The lesion is umbilicated...  As you probably remember, umbilicated lesions in the duodenum are usually NET's.  For some reason, they look different to NET's found elsewhere in the gut, where thin vessels are usually seen crawling up their sides.  In the photograph below, you can see that at the very earliest stage (first image), there is no central indentation. However, very soon a subtle central dip develops (second image) which at the advanced stage (when I was tasked with removing the lesion), it has a deep central pit which was intermittently bleeding. 

This proved to be a WHO grade I NET (neuroendocrine tumour, with a mitotic rate <2%). These lesions always seem to be tethered down close to the muscle propria and can be difficult to remove without taking a chunk of muscle propria with it. 

Duodenal NET's should of course be discussed at the 'neuroendocrine MDT' and usually end up having: 1) Measurement of plasma chromogranin A (pCgA) levels. CgA is produced by all cells derived from the neural crest and high levels are found particularly in patients with metastatic disease. 2) A ‘gut hormone screen’ (measurement of gastrin, glucagon, vasointestinal peptide, somatostatin and pancreatic polypeptide levels). 3) Finally, an ‘octreotide scan’ (somatostatin receptor scintigraphy) or, even better, a PET/CT scan using peptides that bind to somatostatin receptors (68Gallium-DOTA-TOC/NOC/TATE). 

If it wasn't for the somewhat odd, elongated appearance of the ulcer, it wouldn't seem odd. However, the only clue to the aetiology is that yellowish appearance.  You may be familiar with 'iron pill gastritis' but can you get iron pill duodenitis?  Yes you can!  It has even been reported in the hypopharynx and the oesophagus. Those dark spots on the histology slides are crystalline iron deposits are easy to spot for the pathologists. 

Of course, ferrous sulphate (Fe2+)  tablets is the non-toxic form of iron (naturally).  However, the tablets can produce mucosal inflammation if the iron becomes oxidised to  Ferric iron (Fe3+). You may remember from medical school that most dietary nonheme iron is in the toxic, oxidized form of ferric iron (Fe3+).  To absorb iron it first needs to be reduced, in the presence of acid, to ferrous iron (Fe2+).  After this, the iron is easily transported across the enterocytes via a divalent metal-ion transporter 1 (DMT1) protein.  Here is a link to an open access article if you want to brush up on iron absorption. 

Naturally, the treatment is to change to the liquid form of iron supplement which is much less toxic to the mucosa than solid iron tablets. 

Initially, the patient appeared to have classic coeliac disease, with diarrhoea and weight loss together with a positive serology.  The endoscopy was crucial as the duodenal biopsies identified the presence of a thick band-like deposit of collagen just below the duodenal epithelium. On the basis of this finding, collagenous sprue was diagnosed.
 
Collagenous sprue was first described in 1947,1 but it was not until 1970 that Weinstein et al. introduced it as a diagnostic term to the medical nomenclature.2 Collagenous sprue is more frequent in females and in individuals who have other autoimmune diseases.3 It is now recognised that collagenous sprue shares similar clinical features with coeliac disease, such as chronic diarrhoea, anaemia and weight loss. In addition, the endoscopic and histological features of both diseases are similar, with an atrophic and scalloped duodenal mucosa. However, the histological hallmark of collagenous sprue is the presence of a thick subepithelial collagen band.  Such collagen bands may also be found in collagenous gastritis and collagenous colitis.4
 
It has been proposed that collagenous sprue may be a heterogenous condition of collagenous gastroenteritides, including conditions such as collagenous colitis and coeliac disease.5,6 Unlike coeliac disease, in collagenous sprue, the typical histological changes may also be found in the stomach and colon.7 Furthermore, greater numbers of IgG4 plasma cells have been reported in the duodenal mucosa of patients with collagenous sprue when compared with the numbers in patients with coeliac disease, duodenitis or normal duodenal mucosa.8
 
As in our case, patients with collagenous colitis usually also have positive coeliac serology.  For this reason, some believe that the collagen band is simply a marker of particularly severe coeliac disease, which may also be associated with ulceration, perforation and T-cell or B-cell lymphoma.9
 
An interesting report of paraneoplastic collagenous colitis was reported by Freeman et.al.10 In this case a patient with collagen deposits in both the small and large intestine was also found to have a coincidental colon cancer. After surgery, both the malabsorption and the histopathological changes completely resolved!  For this reason, a search for underlying malignant disease should be considered.
 
