Pituitary adenomas account for 10%–20% of all intracranial tumors, with an increasing incidence in the United States.^4,6,8 A large number of adenomas are clinically nonfunctioning pituitary adenomas (NFPAs) without clinical or laboratory evidence of excess hormone secretion, but with symptoms related to mass effect of the tumor, including headaches and visual field defects, that frequently prompt their discovery.^5,17 NFPAs are considered to be the most common type of adenoma after prolactinomas, and are thought to represent 14%–54% of all pituitary adenomas, although estimates vary depending on the study.^4,17,21

Pathological analysis of NFPA tissue enables NFPAs to be divided into silent pituitary adenomas (SPAs) versus null-cell adenomas (NCAs), a distinction that was formalized in 2017 when the WHO criteria were modified to include these entities. SPAs are adenomas with evidence of hormone or transcription factor staining on final pathology results, but otherwise no clinical or biochemical evidence of significant hormone excess.^6,16 Thus, 90%–99% of NFPAs are found to be SPAs following final pathology results, frequently depending on the number of pathological stains performed.^6,16 All types of adenomas can be silent, including somatotroph-, lactotroph-, thyrotroph-, corticotroph-, gonadotroph-, and PIT-1–positive adenomas, with biochemical expression of hormones seen occurring on a spectrum rather than in a binary manner.^1,6,13,14 NCAs are NFPAs that on final pathology findings do not stain positive for any pituitary hormones or transcription factors.^6,14 In 2015, Nishioka et al. demonstrated the importance of including the transcription factor staining in addition to hormone staining in the differentiation of NCAs and SPA subtypes.¹⁶ Indeed, this was especially true for silent gonadotroph adenomas (SGAs), with 79/119 hormone-negative adenomas staining positive for SF-1, a marker of gonadotroph lineage. This study contributed to the 2017 WHO classification scheme differentiating between SF-1–positive SPAs and NCAs, although comprehensive transcription factor staining is still not available in many pathology laboratories.¹⁴

Classification of the SPA subtype is critical given the potential clinical ramifications for patients. Studies have highlighted its importance by showing that silent corticotroph adenomas, for example, tend to present with increased cavernous sinus invasion and when patients are younger.^11,16 Across all subtypes, SPAs are more aggressive relative to their functional adenoma counterparts.⁶ Given that many of the tumors previously considered to be NCAs are in fact SF-1–positive SGAs, it is vital to consider the differences between these 2 groups.^2,6 In their cohort of 31 NCAs and 63 SGAs, Balogun et al. detailed differences in tumor behavior between the 2 subtypes, demonstrating higher MIB-1 staining and a positive correlation between pre- and postoperative tumor volume doubling times (TVDTs) in NCAs.² Their study was limited to behavioral characteristics of the 2 adenoma types and did not assess clinical outcomes aside from resection rates.¹² There remains a severe paucity of data investigating the difference between NCAs and SGAs with regard to clinical outcomes.

Our study sought to describe how patient characteristics as well as short- and long-term clinical outcomes, including the need for adjuvant radiation therapy and incidence of diabetes insipidus, differ between SGAs and NCAs. We leveraged a database of 1166 patients from a single institution with pituitary adenomas and extensive follow-up periods spanning 2012 to 2019 to assess any potential differences between the groups.

Methods

Results

Comparison of Long-Term Outcomes

We next compared long-term outcomes between our adenoma subgroups (Table 3). NCAs were more likely to receive adjuvant radiation (23.4%) when compared to SGAs (9.4%, p = 0.002) (Table 3, Fig. 1), with most of this seemingly driven by a higher incidence of adjuvant radiation therapy, rather than radiosurgery (12.1% vs 3.4%, p = 0.007) (Table 3). There was no significant difference between NCAs and SGAs in the incidence of adjuvant radiation provided for residual tumor following resection (15.0% vs 7.4%, p = 0.052). However, NCAs received significantly more adjuvant radiation for recurrence than SGAs (8.4% vs 2.0%, p = 0.017) (Table 3). Indeed, NCAs were significantly more likely to recur than SGAs (9.3% vs 2.7%, p = 0.021). To further explore this, we then generated a Kaplan-Meier plot for progression-free survival for the 2 groups (Fig. 2). The Kaplan-Meier plot demonstrated a significant difference on log-rank analysis (χ² = 4.73, p = 0.030). A subsequent backward likelihood ratio Cox regression analysis also demonstrated significantly higher recurrence in NCA tumors versus SGAs (HR 3.64, 95% CI 1.07–12.30; p = 0.038). Other variables that were significantly independently associated with recurrence included cavernous sinus invasion (HR 3.97, 95% CI 1.04–15.14; p = 0.043) and AP dimension of the tumor (HR 2.23, 95% CI 1.09–4.59; p = 0.030) (Table 4).

