Pituitary adenomas (PAs) account for 10%–25% of all intracranial neoplasms, have different morphology and expansion directions, and about 35% are invasive [1, 2]. PAs usually originate from the anterior pituitary gland (adenohypophysis). The adenohypophyseal tumors are recently renamed pituitary neuroendocrine tumors (PitNET), which may present various behaviors, such as invasive, aggressive and malignant with metastases. They are classified into seven morphofunctional types and three lineages: lactotroph, somatotroph and thyrotroph (PIT-1 lineage), corticotroph (T-PIT lineage) or gonadotroph (SF-1 lineage), null cell or immunonegative tumor and plurihormonal tumors [3]. The World Health Organization (WHO) 2017 classification of PitNET is mostly based on immunohistochemistry for pituitary hormones, pituitary-specific transcription factors, and other immunohistochemical markers commonly used in pathology practice [4, 5].

Transsphenoidal surgery remains the first-line approach for most PAs (except for prolactinoma) resection because of its minimal invasiveness and low morbidity [6]. The most frequent complications associated with transsphenoidal pituitary surgery are normal pituitary tissue damage and cerebrospinal fluid leakage, leading to partial or complete insufficiency of one or more hormonal axes (hypopituitarism) [7]. PAs originate from the adenohypophysis and combine with the normal pituitary tissue by deformation and displacement. The average adult gland measures 6 mm superior-inferior $\times$ 9 mm anteroposterior $\times$ 13 mm medio-lateral and weighs 0.6 grams [8, 9]. Magnetic resonance imaging (MRI), especially gadolinium-enhanced MRI, has become the definitive radiological procedure for diagnosing and following pituitary tumors. It provides a detailed anatomic view of the sellar region [10]. The residual adenomas and normal pituitary parenchyma can be well recognized on MRI [11].

However, few reports have investigated the pre-and post-operative MRI appearance of the normal pituitary tissues in patients with pituitary macroadenomas, including size, shape, location, and its clinical correlation [12]. The normal pituitary tissues, including adenohypophysis and neurohypophysis, can be identified easily on MRI T1-weighted images (T1WI) in healthy adults or microadenoma patients [13]. Previous studies are mainly focused on the posterior pituitary [14, 15]. Our previous studies suggest that neurohypophysis exhibits high signal intensity on MRI T1WI, related to the neuro-secretory granules containing antidiuretic hormone (ADH). As central diabetes insipidus (DI) is characterized by reducing ADH release, this entity may be reflected by reducing the posterior pituitary bright spot (PBBS) signal [16, 17].

The purpose of this work was to investigate the pre-and postoperative magnetic resonance imaging of normal pituitary tissues in the context of transsphenoidal resection of pituitary macroadenomas. The morphology and remodeling of residual normal pituitary tissues following transsphenoidal resection were observed on contrast-enhanced MRI, and their relationship with PAs was analyzed. The appearance of pituitary stalk was also analyzed as well as its clinical significance.

Until now, MRI studies of pituitary adenomas have mainly focused on the preoperative diagnosis and evaluation of the extent of tumor resection, while few studies of normal structures such as pituitary stalks and residual pituitary tissues have been done [20, 21, 22]. In this study, MRI features of pituitary stalk and residual pituitary tissue before and after transsphenoidal surgery were analyzed, and some significant results were observed. As we all know, pituitary stalk connects hypothalamus to the pituitary gland and plays an important role in maintaining the normal physiological function of the pituitary gland [23, 24]. In pituitary adenoma patients, pituitary stalk maybe relatively thickened or compressed by the pituitary adenomas, and even be curled up into “C” or “S” type, some of which are indistinguishable from the saddle diaphragm. The pituitary stalk is thinned, deflected and pulled by the tumor, which may affect the blood supply of the pituitary portal system and thus change the pituitary function. Furthermore, surgery or tumor resection can also cause traction or change in position of the pituitary stalk, thus affecting the function of pituitary stalk, resulting in postoperative complications, such as postoperative diabetes insipidus.

