**4. Endoscopic recognition of dysplasia**

150 Gastrointestinal Endoscopy

Barrett's oesophagus is most often identified incidentally in patients who are undergoing an upper endoscopy for investigation of reflux symptoms. Barrett's oesophagus has a classical endoscopic appearance of 'salmon pink' columnar mucosa arising proximally from the oesophago-gastric junction (OGJ), often with characteristic 'tongue' extensions. There may also be readily identifiable islands of columnar mucosa. Following endoscopic recognition, the extent of proximal extension above the OGJ should be measured and documented, taking care to accurately identify any sliding hiatus hernia which may confuse this measurement. The diagnosis must then be confirmed / corroborated histologically by multiple pinch biopsies of the affected segment. When biopsies are obtained it is crucial that they originate from the oesophagus and that their site is recorded as accurately as possible. The 'Prague C and M criteria', defined by an International Working Group on Barrett's oesophagus, offers a validated method of classifying Barrett's based on its endoscopic appearance. (Sharma et al., 2006b) The extent of circumferential involvement in centimetres from the OGJ should be recorded, as should the maximum length of the Barrett's segment

Difficulties arise in diagnosis particularly in 'ultra-short' segment Barrett's oesophagus. The original description of Barrett's oesophagus was of columnar metaplasia extending for at least 3cm from the OGJ. Although the risk of malignant progression is greater in long Barrett's segments (>8cm), it is now recognised that shorter lengths, even below 3cm have malignant potential. (Hirota et al., 1999; Schnell et al., 1992; Sharma et al., 1997; May et al., 2002) However, what appears endoscopically to be a short segment of Barrett's oesophagus in the distal oesophagus or an irregular z-line may in fact represent intestinal metaplasia of the gastric cardia known as cardia intestinal metaplasia (CIM). (BSG Working Party, 2005) This can lead to misclassification of CIM as short segment Barrett's. For this reason, the endoscopist has a crucial role in defining the exact position from which biopsies are taken to

Dysplasia is defined as "an unequivocal neoplastic alteration of epithelium which has the potential to progress to invasive malignancy but remains confined within the basement membrane of the epithelium within which it arose." (Shaheen and Ransohoff, 2002; Riddell et al., 1983) Dysplasia is classified as either low grade (LGD), or high grade (HGD) (often also termed high-grade intraepithelial neoplasia HGIN)), based on its histological appearances. As already described, HGD has a higher malignant potential than LGD and malignant transformation classically occurs through a stepwise progression of pathology from metaplastic Barrett's oesophagus, to LGD, then HGD, and finally invasive

Understanding the pathogenesis and natural history of Barrett's oesophagus is key to understanding the malignant potential and clinical significance of the various dysplastic stages. Surveillance studies have shown that the risk of developing adenocarcinoma varies between 0.4% and 1% per year (in the US and UK respectively). However, it is clear from cohort studies that not all Barrett's oesophagus progresses to dysplasia. In fact, in most longterm studies fewer than 10% of patients show evidence of progressive disease. (Schnell et al., 2001a) Patients with Barrett's oesophagus are thought to have a lifetime risk of developing oesophageal adenocarcinoma of 3-14% (approximately 0.5-1% per year following diagnosis).

**2. Endoscopic detection of Barrett's oesophagus** 

(including tongues of Barrett's but excluding isolated 'islands').

**3. Definition and clinical significance of Barrett's dysplasia** 

prevent misdiagnosis.

adenocarcinoma.

Endoscopic recognition of dysplasia within Barrett's oesophagus is difficult and unreliable, even for skilled endoscopists. A rigorous biopsy protocal such as that recommended by the British Society of Gastroenterologists (Box 1) is therefore necessary to identify any dysplastic change so that apropriate surveillance / endotherapy / surgery can be considered. Sites of biospy must be accurately recorded and when possible macroscopic lesions should be classified using the Paris classification (Box 2).