The response to a GFD is usually disappointing and for this reason, the outlook used to be grave for patients with collagenous sprue.  However, remission of the condition has been reported with courses of corticosteroids11 or immunosuppressive agents such as infliximab.12
 
References

1. Schein J. Syndrome on non tropical sprue with hitherto undescribed lesions of the intestine. Gastroenterology 1947; 8: 438–460.
2. Weinstein WM, Saunders DR, Tytgat GN, et al. Collagenous sprue—an unrecognized type of malabsorption. N Engl J Med 1970; 283: 1297–1301. 
3. Rubio-Tapia A, Talley NJ, Gurudu SR, et al. Gluten-free diet and steroid treatment are effective therapy for most patients with collagenous sprue. Clin Gastroenterol Hepatol 2010; 8: 344–349.e3. 
4. Zhao X and Johnson RL. Collagenous sprue: a rare, severe small-bowel malabsorptive disorder. Arch Pathol Lab Med 2011; 135: 803–809. 
5. Maguire AA, Greenson JK, Lauwers GY, et al. Collagenous sprue. A clinicopathologic study of 12 cases. Am J Surg Pathol 2009; 33: 1440–1449. 
6. Nielsen OH, Riis LB, Danese S, et al. Proximal collagenous gastroenteritides: Clinical management. A systematic review. Ann Med 2014; 46: 311–317. 
7. Robert ME, Ament ME and Weinstein WM. The histologic spectrum and clinical outcome of refractory and unclassified sprue. Am J Surg Pathol 2000; 24: 676–687. 
8. Schoolmeester JK, Jenkins SM, Murray JA, et al. Increased immunoglobulin G4-positive plasma cells in collagenous sprue. Hum Pathol 2013; 44: 1624–1629. 
9. Freeman HJ. Collagenous sprue associated with an extensive T-cell lymphoma. J Clin Gastroenterol 2003; 36(2): 144–6. 

You'll know that NSAID induced strictures are classically 'membrane-like' as in Photograph D. 

The 'Peptic stricture' is shown in photograph A.  The healed duodenal ulcer had been at 9 O'clock in that stricture.  

Chronic pancreatitis causes external compression of D1 or D2 usually with swollen villi. Of course there are two photographs to choose between - Photo C or E.  Lets leave that one and see if the Crohn's stricture is easier to find.

Of course, photograph B shows ulceration and inflammation and must be the CD stricture. 

That leaves the malignant stricture which could be photograph C or E.  It's usually relatively easy to squeeze past the external compression caused by a swelling in the head of the pancreas. In contrast, malignant duodenal strictures are usually impassable. Which one looks the tightest?  Photo C or E?  Well, C is the malignant stricture which I had to dilate up to 10mm before I could take samples a little deeper into the stricture to confirm the diagnosis.  Photograph E was a case of severe pancreatitis which we scoped as he had dropped his Hb (never found out why or how). 

The 3 polyps in column A are all adenomatous polyps.  Flat and plaque-like in the stomach,  pale in the duodenum and a sessile TVA in the colon. 

Column B shows NET's with the prominent vessels along it's side in the stomach and colon whilst the duodenal NET has the typical central depression.

Column C shows 'hamartomatous' polyps seen in Peutz-Jeghers syndrome. There is hardly ever any dysplasia within these polyp and it's uncertain what the 'precursor lesion' is for these patients GI cancers

Column D show inflammatory polyps which are typically angry red and sometimes with pretty white spots ! 

