Clinical Course of Nonfunctional Pituitary Microadenoma in Children: A Single-Center Experience

Abstract

Context

Pituitary lesions consistent with microadenomas are increasingly discovered by MRI. Sparse data are available on the long-term clinical and imaging course of such lesions in children.

Objective

The aim of this study was to define the clinical and imaging course of pituitary lesions representing or possibly representing nonfunctioning microadenomas in children to guide clinical management.

Design

Retrospective observational study.

Methods

The clinical data warehouse at a tertiary care academic children’s hospital was queried with the terms “pituitary” AND “microadenoma” and “pituitary” AND “incidentaloma.” The electronic health records of the identified subjects were reviewed to extract data on the clinical and imaging course.

Results

A total of 78 children had nonfunctioning pituitary lesions incidentally discovered during clinical care, of which 44 (56%) were reported as presumed or possible microadenomas. In the children with microadenoma (median age 15 years, interquartile range 2), a majority (70%) underwent imaging for nonendocrine symptoms, the most common being headache (n = 16, 36%). No significant increase in the size of the microadenoma or cysts or worsening of pituitary function was seen over the average clinical follow-up of 4.5 ± 2.6 years. Four cases of drug-induced hyperprolactinemia resolved with discontinuation of the offending medication.

Conclusions

Asymptomatic pituitary lesions representing cysts, microadenomas, or possible microadenomas follow a benign course in children. In the absence of new endocrine or visual symptoms, repeat MRI may not be needed, and if performed, should be done in no less than a year. When possible, it is prudent to discontinue hyperprolactinemia-inducing medications before imaging.

This study examined the clinical course of incidentally identified pituitary microadenomas in children, revealing a benign course. Patients can be followed clinically with infrequent or no imaging.

The presence of silent pituitary adenomas in humans has been known for almost a century (1). Estimates of their prevalence on autopsy studies have varied between 14.7% and 37%, with an average of 10.7%, predominantly prolactinomas, on immunohistochemistry (1–5). In MRI studies of normal adult volunteers, pituitary “hypointensities” have ranged from 10% to 40% (6, 7). One study of MRI scans in children between 2 and 16 years noted “certain” microadenomas in 28.6% and “possible” microadenomas in 21.4% of children without hormonal abnormalities (n = 28) (8). In another study, 41 children were identified with pituitary “incidentalomas,” of which 6 (14.6%) were microadenomas (9).

The natural history of pituitary microadenomas is difficult to ascertain because of the lack of accessibility of the tissue during life. The tumors are thought to arise from single precursor cells that possess a unique proliferative advantage (10), but the tumor-promoting factors are not well understood (11). There is wide variability in the reported rates of progression and the presence of hormonal disturbances. In one aggregate report of 9 studies that included 166 tumors, 83% of microadenomas were unchanged, 10% increased, and 7% decreased in size at a follow-up interval ranging from 2.3 to 8 years (12). In some reports, there was no alteration in the pituitary function (13–15), whereas Yuen et al. (16) reported at least one hormone abnormality in 50% of the patients in their series.

In the Endocrine Society guidelines published in 2011, a pituitary incidentaloma was defined as a previously unsuspected pituitary lesion that is discovered on an imaging study performed for an unrelated reason (e.g., other than visual loss or a clinical manifestation of hypopituitarism or hormone excess) (17). According to these guidelines, all patients with radiologically diagnosed pituitary adenomas should undergo clinical and laboratory investigation to evaluate hypersecretion and hyposecretion. A formal visual field examination is advocated for all patients with a lesion abutting the optic nerve and chiasm. Follow-up MRI is recommended at 6 months and annually for a macroadenoma and every 1 to 2 years for a microadenoma. Clinical and biochemical testing at periodic intervals for nonprogressive adenomas can be tailored to the presence of symptoms and rate of progression of the tumor size.

Although the identification of such “incidental” or “clinically nonfunctioning” pituitary microadenomas is routine in clinical practice, there are insufficient data on the long-term outcomes, especially for children. Furthermore, it is not known whether adult guidelines provide a suitable framework for follow-up in the pediatric population. This study was designed to assess the experience at a tertiary care pediatric hospital with the clinical and imaging course and management in a cohort of children and adolescents with pituitary lesions that were thought to represent clinically nonfunctioning microadenomas or with lesions for which a microadenoma was a diagnostic consideration.

