Barrett s oesophagus: specimen handling and reporting

Barrett s oesophagus: specimen handling and reporting Professor Neil A Shepherd Gloucester and Cheltenham, UK

The role of the pathologist in Barrett s oesophagus 1. Diagnosis 2. Typing 3. Assessing response to treatment 4. The diagnosis of neoplasia and surveillance

Barrett s oesophagus at endoscopy

Endoscopist vs. the pathologist long segment CLO is on the whole an endoscopic diagnosis short segment may need pathology much more USSCLO (cardia IM) is entirely a pathological diagnosis

The pathologist reporting CLO Perfect world download endoscopy report & images Real world Clinical information:? Barrett s oesophagus full knowledge of site of generous biopsies with diagnostic features two (if you re lucky) scrappy biopsies without specific features

Reporting CLO Hellier & Shepherd, 2005 Diagnostic for CLO: Native oesophageal structures are present with juxtaposition to metaplastic glandular mucosa, whether intestinalised or not About 10 to 15% of biopsy sets Takubo et al 1995, UKBOR 2003 Biopsies corroborative of an endoscopic diagnosis of CLO: Intestinalised metaplastic glandular mucosa with or without non-organoid arrangement, villous architecture, patchwork of different glandular types, etc. This could potentially still represent incomplete intestinal metaplasia in the stomach, especially in a hiatus hernia or IM at the cardia. Biopsies in keeping with, but not specific for, CLO: Gastric-type mucosa of either fundic or cardiac type without IM. Patchwork appearance is still possible, as is a non-organoid arrangement. Such appearances could, however, represent the oesophago-gastric junction or the stomach within or without a hiatus hernia.

the IM/goblet cell question should Barrett s be defined/diagnosed by showing intestinal metaplasia/goblet cells on biopsy?

90 80 70 60 50 40 30 20 10 The influence of the number of biopsies on the demonstration of IM in classical CLO Harrison et al, 2007 0 1 2 3 4 5 6 7 8 9 10 11 % with IM vs number of biopsies taken

The influence of the number of biopsies on the demonstration of IM in classical CLO 100 90 80 70 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 >9 No of Biopsies United Kingdom Barrett s Oesophagus Registry (UKBOR) study diagnostic biopsies of 200 patients Mandalia & Shepherd, 2010

Definition of Barrett's esophagus: time for a rethink is intestinal metaplasia dead? Riddell RH, Odze RD: Am J Gastroenterol, 2009 goblet cells are uncommon in pediatric patients with BE small percentage of adults have CLO without goblet cells the chances of detecting goblet cells proportional to the length of CLO sampling error common interpretation & differentiation of goblet cells vs. pseudogoblet cells difficult in some circumstances. goblet cells have been shown to wax and wane over the natural history of BE background non-goblet epithelium in BE is biologically intestinalised background non-goblet epithelium in BE shows molecular abnormalities similar to the goblet cell-containing epithelium well-defined risk of neoplasia in patients with esophageal columnar metaplasia without goblet cells a diagnosis of BE should not require demonstration of goblet cells in mucosal biopsies guidelines for the diagnosis of BE need to consider revisions that take into account new data

Intestinal differentiation in metaplastic, non-goblet columnar epithelium in the esophagus 89 patients with CLO immunohistochemistry for markers of intestinal differentiation: MUC2, DAS-1, villin and cdx2 metaplastic esophageal columnar epithelium without goblet cells shows phenotypic evidence of intestinal differentiation supports the theory that squamous epithelium converts initially to non-goblet columnar epithelium before goblet cell metaplasia Hahn HP et al, Am J Surg Pathol, 2009

Pathology and the diagnosis of CLO the IM/goblet cell/specialised IM story was always tosh and has now been shown to be by the two best known GI pathologists worldwide the BSG reporting guidelines are still valid there are pathological features of CLO that are still diagnostic (juxtaposed native structures, hybrid glands,? MLE) but these are all unusual I (and others) am becoming more and more convinced that pathology has little or no role in the diagnosis of CLO except when there is stricturing, ulceration or short segment disease (but even then it is difficult pathologically)

Typing Barrett s oesophagus: Paull et al, 1976 fundic is it really hiatus hernia? cardiac it may be intestinalised, it s just that we can t see it on H&E intestinal incomplete, complete pancreatic half-way houses hybrid glands, multilayered epithelium

Surface squamous re-epithelialisation, especially after ablation therapy

Squamous re-epithelialisation with mimicry of glandular dysplasia

Squamous re-epithelialisation over glandular dysplasia mimicking invasive adenocarcinoma

EMR in Barrett s oesophagus

Management of dysplasia H Barr, NA Shepherd Indefinite for dysplasia early re-evaluation with extensive biopsies following a course of PPI. Low grade dysplasia extensive biopsies after intensive acid suppression for 8-12 weeks. 6-monthly surveillance as long as disease is stable. High grade dysplasia if changes persist after intensive acid suppression and HGD is confirmed by two pathologists, oesophagectomy in specialised unit recommended. ablative therapy for patients unfit for surgery.