This polyp doesn't look adenomatous. Remember that in the duodenum, the adenomatous polyp are usually flat, often with white white edges and with a slit-like or gyrate crypt pattern. Of course, in this case the crypt pattern is gyrate, almost exaggerated gyrate!  The rather beautiful surface rules out an inflammatory polyp and an NET, both of which don't have this pretty surface structure.  Even the histology is rather beautiful, see example below!   

In addition, as you probably know, most duodenal NET's will have a central indentation and will not have a distinct crypt pattern. 

Cutting to the chase, this patient has Peutz-Jeghers syndrome !   This is a good example of the hamartomatous polyps these patients develop throughout their GI tract.  Similar hamartomatous polyps also develop in the 'Cronkhite-Canada syndrome and Juvenile polyposis.

Hamartomatous polyps are characterized by 'disorganized overgrowth' of the normal tissue but without visible cellular instability (dysplasia). Of course, these patient have an appalling lifetime risk of developing cancer which can be as high as 90%, including gastric, colorectal, breast, colorectal, and pancreatic malignancies. 

Unfortunately, there is no no evidence that removing hamartomatous polyp will reduce the subsequent risk of cancer.  For this reason the only justification for attempting the somewhat hazardous endeavour to remove PJS polyps deep within the small bowel is to reduce the risk of subsequent intussusception.   Rather surprisingly, a quick medline search has not revealed any RCT evidence that this actually is worthwhile !  

​Umbilicated lesions in the duodenum are usually NET's. Even when the NET is as small as 7-8mm they will usually start to develop a central dimple as in the example below !

Ampullary NET's have a far worse prognosis than NET's situated elsewhere in the duodenum and are therefore first considered for a surgical resection.  However, a small non-functioning, non-ampullary NET's may be considered for endoscopic resection.  

I don't think that I was thinking entirely clearly when I accepted the referral from the Neuroendocrine 'multi-disciplinary team' for an attempt at endoscopic resection.   The lesion is about 2cm and I quoted a 1:10 risk of requiring an emergency surgery in case of a perforation. This was probably an underestimate because NET's develop in the submucosa and you almost have to scrape them off the muscle propria layer which is VERY thin in the duodenum.

Furthermore, I think that a small unrecognised 'micro-perforation' in the stomach is probably often 'sorted out' by a timely intervention by our great friend - the omentum!  The duodenum on the other hand is retro-peritoneal and bathing in pancreatic juices.  After 12 hours, that 1mm perforation will be 1 inch across and the patient will be in a desperate state. For this reason, if you do undertake duodenal resections you should be backed up with a 24-hour/day 364 day/year specialised pancreatic emergency surgery service.  It's no good having an breast surgeon trying to patch things up! 

Now being an older and more experienced endoscopist, I believe that a far more sensible cut-off for endoscopic resection of a duodenal NET is probably 10-15mm.  Personally, in an old person with comorbidities, who are unlikely to survive an emergency laparotomy, I would decline if I thought that the risk of emergency surgery was <1:100. Of course the NET could grow but is unlikely to turn malignant and kill an 85 year old patient with a previous myocardial infarction and dodgy kidney function. Particularly if the lesion is 'WHO grade is 1' only !  

Of course Rembackens Resection Rule #15 applies: "A good endoscopist knows when to resect and a great endoscopist when to reassure!  Also known as "Disengage if there is Disease".

This is exactly what happened !   The 'lesion' was soft to the touch and previous samples had not revealed any abnormality.  Of course a Brunner's gland hyperplasia could be a possibility but the fold was too soft and in any event, these don't need to be removed!   On the other hand, the patient was standing in the endoscopy room expecting his 'lesion' to be removed, to 'find out what it is! '  

At times you may feel that you are 'put on the spot' in endoscopy, being asked to do things which you don't agree with or perhaps procedures you don't feel qualified to carry out unsupervised. 