Methods

This is a retrospective observational study of the radiological data set of the clinical data warehouse at Boston Children’s Hospital, a tertiary care pediatric hospital, from January 1990 to December 2013. Neuroradiology reports were queried with the words “pituitary” AND “microadenoma” or “pituitary” AND “incidentaloma” for children between 0 and 18 years of age. The inclusion criteria for the cohort were a pituitary “incidentaloma,” defined as any pituitary lesion identified on MRI performed for an ailment unrelated to a pituitary symptom or hormonal abnormality in keeping with the Endocrine Society guidelines (17); and “nonfunctioning” pituitary microadenomas, defined as lesions in subjects who underwent imaging for an endocrine-related reason, but the identified microadenoma was considered nonfunctional by the treating clinician, or a cystic lesion of the pituitary (Rathke cleft or pars intermedia cyst). Patients with MRI reports indicating other brain abnormalities such as empty sella, central nervous system malignancy, congenital malformation, central nervous system diseases such as neurofibromatosis 1 or 2, tuberous sclerosis, Langerhans histiocytosis, previous brain surgery, and MRI scans performed as part of a research protocol were excluded. Patients treated for hyperprolactinemia or those diagnosed with Cushing disease were also excluded. The clinical course of children identified with microadenoma or possible microadenoma was determined by the review of medical records from January 1990 to December 2018. Patients with <1 year follow-up were excluded.

The subjects underwent dedicated pituitary MRI scans for imaging follow-up. MRI parameters were somewhat heterogeneous but generally included a small field of view; T1- and T2-weighted sequences of the pituitary gland, often with gadolinium-enhanced coronal and sagittal T1-weighted sequences; and slice thicknesses of 2 to 3 mm with little or no interslice gap. MRI reports either provided a diagnosis of a pituitary microadenoma or gave a differential diagnosis for the pituitary lesion that included Rathke cleft cyst, pars intermedia cyst, or pituitary microadenoma. The largest dimension of the pituitary lesion reported in the MRI report was registered as the lesion size. Any lesion with a difference in lesion size of ≤1 mm between two follow-up MRI reports was considered stable to account for a margin of error attributed to minor differences in scanning parameters such as the field of view, interslice gap, and patients being scanned on MRI machines of different magnet strengths. Because this study was designed to evaluate the clinical management of pituitary lesions based on the existing MRI report, no repeat radiological validation of the imaging interpretations was obtained.

Detailed information including demographic data, clinical presentation, and follow-up data were extracted from the electronic medical records only for the subjects with a diagnosis of “microadenoma” or “possible microadenoma” but not for those with the cystic lesions. All patients had at least one clinical visit, with varying degrees of hormonal testing. Visual field perimetry was performed by an ophthalmologist, where applicable. Hormonal dysfunction was defined with the laboratory assays and reference ranges of the Division of Laboratory Medicine at Boston Children’s Hospital: chemiluminescent immunoassay on the Siemens Immulite 2000 platform for GH, IGF-I, IGF binding protein 3, and ACTH; electrochemiluminescent immunoassay on the Roche e601 platform for cortisol, dehydroepiandrosterone sulfate, estradiol, FSH, TSH, free T4, total T4, and prolactin; and liquid chromatography mass spectrometry for testosterone.

The upper range of normal prolactin level for clinical care was 18 ng/mL. Elevated prolactin levels ≤200 ng/mL, temporally associated with a known inciting medication, were considered as medication induced unless proven otherwise (18). Biochemical GH deficiency was defined as IGF-I levels below the reference range for age if the patient was preadolescent and below the reference range for Tanner staging if the patient was of adolescent age. Children with GH deficiency underwent dynamic testing with arginine and glucagon stimulation. A peak stimulated GH level <10 ng/mL was considered GH deficiency. Primary hypothyroidism was defined as an elevation of TSH level for age with or without a low free T4 level. Central hypothyroidism was defined as an inappropriately low TSH level for a given free T4 or total T4 level. Adrenal hypofunction was defined as a morning cortisol level <5 μg/dL, a peak cortisol level of <18 μg/dL after a 250-μg injection of cosyntropin, or a minimum change in cortisol level from baseline to stimulated testing of 10 μg/dL. A 24-hour urine free cortisol level with adequate urine creatinine was used to define cortisol overproduction. Children noted to have high 24-hour urine free cortisol underwent overnight 1-mg dexamethasone suppression testing and midnight salivary cortisol evaluation. Gonadotropin deficiency was defined as low serum estradiol or testosterone with low or inappropriately “normal” LH and FSH levels.

The patients were evaluated clinically annually or as planned by the treating physician. The assessment of biochemical parameters and follow-up imaging were at the discretion of the treating physician. The duration of follow-up was ascertained from the date of the first MRI study to the last day of visit in the outpatient clinic.

