Introduction Vasculogenic mimicry (VM) refers to the formation of vascular channels within genetically dysregulated tumor cells. It occurs de novo without the presence of an endothelial cell lining. Since its introduction in 1999, it has been observed in several tumor types and is proposed to be an alternative means of providing blood perfusion to tumors. Pituitary tumors are generally slow growing, benign adenomas. To date, VM in pituitary adenomas has not been described. Herein, we studied whether these channels occur in clinically non‐functioning pituitary adenomas (NFPA). Materials and Methods We studied the presence of VM in 49 randomly selected cases of surgically‐removed clinically NFPA using CD34 and Periodic Acid‐Schiff (PAS). Tumors were divided into 2 groups (19 recurring and 30 non‐recurring). 10 surgically removed and 5 autopsy‐obtained non‐tumorous pituitaries from patients with no endocrine diseases and no hormone treatment were also included. To demonstrate vascular change, CD34‐PAS double immunostaining on paraffin‐embedded sections was used. Slides were assessed for the presence of CD34‐negative and PAS‐positive channels indicating that they were not lined by endothelial cells. Immunohistochemical findings were evaluated under low and high magnification. Results Immunopositivity for CD34 demonstrated endothelial cells lined blood vessels, which were localized in all 49 cases of clinically NFPA. The PAS‐ positive pattern, which stained sub‐endothelial matrix, was also visible in all cases. Slides were assessed for the presence of CD34‐negative and PAS‐positive channels indicating that they were not lined by endothelial cells. This staining pattern suggestive of VM was observed in 22/49 (45%) of the clinically NFPA. The amount of staining pattern varied both from case to case and within groups. No statistically significant correlation was found between the presence of this staining pattern within each group and the gender of the patients. An intriguing finding was the correlation between staining pattern and age. In 22 cases studied where the staining was identified, 14 cases (64%) were in the age range between 52–59 yrs while the remainder of the cases were equally distributed between the age range 25–75 yrs. In 27 cases where the staining pattern was not identified, 12 cases (44%) were in the age range between 51–57 yrs and 22/27 (81%) of the negative cases in the age range between 31–57 yrs. This was an intriguing finding since cases studied were randomly selected. Conclusion This finding carries potential clinical implications with respect to patient response to therapy as well as clinical course. More work is needed to prove that the presence of sub‐endothelial matrix without endothelial lining represents VM. Support or Funding Information Acknowledgement Authors are grateful to the Jarislowsky and Lloyd Carr‐Harris Foundations for their generous support.