Of course, you are more than a 'technician' and it's your duty to do what's best for the patient.  That patient is 'your patient' in the endoscopy unit and should something go wrong, you'll have to take full responsibility.   Don't let yourself be pushed around by either the patient or his physician.    "I was just following orders" doesn't wash if there is a complaint !     

Anyway, I diagnosed an 'odd looking but entirely innocent duodenal fold', took another couple of biopsies (to calm things down) and reassured the patient !   

The Spigelman score is a little peculiar in that patients with scores of 0 to 3 have a very low risk of developing cancer. However, when patients accumulate a score of 4, surgery should be considered as there is a high risk of cancer.  

Of course, many patients with FAP have desmoids which precludes surgery. In those cases I consider intervening endoscopically if: 1) the patient is fit enough to survive emergency surgery to sort out a perforation or cope with a 4+ unit blood loss AND/OR  2) there is a higher than average risk of cancer, for example in a patient with extensive HGD or confirmed IMca. Of course, the lesion also has to be 'resectable' within your level of expertise. Personally, I find lesions situated at 12 O'clock the most challenging to remove.  Furthermore, a general anaesthetic is probably a prerequisite for resections taking longer than 20 min or so (which they always seem to do). 

Thus, I consider the "correct" answer to be 'Abort'!  I decided that the lesion was likely to be too large for me to resect.  Fortunately, the patient was a good surgical candidate.  A 'pancreas sparing duodenectomy' was carried out and a TVA with mainly LGD but also areas with HGD was completely removed. Postoperatively the patient developed pancreatitis and was discharged 4 months later...  Clearly this is 'Tiger Country' for surgeons as well !  

The ampulla does indeed seem a little plump! There must be something inside it which is pushing it out like this!  We organised an EUS which confirmed a nodule inside the ampulla. FNA confirmed adenocarcinoma. As the lesion extended deeper than 5-6mm (on EUS) and because cancer was confirmed, there was no question about attempting an ampullectomy.  

As the patient was jaundiced and there was no suggestion of cholangitis, an ERCP would not be indicated.  A subsequent MRI also identified a small mass at the bottom end of the bile duct but of course could not provide the tissue diagnosis.  

>95% of our facebook group correctly recognised that this lesion is a sporadic duodenal adenoma.  Of course, you should now consider offering your patient a colonoscopy as there is a link with colorectal adenomas. 

Actually, it turned out to be tubular adenoma with low-grade dysplasia (TA+LGD). The risk of malignant conversion in a 'sporadic duodenal polyp', is probably very low. Nevertheless, I do remove these lesions in patients who are at least "reasonable fit" with a life expectancy of at least 10 yrs. 

Because there is a 1:20-30 risk of late bleeding even after removing a small polyp, I always apply clips to the mucosal defect. In addition, I tend to reject patients who I think may not 'cope well' with a 2 unit blood loss or an emergency laparotomy. 

By the way, I think that using a banding device such as the Cook 'Duette' or the Boston Sci 'captivator', is probably dangerous in the thin walled duodenum.  You should also be careful with the use of the haemostatic forceps. There is a real risk of a late perforation if too much heat is applied to the thin mucosa. 

Finally, immediate perforations are often difficult to recognise in the duodenum. Not sure why this is but my last perforation was pinhole and entirely invisible at the time of the procedure. I know this because I watched the video clip several times afterwards.   After a few hours exposure to pancreatic juice, the hole was an a centimetre in diameter (measured at the time of laparotomy) and the poor patient spent another 2 months in hospital recovering. Fortunately, 12 months later the external drains could be removed. 