Statistical analysis

Data for continuous variables are presented as mean ± SEM (SE) if normally distributed and median and interquartile range for skewed distribution. Data for categorical variables are presented as numbers or percentages. A Kolmogorov-Smirnov test and a Q-Q plot were used to check the normality of distribution of continuous variables. A χ² or Fischer exact test was used to compare categorical variables. An independent sample t test or Mann-Whitney test was used to identify the difference between two independent groups of continuous variables. A comparison of change in size was made when two or more MRI scans were available. A P <0.05 was considered statistically significant. Statistical analyses were performed in R version 3.4 (The R Foundation for Statistical Computing).

Results

Based on the a priori search criteria, a total of 843 imaging records of 643 unique patients were retrieved. According to the predefined screening criteria, 78 clinically nonfunctioning pituitary lesions were identified, of which 44 were classified as microadenoma or possible microadenoma (Fig. 1). The lesions were almost equally distributed on the left or the right side of the pituitary, and some were in the paramedial region. Sixteen of these patients underwent the imaging for an endocrine-related problem or hormonal abnormality, but the adenoma was considered a nonfunctioning pituitary adenoma (NFPA) by the treating clinician. The presenting symptoms and demographic distribution of the children in the cohort are elaborated in Table 1. The most common indication for imaging was headaches, including those for postconcussion symptoms (n = 17, 39%). Pubertal abnormalities were the most common endocrine disorder in the cohort (n = 9, 20%), including primary or secondary amenorrhea, precocious puberty, and delayed puberty. Four subjects underwent imaging for prolactin levels between 18 and 200 ng/mL while on neuropsychiatric medications.

Figure 1.

Summary of the cohort review that resulted in identification of the incidentaloma or nonfunctioning pituitary microadenomas identified during clinical care.

Table 1.

Demographics, Presenting Symptoms, and Pituitary Hormone Function Status in Children and Young Adults With Nonfunctioning Pituitary Microadenomas Incidentally Identified During Clinical Care

	n
Total patients assessed	44
Females, n (%)	36 (82)
White, n (%)	31 (70)
Age at first imaging, y (SD)	13.8 (3.8)
Clinical follow-up, y (SD)	4.5 (2.6)
MRI scans, median (range)	2 (1–9)
Indication for MRI
Headache	17
Precocious puberty	6
Delayed puberty	3
Hyperprolactinemia	4
Seizure	2
Short stature	2
Tall stature	1
Vision abnormalities	2
Focal numbness	1
Vertigo	1
Macrocephaly	1
Behavioral abnormalities	2
Fatigue	1
Temporomandibular joint assessment	1
Memory loss	1
Pituitary function assessment
Normal	24
Hyperprolactinemia (<200 ng/mL)	7
Abnormal gonadotrophin profile	5
Central hypothyroidism	1
Failed GH stimulation test	1
Not available	6

None of the children with “incidentaloma,” as defined by the Endocrine Society, had any hormonal abnormality. Four adolescents with mild hyperprolactinemia returned to normal after discontinuation of the offending neuropsychiatric medication, primarily Risperdal, and the clinicians maintained the medication (and not the microadenoma) as the cause of symptoms. One child with short stature was treated with recombinant human GH therapy, and another was followed with clinical observation. Three girls with precocious puberty were treated with GnRH therapy, one was followed clinically, and one was lost to follow-up. Two girls with delayed puberty were started on hormonal replacement therapy (19).

There was no temporal difference in the number of lesions noted per year from 2000 to 2013 (P = 0.47). Children with precocious puberty or short stature were younger, with a peak between 8 and 9 years, whereas adolescents with delayed puberty, primary or secondary amenorrhea, or medication-induced hyperprolactinemia were older, with a peak around 16 years. There was no statistically significant difference between the age of presentation, but the cohort with the endocrine abnormality had a longer duration of follow-up, albeit with the same number of clinical visits. The majority of patients in our series were adolescent girls. The median number of imaging studies was two (range 1 to 10), and four children had only one MRI. One of these children had juvenile rheumatoid arthritis and incidental identification of a pituitary lesion while being evaluated for temporomandibular joint integrity.

