general the more receptor subtype 2 you have, the more slow-growing, the more benign the tumor.

Holly Atkinson MD: Let's move along now and assume that we've located the tumor, we've biopsied it. What then, Dick? How do you go after further evaluating?

Richard Warner MD: I like to have other stains and immunohistochemistry done in order to clarify just what it is. It has to be recognized that almost all of these neuroendocrine tumors by standard H&E stain look the same. You can't tell the difference under the microscope with a regular H&E preparation between a carcinoid and a gastrinoma or a VIPoma or one of the other tumors. The only way you can tell is to delineate what the main hormone is that's produced, so you have to do immunohistochemical staining for gastrin or for serotonin. You would suspect from doing the appropriate blood tests, urine tests, and then you would have this stain done.

In addition, besides the appearance of the standard stain, you should do things like staining for synaptophysin and chromogranin A, which confirms that it is a neuroendocrine tumor.

Once having done that, I also like to do other things on the microscopy of an individual tumor in order to tell about its aggressiveness, and to have a prediction of what kind of behavior you can expect. I want a count of the mitoses to determine if this is a very slow-growing tumor. In addition to that, I like to have a transforming thyroid factor 1 stain done, because if it's positive, especially strongly positive, it means it's of foregut origin.

I also would like to have KI-67 or MIB-1 stained in order to determine the proliferation index. If the proliferation index is very high, it means we have a very aggressive tumor, and if that, coupled with a high mitotic count, is present, that speaks for more aggressive chemotherapy.
Nowadays, we also like, especially in tumors we think are very aggressive, to stain for C-KIT and to also determine if the P53 abnormalities are present on immunohistochemistry.

Holly Atkinson MD: Obviously, what you've all been pointing out is how difficult at times and how challenging diagnosis can be. Lowell, to wind up this segment, review for us what you think are the major challenges in making the diagnosis.

Lowell Anthony MD: I really think it's listening to the patient. It boils down to what is present, what is responding to the expected therapy and what's not. You may have a complex of symptoms, an evolution of symptom complexes that are not following a typical pattern. We've talked about a lot of things here that are complex, but I think that practicing physicians, if they do have somebody who is not following the expected course of, say, Crohn's disease or some idiopathic flushing category, must think a little bit outside of what they normally might think of. They must measure some of these biochemical markers or do some of this testing that we've talked about. Literally, I think it comes down to looking at signs and symptoms.

Holly Atkinson MD: And listening to the patient.

Lowell Anthony MD: Exactly.

Thomas O'Dorisio MD: And listening to the patient, yes. Excellent.
What is carcinoid syndrome?

Moderated panel discussion: diagnosis of neuroendocrine tumors

Thomas O'Dorisio MD