ABSTRACT
AIM: We determined the pattern of immunohistochemical expression in invasive adenocarcinoma in endoscopically resected adenoma, its relationship with the risk of residual or recurrent disease and the related factors. METHOD: We included individuals with malignant polyps resected endoscopically in the period 1999-2009. Clinical and endoscopic data were collected. All histological specimens were re-analysed. CD44, matrix metalloproteinase 9 (MMP-9), vascular endothelial growth factor-ß (VEGF-ß), ß-catenin, laminin and cyclooxygenase 2 (COX-2) expression were determined by immunohistochemistry. A multivariate logistic regression was performed to determine variables independently associated with the risk of residual or recurrent disease. RESULTS: One-hundred and fifty-one malignant polyps (114 pedunculated; mean size ± SD=22.61 ± 10.86 mm) were resected endoscopically. Resection was fragmented and incomplete in 26.5% and 8.6% of patients, respectively. Surgical resection was performed on 71 (47%) patients. After a median follow-up of 44 months, residual (n=12) or recurrent (n=6) disease was detected in 17 patients. Conventional histology showed that 32.1% met high-risk histological criteria. Immunohistochemical expression was positive for CD44, MMP-9, VEGF-ß, ß-catenin, laminin and COX-2 in 63.3%, 25.3%, 45%, 38.8%, 79% and 34.5% of specimens, respectively, with no differences between both groups. Variables associated with residual or recurrent disease in the univariate analysis were: nonpedunculated morphology (P=0.07); fragmented (P<0.001) or incomplete resection (P<0.001); margin infiltration (P=0.04); and histological high-risk lesion (P=0.003). Finally, incomplete resection (OR=12.16, 95% CI=3.15-46.98; P<0.001) and histological high risk (OR=4.73, 95% CI=1.33-16.74; P=0.002) were independently associated with the risk of residual or recurrent disease. CONCLUSION: Immunohistochemistry could not predict residual or recurrent disease. Only incomplete excision and histological high risk did so. The factors independently associated were histological high-risk lesion and incomplete resection.