Thirty (68%) of the patients had pituitary lesions that measured ≤5 mm in the largest dimension (Table 2). Three imaging studies noted a maximum diameter of 11 to 12 mm, all reported as hemorrhagic microadenoma, included in the series because these were noted to decrease in size at follow-up. Four of the lesions caused asymmetry, and only one lesion caused impingement on the optic chiasm. The ophthalmological examination for this patient was normal. Forty cases (90%) had two or more MRI scans, allowing size comparison of the pituitary lesion over time; 16 of these had the first follow-up MRI within 6 months of the first imaging study. The 40 cases were divided into tertiles, such that group 1 consisted of lesions ≤3 mm, group 2 consisted of lesions between 3.1 and 5 mm, and group 3 consisted of lesions between ≥5.1 mm. The group with the smallest lesions showed a statistically significant difference in size, whereas the other two groups were similar (summary in Table 2, individual change in Fig. 2). Two cases showed absence of the lesion on the subsequent MRI. Only one MRI report indicated an increase in the size of the pituitary lesion, by 4 mm, in the follow-up MRI report, but it remained <10 mm 6 years after the initial identification and was considered stable on subsequent MRI scans. Ten lesions decreased in size by ≥2 mm.

Table 2.

Summary of the Follow-Up Imaging Findings of Children and Young Adults With Nonfunctioning Pituitary Microadenomas

Group	Size Range (mm)	Sample (n)		Mean ± SD (mm)		P Value	Size Decrease,a n (%)	Size Increase,a n (%)	Endocrine Dysfunction (n)
		Initial	Final	Initial	Final
1	≤3	17	16	2.32 ± 0.18	1.52 ± 0.29	<0.001	5 (12.5%)	0	Hyperprolactinemia (1)
									Precocious puberty (2)
									Delayed puberty (1)
									Short stature (1)
2	3.1–5	13	10	4.42 ± 0.14	4.38 ± 0.68	0.96	2 (5%)	1 (2.5%)	Precocious puberty (2)
									Delayed puberty (1)
									Tall stature (1)
3	≥5	14	14	8.07 ± 0.59	7.5 ± 0.53	0.36	3 (7.5%)	0	Hyperprolactinemia (3)
									Precocious puberty (2)
									Short stature (1)

The total number of subjects was divided into tertiles based on lesion size at the initial study that identified the lesion.

^aDenominator for size change is the final sample size, where ≥2 MRI scans are available for comparison.

Figure 2.

Change in the size of the nonfunctioning pituitary microadenoma from the initial to the final size (n = 40). The lines connect the same subject at initial and final size assessed by radiological imaging. The boxplots represent a summary of the initial and the final size.

Most children (77%) were evaluated by a pediatric endocrinologist, 64% had a neurology or neurosurgery consultation, and 34% had a visual field assessment by an ophthalmologist.

Discussion

The routine use of brain MRI has significantly increased the identification of asymptomatic pituitary microadenomas or possible microadenomas. NFPA is the most common form of incidentaloma in adults, 50% to 90% of which are solid tumors, mostly (95%) adenomas (4, 20, 21).

This study provides a large case series of presumed or possible pituitary microadenomas in children, with longitudinal observation data, specifically focusing on solid or solid/cystic lesions considered at high risk for progression or hormonal disturbance. Based on repeated imaging, only 1 of the 40 cases of microadenoma had an increase in the size of the pituitary lesion on follow-up for ≤10 years, albeit without any hormonal changes. These results are more consistent with previous smaller pediatric studies (8, 9, 22) than with adult studies. The lesions in the group with the smallest lesions had the most significant decrease in the size of the lesions. This probably occurred because it is harder to identify and define smaller lesions.

Despite two decades of evolving imaging technology, no temporal differences in the rates of detection of microadenomas were noted in this study. We hypothesize that the availability of better resolution has allowed radiologists to be more confident in differentiating solid from cystic pituitary lesions that are considered to be benign.

In subjects where an endocrine abnormality such as pubertal disturbances or stature abnormalities prompted the MRI, the clinicians variably pursued follow-up MRI without changing the therapeutic course. A previous report has noted that incidental microadenomas should not be considered a risk factor for treatment with recombinant human GH or therapy for precocious puberty (22).

Souteiro et al. (9) estimated the prevalence of pituitary incidentalomas at 257 per 100,000 patients based on their review of radiology reports >10 years. Such an estimate cannot be obtained from this case series because the search strategy for the reports selectively identified those that mentioned “pituitary” and “incidentalomas,” creating a selection bias. Whereas the search was performed for subjects aged 0 to 18 years at the time of MRI, the median age of subjects with incidentalomas was 14 to 16 years. Furthermore, four subjects had normal reports of MRI before the identification of microadenoma. It is possible that incidentalomas are either absent at younger ages or not large enough to be identified, because autopsy studies have identified pituitary incidentalomas as early as 2 years of age (1).

