Introduction

Pituitary adenomas (PAs) are common, comprising approximately one third of all intracranial tumors [1]. With the increasing access to cerebral imaging, pituitary tumors are increasingly identified by chance, when an investigation is performed for another reason than a suspicion of a pituitary disease—the definition of a pituitary incidentaloma.

Although usually benign, the Non-functioning pituitary adenomas (NFPAs) represent therapeutic challenges because of their location close to the optic chiasm and nerves, and the proximity to the pituitary gland. NFPAs clinically manifest late in the development, at the time when they cause mass effect on surrounding tissue (visual impairment, headache) or affect pituitary hormonal secretion. Visual field abnormality is the presenting symptoms in 87 % of the patients and headaches in 66 % [2]. This is contrary to hormone-secreting PAs normally detected at an earlier stage because of characteristic signs and symptoms [3]. Approximately, half of the patients with NFPA have pituitary deficiency in at least one of the pituitary axes [2, 4]. An example of a patient course is given in Textbox 1 with constituting illustrations of MR-investigations and growth curve in Fig. 1.

Textbox 1 NFPA patient course
Fig. 1
figure 1

MRI performed routinely after surgery, here illustrated by one-(Fig. 1 a) and nine-(Fig. 1 b) year postoperative MRI investigations. The pictures clearly show tumor growth. Figure 1 c illustrates the patient's growth curve, which is congruent with a logistic growth pattern

The PAs are in most cases sporadic, however up to 5 % are familial adenomas, either as part of the multiple endocrine neoplasia syndrome (MEN1) or familial isolated pituitary adenomas (FIPAs), often caused by mutations in the AIP-gene [5]. Also, an increasing number of incidentalomas are found on MRI or computed tomography (CT), raising questions on how to be handled [6].

The study and treatment of NFPAs are challenging not only due to the location at the skull base and because of the complex functions of the pituitary gland, but also due to the restricted tissue availability and the lack of animal models adequately mimicking human pituitary tumorigenesis.

Epidemiology

The exact prevalence of PAs is difficult to estimate, because of the variation in clinical presentation and the study methods or the registers used. A recent nationwide, observational study in Iceland spanning almost six decades demonstrated an overall prevalence of NFPA for 2012 of 43/100.000 inhabitants, NFPA being the next most prevalent diagnosed pituitary tumor. For the entire period, NFPAs were the most prevalent, comprising 43 % of all diagnosed adenomas. The standard incidence rate for PA increased during the observed period, estimated to be 5.8/100.000 by the end of the period [7]. The incidence rate and distribution seem to be comparable with recent estimates from the Swedish Pituitary Register Study, demonstrating an overall incidence rate of PA of 3.9/100.000 per year [8]. Also here, NFPAs were the most common (54 %), followed by the hormone-secreting adenomas (i.e., prolactinomas (32 %), growth hormone producing (9 %), ACTH producing (4 %), and TSH producing (0.7 %)). Thus, it seems that the prevalence of NFPAs is fourfold higher than previously reported, suggesting that the tumors have an important burden on the Health Care System. Therefore, optimal resource distribution for both clinical care and research activities aiming to improve the outcome for these patients is needed [7, 9, 10].

Classification

Pituitary adenomas are classified into different systems based on the modalities of investigation. Radiologic investigations describe pituitary adenomas according to their size and invasiveness into surrounding structures. Histological investigations sub-classify the tumors by immunohistochemical staining (IHC) and define the level of atypia. The clinical course of the patient will assess the aggressiveness of the tumor [11]. A major challenge in diagnosing, treating, and following these patients is that there seems to be no reliable correlation between the radiological or histological characterization of the tumors and the clinical course.