The average long segment CLO surveillance case with Seattle biopsies 40 biopsy pots 352 individual sections

From microscope to smart endoscope for the diagnosis of neoplasia in Barrett s oesophagus

Newer techniques for the in-situ detection of neoplastic change in CLO magnification and high resolution endoscopy chromo-endoscopy auto-fluorescence endoscopy narrow band imaging microscopic tools confocal microscopy multiphoton microscopy in situ molecular analysis FISH spectroscopic analysis fluorescence spectroscopy light scattering spectroscopy optical coherence spectroscopy Raman (inelastic) spectroscopy

Standard endoscopic view of CLO with HGD: no lesion identified

Auto-fluorescence image: purple is abnormal and showed HGD on histology

Narrow band image of same abnormal area: HGD on histology

Chromo-endoscopy with indigo carmine dye-spray: HGD on histology - now for EMR

Management of neoplasia in CLO: the present (in some places) and the future less major surgery trimodal endoscopy means the endoscopist spends the time and not the pathologist!! directed biopsies after such endoscopy and not random Seattle protocols EMR to determine the indication for major surgery more sensible patient-directed management

The limitations of biopsy and advantages of EMR Biopsy EMR Size Size is important Depth - no submucosa in stomach - usually no submucosa in oesophagus Depth - usually submucosa in both stomach and oesophagus Artefact Less artefact (depends on size)

Nice low grade dysplasia in EMR EMR makes histological assessment so much easier

Uses of EMR in the upper GI tract good-sized biopsy of lesion which has provided difficult/equivocal pathological results in repeated standard biopsy procedures complete excision of benign tumorous nodules with ability to judge completeness of excision (cf multiple biopsy chews) (eg GCTs of oesophagus, some endocrine tumours of stomach) complete excision and complete pathological assessment of early gastric cancers to provide definitive pathology and thus a guide to further management of neoplastic lesions of the oesophagus

69F. Long history of achalasia. Nodule close to OGJ. Multiple biopsies at three endoscopic sittings had provided equivocal results. EMR.

Oesophageal EMR Olympus/Keymed inject submucosa and form pseudopolyp, direct snare Cook Duette band EMR band ligation and snare methodology and subsequent therapy likely influences how important margins are

Problem with EMR for early CLO neoplasia Metachronous lesions during FU: 30% Combine with ablative therapy.

Oesophageal EMR Practicalities size and number depends on preferred technique of endoscopist (please make em big!) specimen preparation don t let them leave the band still attached! orientation adequate fixation specimen dissection depends on size we are more interested in the deep margin than peripheral margins

close collaboration with endoscopist and assistant to identify important landmarks resect en bloc if possible keep specimen(s) intact pin out on cork mark margins embed whole specimen(s) for histology

Oesophageal EMR by HB: size matters size number less than 6 mm 15 6-10 mm 76 greater than 10 mm 46

Oesophageal EMR size matters

Oesophageal EMR band artefact

Oesophageal EMRs

Margin fragment

Oesophageal EMR The final diagnostic step and the first therapeutic step Jacques Bergman, 2007

The rationale for EMR in CLO Histological assessment of EMR Intramucosal pathology Submucosal involvement by carcinoma Endoscopic treatment and/or survelliance Referral for surgery

The rationale for EMR in CLO: risk of lymph node metastatic disease intramucosal disease Bergman, 2007 submucosal disease adenocarcinoma 2.0% 24.6% squamous cell carcinoma 3.6% 26% For mucosal disease, the surgical mortality outweighs the risk of metastasis

What are the diagnostic pathological issues in oesophageal EMR?