TABLE 3.

Long-term outcomes in 1166 patients with pituitary adenomas

Variable	NCA	SGA	p Value
Any adjuvant radiation	25 (23.4%)	14 (9.4%)	0.002
Radiosurgery	12 (11.2%)	9 (6.0%)	0.137
Radiotherapy	13 (12.1%)	5 (3.4%)	0.007
Indication: recurrence	9 (8.4%)	3 (2.0%)	0.017
Indication: residual tumor	16 (15.0%)	11 (7.4%)	0.052
Tumor recurrence	10 (9.3%)	4 (2.7%)	0.021
Time to recurrence, mos (range)	15.9 (5–41)	13.5 (6–25)	0.705
New postop diabetes insipidus	13 (12.1%)	7 (4.7%)	0.028
Future reop	5 (4.7%)	7 (4.7%)	0.993
Death	1 (0.9%)	2 (1.3%)	0.765

Incidence of adjuvant radiation and new postoperative diabetes insipidus in NCAs versus SGAs. Patients with NCAs are more likely to develop new postoperative diabetes insipidus than patients with SGAs (4.7% vs 12.1%, *p = 0.028). Patients with NCAs are also more likely to require adjuvant radiation therapy (23.4% vs 9.4%, **p = 0.002).

Kaplan-Meier curve of recurrence-free survival in NCAs versus SGAs. The Kaplan-Meier curve of recurrence-free survival in the SGA and NCA patient population demonstrates reduced recurrence-free survival in patients with NCAs on log-rank analysis (χ² = 4.73, p = 0.030).

TABLE 4.

Cox regression of variables associated with tumor recurrence

Variable	HR (95% CI)	p Value
NCA pathology	3.64 (1.07–12.30)	0.038
Cavernous sinus invasion	3.97 (1.04–15.14)	0.043
AP dimension, cm (range)	2.23 (1.09–4.59)	0.030
ML dimension, cm (range)	0.341 (0.11–1.02)	0.056
SI dimension, cm (range)	2.03 (0.95–4.35)	0.066

Compared to NCAs, SGAs had a significantly reduced incidence of new-onset postoperative diabetes insipidus (4.7% vs 12.1%, p = 0.028) (Fig. 1). Using a binary logistic regression model to evaluate the impact of SGA versus NCA on new diabetes insipidus, NCA was independently associated with increased odds of new postoperative diabetes insipidus (OR 2.39, 95% CI 0.892–6.42; p = 0.083), although this was not statistically significant. Intraoperative CSF leakage (OR 2.77, 95% CI 1.03–7.43; p = 0.043) and tumor cavernous sinus invasion (OR 2.83, 95% CI 1.05–7.58; p = 0.039) were both significantly independently associated with incidence of diabetes insipidus (Table 5). There were no significant differences in the need for future reoperations or death when comparing NCAs and SGAs (Table 3).

TABLE 5.

Binary logistic regression of variables associated with new postoperative diabetes insipidus

Variable	OR (95% CI)	p Value
Intraop CSF leak	2.77 (1.03–7.43)	0.043
NCA pathology	2.39 (0.892–6.42)	0.083
Cavernous sinus invasion	2.83 (1.05–7.58)	0.039

Discussion

Our study includes the largest cohort of patients with SGAs and NCAs with detailed patient characteristics and clinical outcomes from a single institution. SGAs are the most common SPAs, and it is critical to gain a more detailed understanding of them in order to provide optimal care to patients.¹⁶ Building on a study showing higher MIB-1 staining and a positive correlation between pre- and postoperative TVDTs in NCAs,² in a 2- to 3-fold larger cohort, we found that NCAs had a higher incidence of postoperative radiation therapy as well as a higher risk of recurrence and diabetes insipidus relative to SGAs.

