Digital analysis of hormonal immunostaining in pituitary adenomas classified according to WHO 2017 criteria and correlation with preoperative laboratory findings

Abstract

OBJECTIVE The authors sought to evaluate clinical and laboratory data from pituitary adenoma (PA) patients with functioning PA (associated with acromegaly [n = 10] or Cushing disease [n = 10]) or nonfunctioning PA (NFPA; n = 10) that were classified according to 2017 WHO criteria (based on the expression of the transcription factors pituitary-specific positive transcription factor 1 [Pit-1], a transcription factor member of the T-box family [Tpit], and steroidogenic factor 1 [SF-1]) and to assess the immunostaining results for growth hormone (GH) and adrenocorticotropic hormone (ACTH) in the corresponding tumors. METHODS Clinical and laboratory data were collected retrospectively. The percentage of tumoral cells positive for Pit-1, Tpit, or SF-1 was assessed and ImageJ software was used to evaluate immunopositivity in PAs with 2 different antibodies against GH (primary antibody 1 [AbGH-1] and primary antibody 2 [AbGH-2]) and 2 different antibodies against ACTH (primary antibody 1 [AbACTH-1] and primary antibody 2 [AbACTH-2]). RESULTS Cells with positive Pit-1 staining were more frequently observed in lesions from patients with acromegaly (acromegaly group) than in lesions from patients with Cushing disease (Cushing group; p < 0.001) and those from patients with NFPA (NFPA group; p < 0.001). The percentage of Tpit-positive cells was higher in the Cushing group than in the acromegaly (p < 0.001) and NFPA (p < 0.001) groups. No difference was detected regarding SF-1 frequency among all groups (p = 0.855). In acromegalic individuals, GH immunostaining levels varied depending on the antibody employed, and only one of the antibodies (AbGH-2) yielded higher values in comparison with the values for NFPA patients (p < 0.001). For all of the antibodies employed, no significant correlations were detected between GH tissue expression and the laboratory data (serum GH vs AbGH-1, p = 0.933; serum GH vs AbGH-2, p = 0.853; serum insulinlike growth factor-1 [IGF-1] vs AbGH-1, p = 0.407; serum IGF-1 vs AbGH-2, p = 0.881). In the Cushing group data, both antibodies showed similar ACTH tissue expression, which was higher than that obtained in the NFPA group (p < 0.001). There were no significant associations between ACTH immunohistochemical findings and ACTH serum levels (serum ACTH vs AbACTH-1, p = 0.651; serum ACTH vs AbACTH-2, p = 0.987). However, ACTH immunostaining evaluated with AbACTH-1 showed a significant correlation with 24-hour urinary cortisol (24-hour cortisol vs AbACTH-1, p = 0.047; 24- hour cortisol vs AbACTH-2, p = 0.071). CONCLUSIONS Immunostaining for Pit-1 and Tpit accurately identified lesions associated with acromegaly and Cushing disease, respectively. Conversely, SF-1 did not differentiate NFPA from lesions of the other two groups. Regarding hormonal tissue detection, results of the current investigation indicate that different antibodies may lead not only to divergent immunohistochemical results but also to lack of correlation with laboratory findings. Finally, PA classification based on transcription factor expression (Pit-1, Tpit, and SF-1), as proposed by the 2017 WHO classification of pituitary tumors, may avoid the limitations of PA classification based solely on digital immunohistochemical detection of hormones.