As documented in this paper, pituitary stalk was thickened or compressed on preoperative MR images, and gradually recovered to normal during postoperative period. It tended to be in the middle position and its length increased gradually until 4–6 months after operation. Among the 57 patients with preoperative pituitary stalk deviation, 55 of the pituitary stalk deviations recovered in 1 week after surgery, and 30 cases were almost in the middle in 4–6 months after operation, with the other cases get better in varying degrees. These results showed that the pituitary stalk tended to recover to the middle position after pituitary adenoma operation. After operation, the sellar septum descended, the pituitary stalk was relieved of compression, the diameter of pituitary stalk returned to a relatively normal state, and the pituitary stalk curled up and bent out gradually. Of course, excessive tamponade of the intrasellar lumen during the operation may result in more pronounced ascending displacement of the pituitary stalk in the early postoperative period. But generally, 4–6 months after the operation, after the absorption of gelatin sponge and effusion, the sellar septum was further decreased, and the pituitary stalk was gradually prolonged and recovered to be in the middle.

The posterior pituitary is usually hyperintense on T1WI, commonly referred to as the “posterior pituitary bright spot, PPBS”, which is currently considered to be a short T1 signal of arginine vasopressin stored in the posterior pituitary [14, 15]. As tumor compression releases after operation, some cases of PPBS(–) on preoperative MRI may show postoperative PPBS(+). It reflects the rebuilding of the neurohypophysis pathway and recovery of posterior pituitary function. To some extent, the signal intensity of neurohypophysis is related to the compression that it receives. The signal intensity of PPBS may also be associated with the amount of ADH stored in the posterior pituitary. Higher compression may reduce the transportation and storage of ADH in neurohypophysis [25].

Preoperative diabetes insipidus is rare in patients with PA [26]. The incidence of postoperative DI is 21.4%, consistent with the reported incidence of 0.5–25% in PAs [26, 27]. As reported in the literature, the absence of hyperintensity in the posterior pituitary may be correlated with postoperative DI [28]. However, some patients with PPBS(–) do not develop DI, some patients with idiopathic DI still have PPBS(+), which leads to doubts about using the absence of PPBS as the diagnostic criterion for central DI [29]. Patients without pre-operative PPBS had a higher incidence of postoperative DI, suggesting preexisting posterior pituitary dysfunction. In general, because postoperative diabetes insipidus may be related to the disorder of antidiuretic hormone secretion and storage, we should pay more attention to the protection of the pituitary stalk and posterior lobe during surgery. As we all know, pituitary stalk is composed of the blood vessels and the axons of the large neurons in the hypothalamic nucleus. The axons have no perineuria and adventitia, and may be damaged or even broken during the pull-down process accompanied by diaphragm subsidence, resulting in the failure of ADH to pass down through the pituitary stalk, showed with diabetes insipidus and water-electrolyte disturbance.

Zada et al. [30] showed that growth hormone-secreting adenomas tend to grow below the sella, and nonfunctioning PAs tend to extend above the sella. Di Maio S et al. [20] studied 79 cases of MRI of PAs before surgery and classified the displacement of residual normal pituitary tissue as superior, superolateral or lateral in location. However, these authors did not find any statistically significant difference between these preoperative normal pituitary tissue locations and tumor types. We found that the location of normal pituitary tissue on preoperative MRI was significantly different between patients with different immunohistochemical types. Normal pituitary tissues were likely located above the adenomas in somatotroph adenoma patients, while in gonadotroph adenoma patients (almost clinically nonfunctioning adenoma), RPT was more likely to be located on one side of the tumor (parasellar). This observation is consistent with the report by Zada et al. [30]. The downward growth expansion pattern of growth hormone (GH) is associated with the presence of GH cells in the anterior flank of the anterior pituitary gland, which may explain why GH adenomas often first appear in this region on MRI. Additionally, patients with gonadotroph adenoma accounted for 31.2%, similar to previous findings [31, 32].

In our findings, 61.6% of the 112 patients showed normal pituitary tissue on preoperative enhanced MRI, while 95% on the postoperative enhanced MRI images. This change may be due to the re-expansion and repositioning of normal pituitary tissues after surgery. The re-expansion degrees are related to the size and location of adenomas. Meyrignac et al. [33] showed that nearly all normal pituitary tissues in patients with PAs of diameter more than 20 mm were pushed out of the sella. Adenomas that expand superiorly tend to push normal pituitary tissues above the sella. The normal pituitary tissues in about 49% of nonfunctioning adenomas are pushed outside the sella due to the large size of the tumor, and 76% of secreting adenomas push the normal pituitary tissues aside outside the sella [34]. Preoperative evaluation of cavernous sinus invasion has an important effect on surgery and prognosis. However, the medial wall of the cavernous sinus cannot generate its image in T1WI at all times [35]. Pituitary tissues with lateral displacement can indirectly indicate that ipsilateral cavernous sinus is not affected by tumor invasion [36] because it is unlikely that adenomas across normal pituitary tissues invade cavernous sinus. Consistent with this, we showed that patients with lateral normal pituitary tissues before surgery did not have ipsilateral cavernous sinus invasion by tumors.