The mean ages of patients with NCAs and SGAs were 58 and 59 years old, respectively (Table 1), similar to what has previously been described in the literature^2,10 and also supporting the literature suggesting that NFPAs present at a later age given the lack of hormonal excess and subsequent symptoms.¹⁷ NCAs were relatively evenly distributed between male and female patients, whereas the majority of patients with an SGA were male, similar to multiple reports in the literature^10,19—although differing with Balogun et al.,² who described a female predominance in their study. As expected, the majority of SGAs and NCAs were macroadenomas (Table 1), similar to what has previously been described in the literature.^2,5 In contrast to the descriptions of Balogun et al.,² there was no statistically significant difference between the MIB-1 index of SGAs and NCAs. This discrepancy may be related to the lack of consistent MIB-1 staining in both NCAs and SGAs, with 64/105 and 101/149, respectively, receiving MIB-1 staining. Previous studies have also failed to demonstrate a significant difference in MIB-1 staining between SGAs and NCAs, and the cutoff for MIB-1 remains variable, without a standard provided in the most recent WHO guidelines.^22,23 Immunopositivity for p53, a marker of TP53 mutations, was significantly higher in SGAs than in NCAs. This finding is in agreement with the higher p27 positivity in NCAs reported by Balogun et al.² Both p53 and MIB-1 staining theoretically indicate a higher risk category of tumor.^3,9,18 As a result, we would expect NCAs to have a higher incidence of p53 staining relative to SGAs, given their more aggressive nature.¹² Previous studies, however, albeit with limited sample sizes, have demonstrated higher p53 staining in nonfunctioning gonadotroph adenomas relative to NCAs, so our findings are not without precedent.¹⁹ Future investigation into the relationship between p53 and p27 positivity is required, potentially involving staining of the same SGA and NCA samples for both p27 and p53 to establish inter-test reliability of the 2 stains in these adenomas. Similar to Balogun et al.,² we found no significant difference in cavernous sinus invasion between NCAs and SGAs, although there was a trend toward higher invasion in NCAs. We also assessed sphenoid sinus invasion and similarly found no significant differences between the groups (Table 1). Rates of neoadjuvant radiation were low in both groups (Table 1), suggesting that surgeons believed resection alone to be clinically appropriate in most cases.

There were no significant differences between the groups in perioperative outcomes, aside from infection rate (Table 2). SGAs had a significantly lower rate of postoperative infection relative to NCAs, including no incidences of meningitis or sinusitis, although this finding is likely limited by the low overall incidence of postoperative infection. To determine the validity of this finding, further investigation with a larger population of patients would be required. There was no difference in GTR rates (Table 2) between NCAs and SGAs.

Long-term outcomes between the groups were also similar, although with some significant differences. Adjuvant radiation therapy following resection was significantly different between the groups, with SGAs having a rate of 9.4%, but with NCAs having a significantly higher rate of 23.4% (Table 3, Fig. 1). Adjuvant radiation therapy is often considered for nonfunctioning adenomas in the setting of STR or potentially patient preference, with good outcomes.^7,12 NCAs were also significantly more likely to recur, which was likely to have contributed to their higher rates of adjuvant radiation therapy (Table 3). When independent predictors of recurrence were assessed using a Cox regression model in patients with either an NCA or an SGA, tumor cavernous sinus invasion, NCA pathology, and AP dimension of the tumor remained independently associated with incidence of recurrence (Table 4). To our knowledge, this represents the first time that NCA pathology has been independently associated with increased risk of recurrence when compared to SGAs.

In agreement with our findings, Balogun et al.² found that NCAs’ pre- and postoperative TVDTs were positively associated, in contrast to SGAs, indicating that NCAs that grew quickly prior to surgery continued to grow at a high rate following surgery. They also found that cavernous sinus invasion and low p27 staining were predictors of NCA tumor growth.¹² Recurrence seems to be impacted by these variables as well; the NCAs in our study had a higher incidence of cavernous sinus invasion, although significantly lower p53 staining. Cavernous sinus invasion was also independently associated with recurrence (Table 4). Further studies could include serial TVDTs to assess any differences in NCAs versus SGAs. Given the relatively low rates of recurrence, future studies with additional patients may also be beneficial. There was no difference in reoperation between the groups (Table 3), suggesting that although NCAs had higher rates of recurrence and adjuvant radiation therapy, those requiring radiation were adequately controlled by the therapy.

