CD61, CD31 and CD34 Improve Diagnostic Accuracy in Gastric Antral Vascular Ectasia and Portal Hypertensive Gastropathy
M Westerhoff, A Noffsinger, J Miller, J Hart. University of Chicago, Chicago
Background: Portal hypertensive gastropathy (PHG) and gastric antral vascular ectasia (GAVE) have overlapping clinical, endoscopic and histologic features that may make their differentiation difficult. In addition, correlation between the endoscopic impression of PHG or GAVE and histologic findings is often poor. Correct distinction between these two entities is important since their therapies differ. As a result, we sought to determine whether the use of CD31 and CD34 (vascular markers) or CD61 (to highlight platelet thrombi) could aid in the diagnosis and differentiation of these two disorders.
Design: The pathology database was searched for all cases with a histologic diagnosis of GAVE or PHG. In addition, all biopsies with an endoscopic impression of GAVE or PHG (not histologically confirmed) were included as separate groups. Controls consisted of histologically normal antral biopsies with no endoscopic features of GAVE or PHG. All biopsies were stained with antibodies to CD31, CD34 and CD61. CD31 and CD34 stained sections were digitally scanned and analyzed with the use of Chromavision software. CD61 staining was graded as present or absent.
Results: The CD61 stain was positive in all 11 (100%) histologically confirmed cases of GAVE, and in 9 of 17 (78%) cases endoscopically felt to be GAVE. CD61 positivity occurred in 2 of 11 (18%) histologically confirmed cases of PHG, and in 4 of 12 (33%) cases of endoscopically suspected PHG. None of the control biopsies exhibited CD61 staining. These differences were statistically significant in all groups (p<.0001). Review of the H&E slides from all PHG (histologic or endoscopic) cases with positive CD61 stains showed other histologic features allowing their reclassification as GAVE. Microvessel densities were significantly higher in GAVE and PHG compared with controls. Microvascular density in histologically confirmed PHG did not differ significantly from endoscopically suspected PHG. Review of the H&E slides from the endoscopically suspected PHG cases showed frequent active gastritis which appeared to obscure recognition of ectatic vessels.
Conclusions: CD61 immunostains reliably differentiate GAVE from PHG. Microvessel densities are significantly higher in both GAVE and PHG compared to normal antrum. In the absence of features other than vascular ectasia, as occurs in PHG, this increased vascular density may be difficult to recognize on routine H&E sections. Vascular markers such as CD31 or CD34 may make identification of aberrant vessels easier in suspected cases of PHG.
Monday, March 9, 2009 9:15 AM
Platform Session: Section C, Monday Morning