Observational/retrospective studies indicate that prostaglandin-endoperoxide synthase-2 (PTGS2) inhibitors could positively affect colorectal cancer (CRC) patients' survival after diagnosis. To obtain an acceptable cost/benefit balance, the inclusion of PTGS2 inhibitors in the adjuvant setting needs a selective criterion. We quantified the 72 kDa, CRC-associated, glycosylated form of PTGS2 in 100 frozen CRC specimens and evaluated PTGS2 localization by IHC in the same tumors, scoring tumor epithelial-derived and stroma-derived fractions. We also investigated the involvement of interleukin-1 beta (IL1 beta) in PTGS2 induction, both in vitro and in CRC lysates. Finally, we used overall survival (OS) as a criterion for patient selection. Glycosylated PTGS2 can be quantified with high sensibility in tissue lysates, but the expression in both tumor and stromal cells limits its use for predictive purposes. Immunohistochemistry (IHC) analysis indicates that stromal PTGS2 expression could exert a protective role on patient OS. Stromal PTGS2 was prevalently expressed by cancer-associated fibroblasts exerting a barrier function near the gut lumen, and it apparently favored the antitumor M1 macrophage population. IL1 beta was directly linked to gPTGS2 expression both in vitro and in tumors, but its activity was apparently prevalent on the stromal cell population. We suggest that stromal PTGS2 could exert a positive effect on patients OS when expressed in the luminal area of the tumor.