Celiac disease is associated with an increased risk for non-Hodgkin lymphoma, especially of T-cell type and primarily localized in the gut. However, the association does not represent a great enough risk to justify early mass screening for celiac disease.

In comparison with CrD-SBC, CD-SBC patients harbour MSI and high TILs more frequently and show better outcome. This seems mainly due to their higher TIL density, which at multivariable analysis showed an independent prognostic value. MSI/TIL status, KRAS mutations and HER2 amplification might help in stratifying patients for targeted anti-cancer therapy.

No abstract

Non-familial small bowel carcinomas are relatively rare and have a poor prognosis. Two small bowel carcinoma subsets may arise in distinct immune-inflammatory diseases (celiac disease and Crohn's disease) and have been recently suggested to differ in prognosis, celiac disease-associated carcinoma cases showing a better outcome, possibly due to their higher DNA microsatellite instability and tumor-infiltrating T lymphocytes. In this study, we investigated the histological structure (glandular vs diffuse/poorly cohesive, mixed or solid), cell phenotype (intestinal vs gastric/pancreatobiliary duct type) and Wnt signaling activation (β-catenin and/or SOX-9 nuclear expression) in a series of 26 celiac disease-associated small bowel carcinoma, 25 Crohn's disease-associated small bowel carcinoma and 25 sporadic small bowel carcinoma cases, searching for new prognostic parameters. In addition, non-tumor mucosa of celiac and Crohn's disease patients was investigated for epithelial precursor changes (hyperplastic, metaplastic or dysplastic) to help clarify carcinoma histogenesis. When compared with non-glandular structure and non-intestinal phenotype, both glandular structure and intestinal phenotype were associated with a more favorable outcome at univariable or stage- and microsatellite instability/tumor-infiltrating lymphocyte-inclusive multivariable analysis. The prognostic power of histological structure was independent of the clinical groups while the non-intestinal phenotype, associated with poor outcome, was dominant among Crohn's disease-associated carcinoma. Both nuclear β-catenin and SOX-9 were preferably expressed among celiac disease-associated carcinomas; however, they were devoid, per se, of prognostic value. We obtained findings supporting an origin of celiac disease-associated carcinoma in SOX-9-positive immature hyperplastic crypts, partly through flat β-catenin-positive dysplasia, and of Crohn's disease-associated carcinoma in a metaplastic (gastric and/or pancreatobiliary-type) mucosa, often through dysplastic polypoid growths of metaplastic phenotype. In conclusion, despite their common origin in a chronically inflamed mucosa, celiac disease-associated and Crohn's disease-associated small bowel carcinomas differ substantially in histological structure, phenotype, microsatellite instability/tumor-infiltrating lymphocyte status, Wnt pathway activation, mucosal precursor lesions and prognosis.