The Endocrine Society guidelines recommend follow-up MRI at 6 months in cases of macroadenoma and at 1 year for microadenomas (17). On the other hand, French guidelines recommend neither radiological nor hormonal surveillance in cases of NFPA ≤5 mm in size (23). The results of this case series agree with the latter. For lesions where a follow-up MRI may be considered, the timing of the imaging remains an unresolved question. Studies in adults have attempted to provide estimates of rates of growth of pituitary tumors. One study of macroadenomas estimated a growth rate of 0.6 mm/y, with continued growth noted at 22 years (24). Honegger et al. (25) found tumor volume doubling time ranging from 0.8 to 27.2 years, with no correlation between initial tumor size and rate of doubling. In this series, cases where the size change was noted, it was seen at 12 months but not at 6-month follow-up. Therefore, the authors suggest that surveillance with MRI scans for microadenoma may not be indicated in children, and if done it should be performed at a minimum 1-year interval.

In this study, four children underwent MRI despite recognition of drug-induced hyperprolactinemia. In a study of 117 adults with hyperprolactinemic pituitary macroadenomas, Hong et al. (26) noted NFPA in 19 of 70 surgical excisions on histochemical staining. All these patients had prolactin levels <200 ng/mL, and only one subject showed a reduction in size, from 6 to 4 mm, on repeat MRI (26). Similar to what was seen in this study, drug-induced hyperprolactinemia ≤200 ng/mL can be seen in children (27, 28). The mechanism of hyperprolactinemia in such cases is primarily by blockade of the endogenous inhibitory influence of dopamine receptors by the drugs causing symptomatic galactorrhea, menstrual disturbance, or impotence (29). Although the number of subjects with drug-induced hyperprolactinemia in this study is small, it is unlikely that a prolactinoma is induced or increased in size by the medications, and thus the authors recommend drug withdrawal before MRI in children and young adults with suspected drug-induced hyperprolactinemia whenever possible, in accordance with recommendations based on previous literature review (29).

The Endocrine Society guidelines recommend initial hormonal evaluation for both hyperpituitarism and hypopituitarism in patients with a macroadenoma and a hyperpituitarism evaluation in patients with a microadenoma. Hyperprolactinemia is the most common hormonal abnormality seen with pituitary adenomas (30, 31). The recommendation for assessment of GH status stems from immunohistochemistry performed on autopsy-identified pituitary microadenomas (32). Screening for glucocorticoid excess may be considered when suspected clinically (17). The French Consensus guidelines recommend measurement of prolactin and IGF-I levels, without endorsing measurement of cortisol levels. King et al. (33) analyzed the cost-effectiveness of various management strategies for pituitary incidentaloma based on the hidden Markov model over a 10-year horizon. The authors included four potential management strategies: observation, testing with prolactin levels only, complete endocrine panel, and clinical follow-up with MRI study. The incremental effect in the quality-adjusted life years between the most effective strategy of complete endocrine panel to the least effective strategy of expectant management was small, accounted for by the interplay of patient anxiety from their incidental pituitary microadenoma, low positive rates of screening tests, low incidence of endocrine and neurologic disease, and effective treatments with low morbidity and mortality (33). In this study, 36 (86%) subjects underwent testing for the pituitary panel at initial evaluation, and four had urine cortisol measurements that were normal. The results endorse the French Consensus guidelines for assessment of prolactin and IGF-I levels coupled with careful clinical follow-up, with additional testing indicated only if other relevant symptoms are seen.

Three subjects seen in this cohort complained of vision disturbances, and one had a tumor abutting the optic chiasm. None had objective vision changes on ophthalmologic evaluation. The data are insufficient to make recommendations on screening for visual field defects. The Endocrine Society guidelines suggest that visual field testing is necessary only for macroadenomas.

This study is limited by the retrospective nature causing a lack of standardized intervals for longitudinal data. After a median of four visits, no further follow-up is documented in many of the charts (n = 28, 64%), although the treating physician noted the desire to continue follow-up. This finding could indicate sufficient reassurance for the families to discontinue care, assumption of care by another provider, or an unexpected outcome, and the need for a prospective study to understand the long-term outcomes remains. The study was limited to institutional records, and therefore any clinical follow-up outside the system when a report was not scanned in the medical records is not known.

The data obtained in this case series provide reassurance that incidental or pituitary lesions representing NFPA identified in children and adolescents undergoing imaging studies follow a benign course. Although this study is limited to one center, the large number of patients reported in this series provides valuable evidence toward the management of NFPAs in children and adolescents. The study addresses a challenging management dilemma, and we hope that the results will spark further research, especially for children and young adults with medication-induced hyperprolactinemia.

Glossary

Abbreviation:

NFPA

nonfunctioning pituitary adenoma