SGAs also had significantly lower levels of diabetes insipidus relative to the NCAs (Table 3). Risk factors for postoperative diabetes insipidus from pituitary surgery include the following: visual abnormalities, suprasellar extension of the tumor, intraoperative CSF leakage, and maximal tumor diameter.^15,20 On multivariate analysis using binary logistic regression modeling, NCA pathology was independently associated with increased incidence of new diabetes insipidus, although this only trended toward significance (p = 0.083). Corresponding with what has previously been described in the literature,^15,20 intraoperative CSF leakage was independently and significantly associated with diabetes insipidus. Cavernous sinus invasion was also significantly associated with new postoperative diabetes insipidus on multivariate analysis. We suspect that this may be a result of the additional dissection required to remove these tumors, potentially increasing the risk of disturbing the pituitary stalk. Additional studies with larger patient sample sizes are required to fully elucidate the association between cavernous sinus invasion and postoperative diabetes insipidus. Because our study only included NFPAs, studies involving functional adenomas as well should be performed to determine if the association between cavernous sinus invasion and diabetes insipidus is NFPA-specific. Further investigation into differences in new postoperative diabetes insipidus between NCAs and SGAs is also required, given the small incidence of diabetes insipidus in our study, to determine if patients with NCAs are at higher risk for postoperative diabetes insipidus relative to SGAs.

Although the behavioral characteristics of SGAs and NCAs have been compared in the past,¹² there remains a paucity of data comparing the clinical outcomes between the 2 groups. Our study, to our knowledge, is the largest cohort of SGAs and NCAs with detailed patient characteristics as well as short- and long-term clinical outcomes, including diabetes insipidus, the need for adjuvant radiation therapy, time to recurrence, and incidence of reoperation, with an average follow-up of more than 2 years for all subgroups. We found that patients with NCAs have a higher incidence of postoperative diabetes insipidus and are at higher risk for recurrence relative to patients with SGAs.

Limitations of our study include its retrospective nature, lack of tumor volumetric data, and the relatively low number of patients in each group, albeit that limitation was due to the rarity of the pathological entities. Future directions could involve including tumor volumetric data and larger numbers of patients, perhaps using a multi-institutional database, to provide further insight into the differences between NCAs and SGAs. As transcription factor staining for SF-1, PIT-1, and TPIT becomes more widespread, additional studies could expand on this study by using the full repertoire of transcription factor staining when determining SGA and NCA subgroups. Follow-up studies could also investigate underlying biological mechanisms differentiating these 2 groups and could leverage techniques such as regional biopsies and single-cell RNA sequencing to investigate whether there is heterogeneity at various levels within these tumors.

Conclusions

Our study highlights differences in patient characteristics and clinical outcomes between NCAs and SGAs, supporting the concept that they should be considered distinct groups, as recognized in the 2017 WHO guidelines. SGAs have a much higher incidence in male patients. Relative to SGAs, NCAs have increased risk of recurrence and postoperative diabetes insipidus, both of which could reflect more biologically aggressive tumors. The rarity of these 2 subgroups limits the potential sample size of any study, highlighting the need for further multi-institutional studies to support our findings. Future studies investigating the underlying biological differences between NCAs and SGAs should also be considered.

Acknowledgments

We thank the UCSF Brain Tumor Research Center for its continued support.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Aghi, Haddad, Young, MP Pereira, El-Sayed. Acquisition of data: Haddad, MP Pereira, Joshi, KM Pereira, Osorio, Donohue, Peeran, Sudhir, Jain, Beniwal, Chopra, Sandhu, Theodosopoulos, Kunwar, Gurrola, Blevins. Analysis and interpretation of data: Haddad, MP Pereira. Drafting the article: Haddad, Young, Oh, Joshi. Critically revising the article: Aghi, Haddad, Oh, MP Pereira, Joshi. Reviewed submitted version of manuscript: Aghi, Haddad, Oh, Theodosopoulos, Kunwar, El-Sayed, Gurrola, Blevins. Statistical analysis: Haddad, MP Pereira. Administrative/technical/material support: Haddad, Theodosopoulos, Kunwar, El-Sayed, Gurrola, Blevins. Study supervision: Aghi, Haddad.