HGD versus intramucosal carcinoma

Entrapped and submucosal glands mimicking submucosal adenocarcinoma

Reporting oesophageal EMRs is it Barrett s? are there treatment effects? neoplasia diagnosis depth of spread lymphovascular spread peripheral margins status deep margin status

Oesophageal EMR by HB diagnosis and complete excision CLO only 21 Squamous only 9 Gastric mucosa only 7 LGD excised 5 LGD at margins 9 HGD excised 9 HGD at margins 36 IMC excised 4 IMC at margins 12 Ca into SM excised 8 Ca into SM at margins 17 Total 137

Low grade dysplasia at margins

Oesophageal EMR in Gloucestershire we have now done about 600 number of normals reflects endoscopic difficulties benign nodules in CLO, hiatus hernia, etc LGD and HGD often at margins reflects endoscopic difficulties matters less because of subsequent ablative therapy IMC often at margins don t know the implications of this but one suspects that this, too, will be successfully ablated submucosal adenocarcinoma (ironically) more often clear of margins but doesn t matter much as this is an indication for radical surgery

Oesophageal EMR pathology is important, mainly to confirm or refute: the presence of malignancy if present, depth of malignancy margins matter less (but this does depend on subsequent management strategy)

CLO and the double muscularis mucosae and entrapment of dysplastic epithelium Takubo et al, 1991, Lewis et al, 2008

CLO and the double muscularis mucosae Takubo et al, 1991, Lewis et al, 2008

Lewis JT, Wang KK, Abraham SC. Muscularis mucosae duplication and the musculo-fibrous anomaly in endoscopic mucosal resections for Barrett esophagus: implications for staging of adenocarcinoma. Am J Surg Pathol 2008; 32: 566-71. 122 EMR specimens from 100 patients from 1999 to 2006 prolapse changes in 65 (54%) cases & gland entrapment in 67 (56%). 33 cases of IMAC, tumor invaded lamina propria in 10 (30%), inner or single MM in 14 (42%), space between duplicated MM in 5 (15%), and outer MM layer in 4(12%). lymphatic invasion was seen in 2 (10%) cases in which tumor reached the space between MM layers. overstaging of carcinoma in 7% due to misinterpretation of anatomy one case where deep MM was interpreted as muscularis propria.

Staging of adenocarcinoma in EMRs European/Bergmann pt1 m1, m2, m3 Bayreuth/Vieth pt1 m1, m2, m3, m4 spread into the true submucosa remains the most important arbiter for resection need more data on the niceties of intramucosal staging

pt 1 (m1) pt 1 (m2) pt 1 (m3) pt 1 (m4) pt 1 (sm1)

Frequency of vessel permeation in Barrett s adenocarcinoma Invasion n L1 Barrett-Ca pt1 m1 116 1 (0,8%) m2 37 0 m3 19 0 m4 35 1 (2,8%) Barrett-Ca pt1 sm1 30 4 (13,3%) sm2 16 2 (12,5%) sm3 19 5 (26,3%) Vieth et al, 2005

Is it in the submucosa?

retrospective analysis of post-emr biopsies in 33 patients who underwent gastric EMR inflammation (100%), stromal edema (97.0%), foveolar hyperplasia (78.8%), ectatic vessels (66.7%), epithelial atypia (60.6%), increased glandular mitoses (57.6%), epithelial anisonucleosis (54.5%), fibrinopurulent materials (51.5%), ischemia (48.5%), stromal hemorrhage (33.3%), mucin depletion (12.1%), clear cell degeneration (15.2%), and signet-ring cell-like change (6.1%) clear cell degeneration and signet-ring cell-like change in areas of ischaemia (1-16 days post-emr) degenerative glands were usually embedded in a nondesmoplastic stroma and showed anisonucleosis of glandular epithelia. mimics of residual adenocarcinoma, namely clear cell degeneration and signet-ring cell-like change, should be judiciously assessed to avoid unnecessary surgery Mitsuhashi T, Lauwers GY, et al, Am J Surg Pathol 2006; 30: 650-6

Mitsuhashi T, Lauwers GY, et al, Am J Surg Pathol 2006; 30: 650-6

Artefact to the left and carcinoma to the right Mitsuhashi T, Lauwers GY, et al, Am J Surg Pathol 2006; 30: 650-6

The pathology of oesophageal EMR: take home messages the most important assessment is the differentiation of in-situ and intramucosal disease, on the one hand, from submucosal adenocarcinoma, on the other lymphovascular spread is important and is proportional to depth of spread margins may be less important depending on local management strategy duplicated mm, entrapped glands & prolapse changes cause consternation the jury is still out on the utility of intramucosal staging & which system EMRs ease the pathological burden and are generally easier to assess than straggly biopsies don t let them biopsy shortly after an oesophageal or gastric EMR!