A recent study published in GIE [Fukuhara S et al. Management of perforation related to endoscopic submucosal dissection for superficial duodenal epithelial tumors. Gastrointestinal Endoscopy 2020;91(5):1129-11370] reinforced my view.   This Japanese multicentre study reported on the management of perforations following ESD resections of duodenal polyps.  A perforation complicated 14% of ESD resections and only  1/3 of perforations could be closed completely endoscopically.  Fortunately, even perforations which were not completely closed, only 2/36 perforations required surgery. The remainder could be managed with the placement of an 'endoscopic naso-biliary and pancreatic duct drainage tube.  

The duodenum is "Tiger Country" and I recommend that you avoid polypectomy here unless you have full 24 hour/364 days/year backup by expert upper GI radiology and specialist surgical backup.  You need the reassurance that it will be a friendly and support specialist upper GI surgeon who is sorting out your patients biliary peritonitis at 3 O'clock in the morning! 

OK, there is no way to beat around the bush. This turned out to be a 'pyloric gland adenoma'. I don't think that there is any way that I could have distinguished this from a 'normal' duodenal adenoma endoscopically (you also have the histology to help you though). A Brunner's gland polyp is unlikely as the lesion isn't a beautiful, symmetrical submucosal ball, covered with normal duodenal mucosa. Gastric heterotopia is never this 'polypoid' or 'flamboyant'. Finally, a NET is unlikely as these should also be 'balls', classically with a central depression in the duodenum. 

The first time I came across a 'Pyloric gland adenoma', I objected to the histology because the polyp was situated in the stomach. Our lovely pathologist patiently educated me and explained that in spite of the name, they don't necessarily arise from the pylorus.  Instead, they got their name from the fact that they arise from 'pyloric glands' which are found in the oesophagus, stomach, pylorus and the duodenal cap.  

Furthermore, I didn't know that gastric adenomas are broadly classified into two types: 1) Gastric type and 2) the Intestinal type of gastric adenoma.   Gastric-type adenomas are further subclassified into 'pyloric gland adenomas' and 'foveolar-type adenoma'.  However, it's the intestinal-type adenomas which we endoscopists usually refer to as 'adenomas' in the upper GI tract. 

Pyloric gland adenomas are found in patients with Helicobacter pylori associated gastritis and 'chemical gastritis'. However, they are indeed 'neoplastic' (the cells are unstable) and may harbour low-grade dysplasia, high-grade dysplasia or even cancer.  If you would like to learn more about 'Pyloric gland adenomas', this is a good link: http://www.archivesofpathology.org/doi/pdf/10.5858/arpa.2013-0613-RS

Of course the endoscopic appearances are that of a submucosal lesion, either a lipoma, a lymphangioectatic cyst (if soft when prodded with a biopsy forceps), or firms as you would expect with a GIST, a leiomyoma, an ectopic pancreas (although have never seen one in the duodenum, a NET or a 'Brunner's gland hyperplasia'.
 
Of course, if a submucosal lesion is firm, the only way to make a diagnosis is to slice the mucosa open with a needle knife (or if like me you don't want to spend hundreds of pounds, use the tip of a snare with the diathermy set on 'sphincterotomy' or EndoCut I ) and then sample from deep within the lesion (don't forget to place a couple of clips afterwards). 
 
Although on balance of probabilities, a Brunner's gland hyperplasia would be the most likely diagnosis, it could make sense to obtain some samples to rule out a GIST
.  Seeing >10 mitotic figures/HPF is a sign of a 'bad GIST' requiring surgery.  Of course, if the patient is too old for surgery, it would make little sense in obtaining any samples.  I leave the decision about sampling to the 'upper GI cancer team'.
 
Anyway, Brunner's gland hyperplasia (sometimes called a 'hamartoma' or 'adenoma' even though they are not neoplastic) are supposedly rare, only accounting for some 5% of all duodenal masses seen on radiology.  In my experience, excluding lipomas and lymphoangioectatic cysts (which are both soft of course) Brunner's gland hamartomas are the most common 'firm' submucosal lesion. I usually find these in the duodenal bulb and less commonly just beyond the D1/D2 junction.