Pituitary adenomas are also split into clinically functioning and non-functioning adenomas. The latter again divided into silent pituitary adenomas, staining immunohistochemically for one or several pituitary hormones but without hormonal secretion, and null-cell adenomas. Some tumors not manifesting symptoms may, however, have a low but detectable overproduction of hormones; further, challenging the classification of the tumors [12]. Identification of transcription factors to determine adenohypophyseal cell lineage as a supplement to hormone staining in the sub-classification of NFPAs has recently been suggested. This might give a more precise classification, in particular, for the hormone-negative, null-cell, adenomas [13].

WHO published criteria for atypical pituitary adenomas in 2004 according to mitotic activity, Ki-67 index, and p53 immunoreactivity [14]. A high Ki-67 value is associated with greater tumor size and an increased risk of tumor recurrence [15, 16]. Ki-67 is to date the most reliable cell proliferation marker [11]. It has, however, a low accuracy in prognosticating an aggressive clinical course [17], and there is a lack of standardization of method and sampling [11]. In recent years, Trouillas et al. proposed a new histopathological classification system taking size and IHC subtype, in addition to tumor grade based on proliferation rate and invasiveness, into account [17]. They found a significant difference between the grades related to the rate of recurrence/progression at the end of follow-up in their cohort. The effect of grading on recurrence differed between the subtypes of tumors [12].

In search for a reliable biomarker of aggressiveness, growth factors (e.g., FGFR4), proteolytic enzymes (e.g., MMPs), angiogenetic factors (e.g., VEGF, PTTG) in addition to the proliferation factors have been studied. However, more investigations in well-characterized cohorts are needed before these or other new markers can be introduced [11, 15]. A potential marker of aggressiveness is E-cadherin. Epithelial to mesenchymal transition (EMT) is a tissue phenotype modulation, characterized by disruption of cell contacts and enhanced motility. EMT is associated with the invasion and dissemination of cancer cells in several epithelial cancers [18]. E-cadherin is the main cell–cell adhesion contact in epithelial tissue, and down regulation of E-cadherin is the hallmark of EMT, leading to the destabilization of the cellular architecture. Previous studies have shown an association between EMT and aggressiveness of somatotroph and corticotroph pituitary adenomas [1921]. So far, little has been described about growth rate and EMT in NFPAs.

Present and future therapies

Transsphenoidal surgery is the mainstay of treatment for large and symptomatic NFPAs, with an acceptable complication rate [22]. There are great differences between surgical outcomes with reported gross total removal ranging between 83 [23] and 19 % [24]. Postoperative irradiation is effective for residual tumors with growth potential, but at the price of an increased risk of hypopituitarism [2527] and other possible, though more seldom, radiation induced long-term complications [28]. Fractionated radiation therapy or single-fraction radiosurgery (e.g., stereotactic gamma knife treatment) is the most common alternatives of irradiation [27, 29]. Novel techniques in both radiotherapy and radiosurgery are promising in increasing accuracy of the target tissue, while diminishing side effects [30]. Proton therapy has been hypothesized to be efficient while sparing adjacent radiation-sensitive tissue. There is, however, a lack of prospective, randomized controlled studies on the effect and studies on long-term complications [31, 32].

To date, there have been no large RCTs on medical treatment of NFPAs. For the functional pituitary tumors, two drug groups have proven effective, with an acceptable safety profile: dopamine agonists (DA) and somatostatin analogs (SA). Dopamine receptor subtype 2 is expressed in NFPAs both on mRNA and protein level [33]. DAs seem promising in achieving tumor stabilization, as recently summarized in this journal [34]. A phase three RCT addressing the effect of cabergoline (DA) on change in tumor volume is currently recruiting participants [35].

In somatotroph tumors, there is a correlation between the long-term effect of octreotide on IGF-1 and the IHC expression of somatostatin receptor (SSTR) 2 [36]. Octreotide has a high affinity for SSTR2, while the second-generation SA, pasireotide (SOM230) has a high affinity for SSTR1, SSTR2, SSTR3 and the highest affinity for SSTR5 [37]. The expressions of the various SSTRs in the tumors seem to be indicative for the effect of different SAs [38]. NFPAs express SSTRs [39, 40], but the expression of SSTR2 was found to be lower than in somatotroph adenomas [41]. This may partly explain the relatively disappointing results of octreotide. Most studies of octreotide on NFPAs have been of short duration and with relatively few patients [42], but the use of longer-acting octreotide and longer observations did not seem to improve the outcome [43].

