Optimizing the Detection of Fundic Gland Polyp Dysplasia in Subjects with Famililal Adenomatous Polyposis

Abstract

Purpose: Dysplasia in fundic gland polyps (FGPs) is common in familial adenomatous polyposis (FAP) and gastric cancer is reported arising from FGPs in FAP. Optimal biopsy protocols are elucidated for dysplasia detection in IBD and Barrett's, but have not been investigated for dysplasia detection in FGPs in FAP. Our aim was to determine the most effective biopsy strategy for the detection of dysplasia in FAP FGPs. Methods: A systematic protocol was utilized to biopsy FGPs from consecutive pts with FGPs undergoing routine endoscopic surveillance for FAP. FGP number (1–20, 21–30, >30) and size (1–4 mm, 5–10 mm, >10 mm) was recorded. The stomach involved with FGPs was visually divided into equal segments: proximal, middle, distal. Equal numbers of FGPs were biopsied from each segment. If 1–20 FGPs, 9 biopsies obtained. If 21–30 FGPs, 15 biopsies obtained. If > 30 FGPs, 21 biopsies obtained. Directed biopsies were obtained from large (>1 cm) and irregular appearing FGPs. One gastrointestinal pathologist blinded to the endoscopic findings interpreted all histology. Results: 66 subjects (48% female) with a mean age of 43.6 yrs. 8% reported a family history of gastric cancer. Dysplasia was detected in 41% (2 subjects with high grade dysplasia (HGD), 25 subjects with low grade dysplasia (LGD)). 12% (3/25) had LGD in all three sets of polyps biopsied; 32% (8/25) had LGD in FGPs biopsied from the proximal segment; 52% (13/25) had LGD in FGPs from the middle segment, and 48% (12/25) with LGD in FGPs from the distal segment. Biopsies from proximal and middle segments missed 36% of subjects with LGD; biopsies from middle and distal segments missed 20% of LGD. 8% (2/25) of subjects with LGD were detected on directed biopsies alone. 100% (2/2) of subjects with HGD were detected in directed biopsies of large or irregular FGPs. 89% of subjects with dysplasia were detected by either directed biopsies of large or irregular FGPs or random biopsies from the distal 2/3 of the stomach. Conclusion: LGD was not uniformly distributed among FGPs. HGD was detected only in directed biopsies. The yield for dysplasia detection was greatest when directed biopsies of large (>1 cm) and irregular FGPs were combined with random biopsies obtained from FGPs in the distal 2/3 of involved mucosa.