Postoperative dynamic MR imaging was used not only for early detection of residual or recurrent pituitary adenomas, but also for evaluation of postoperative RPT recovery. Early postoperative MRI (1 week after operation) can roughly judge the degree of tumor resection and provide imaging reference for subsequent follow-up. On postoperative MRI, RPT showed obvious “re-tensioning”, and on T1 enhanced imaging, the pituitary gland was obviously enlarged. While in the 4–6 months after operation, with the absorption of gelatin sponge, the septum sellae decreased gradually, RPT was reshaped, and some patients could appear partial empty sellae on MRI. The residual tumor is usually located in suprasellar or bilateral cavernous sinus, showing low signal intensity on T1WI, high signal intensity on T2 weighted imaging (T2WI) and weak enhancement on T1WI. In this study, 71 patients (63.4%) underwent total tumor resection and 41 patients (36.6%) underwent residual resection, which was approximately the same as that reported previously [30]. The primary cause of tumor remnant was the invasion of cavernous sinus, followed by the obvious suprasellar extension of the tumor.

Understanding postoperative pituitary remodeling is helpful to evaluate postoperative pituitary recovery. RPT was often compressed by the tumor and was usually located in the periphery of the tumor (continuous or discontinuous) in a flaky manner, which can generally be distinguished on enhanced MRI [34]. Preoperative evaluation of the RPT location was helpful to identify and protect the normal pituitary tissue during the operation and reduce the incidence of postoperative hypopituitarism. The pituitary stalk was also used to help locate pituitary tissue, which has a high signal intensity on enhanced MRI.

Our research showed that the postoperative pituitary volume recovery was observed in most cases, and flaky RPT may present volume enlargement in some cases even in the early postoperative period. During postoperative follow-up, attention should be paid to distinguish pituitary volume recovery from residual tumors or organized tissues. On MRI T1WI enhanced scan, normal pituitary was usually enhanced significantly, while residual tumors and organized tissues showed weak enhancement due to poor blood supply. Di Maio S et al. [20] suggest that the period between the 4th–6th month after surgery is ideal for evaluating pituitary MRI, preventing the influence by postoperative tumor residuals or hemostatic materials. At the early stage after surgery, residual hemostatic materials may cover the remaining pituitary tissues [13, 37]. Our study also found that the postoperative RPT volume recovery is not related to the patient’s age, but negatively related to the size of the tumor. The larger the tumor, the more difficult the postoperative pituitary volume to recover, possibly because the larger the pituitary tumor, the more severe the pituitary tissue compression, the longer the compression time, resulting in the postoperative pituitary morphology more difficult to recover. Pituitary tissues with lateral location have better recovery compared with other types. The results suggest that suprasellar extension of PA may cause normal pituitary tissue (adenohypophysis and neurohypophysis) to receive more compression. Di Maio S et al. [20] showed that the degree of pituitary morphological remodeling was related to pituitary function after pituitary adenoma surgery.

In conclusion, our findings demonstrate that for the first time that pituitary stalk tends to be in the middle position and its length increased gradually until 4–6 months after operation. Somatotroph adenoma patients tend to have normal pituitary tissues with superior and superolateral displacement before surgery, while gonadotropin adenoma patients tend to have normal pituitary tissue with lateral displacement. Postoperative pituitary tissue remodeling and volume recovery were related to tumor size. The larger the tumor, the worse the pituitary tissue recovery.

PPBS, posterior pituitary bright spot; MRI, Magnetic resonance imaging; T1WI, T1-weighted images; ADH, antidiuretic hormone; DI, diabetes insipidus; RPT, residual pituitary tissues; PGV, Pituitary gland volume.

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

SW conceived and designed the research study; SW, DX, SZ, KL, LZ and LW performed the research; SZ and DX analyzed the data; DX and SW wrote the paper. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

All procedures were approved by the Ethics Committee of Fujian Medical University. The Ethics approval number is EC of FGH 2012-017. Written informed consent was obtained from every patient.

Not applicable.

This work was supported by Fujian Provincial Natural Science Foundation of China [grant number 2015D014] and Fujian Provincial Key Project of Science and Technology Plan of China [grant number 2018Y0067].

The authors declare no conflict of interest.