Several studies have found a high expression of SSTR3 in NFPAs [40, 41], especially in null-cell adenomas [39] and silent gonadotroph adenomas [39, 44]. With this in mind, the evaluation of pasireotide as an alternative medical treatment of NFPAs is of great interest [45].

Aggressive pituitary adenomas and carcinomas are rare. Temozolomide (TMZ), a lipophilic alkylating agent, is a potential treatment alternative or supplement for this patient group and for tumors where tumor control is not achieved by conventional treatment [30, 46]. Available studies on this matter have included both functioning and non-functioning pituitary adenomas, the latter consisting of only a few cases [46, 47]. There is an ongoing initiative in Europe for collecting information on larger patient cohorts in respect to the use of TMZ [48].

Growth patterns

The progression rate of NFPAs is difficult to estimate. Some NFPAs show a slow growth and necessitate a long observation time, while others act more aggressively and invade the neighboring structures, thereby requiring rapid neurosurgical intervention to prevent long-term impairment on visual field or pituitary deficiency [3]. Information on tumor growth and behavior, before and after surgery, for this group of tumors is sparse [49, 50]. Annual surveillance by MRI for the first five postoperative years, thereafter MRI according to growth, residual tumor and/or clinical signs has been suggested [51]. Annual MRI follow-up has also been considered as cost-effective, although this varies substantially between regions [52]. To date, there are no data, to our knowledge, concluding when the right time to terminate surveillance of NFPAs neither after surgery nor conservatively treated [5356]. Reddy et al. found that more than 20 % of relapses after surgery happened more than 10 years after the first operation. They also found a significant increase in rate of relapse if the residual tumor was present after surgery, and if the residual tumor was extra- rather than intra-sellar [24]. Some studies have found invasion of cavernous sinus to be predicative of future recurrence of NFPAs, though the results have been conflicting [54, 55, 5760].

Information and understanding of adenoma growth kinetics (i.e., linear, logistic or exponential growth) collaborated with data about tumor biology would substantially enhance our clinical acquaintance of tumor behavior and help to adjust and individualize MRI scanning protocols and follow-up (Fig. 2) [61]. The so far only published systematic study investigating specific growth kinetics of NFPAs, although with a limited number of patients (n = 15), concluded that the linear growth model is not suitable and that an exponential or a logistic growth model is more accurate to describe the growth kinetics of NFPAs [61].

Fig. 2
figure 2

Two different growth patterns being applicable for pituitary adenomas [61]. Applying TVDT and thereby the exponential growth pattern will be the most conservative approach. An estimate of TVDT or an indication of logistic growth would markedly reduce the number of investigations as compared current management of patients

The different subgroups of NFPAs show different clinical behaviors, and some are known to be more aggressive than others [3, 62]. Balogun et al. found in their cohort of null-cell and silent gonadotroph adenomas a difference in growth behavior prior to and after surgery between the two subgroups [63]. Brochier et al. showed a significant difference in relapse rate between immunohistochemical subgroups [64]. Thus, it seems reasonable that non-functioning pituitary adenomas are a heterogeneous group of tumors with different growth patterns, and more information is needed to better describe the behavior of these tumors.

Conclusion

Non-functioning pituitary adenomas are a heterogeneous group of tumors, with a variety of clinical manifestations and courses. To date, there is no known credible marker of aggressive growth or behavior, making the monitoring of the patients a challenge. Some studies show a latent growth potential of these tumors, and suggest a prolonged follow-up. More information on growth patterns with constituent markers are needed to tailor the care of this patient group. Studies characterizing the membrane receptors of NFPAs have shown promising results, which may give rise to the development of medical treatment.