The pituitary stalk, known as the infundibulum or infundibular stalk, is the connection between the hypothalamus and the pituitary gland, especially the posterior pituitary gland. The shape of the stalk is smoothly tapering, and the normal pituitary stalk is widest superiorly, measuring 3.25±0.56 mm in transverse diameter at the optic chiasm and 1.91±0.40 mm at its pituitary insertion¹⁾. Signal intensity of the pituitary stalk is less than that of the optic chiasm and posterior pituitary gland on brain magnetic resonance imaging (MRI) T1-weighted images. Deviation/tilt of the pituitary stalk can be seen without any problems¹⁾. Disturbance of the connection between the hypothalamus and the pituitary gland via pituitary stalk can cause anterior and posterior pituitary dysfunction. Because of the critical location and role of the pituitary stalk, patients who have disease involving the pituitary stalk often present with various clinical features of hypopituitarism, DI, or hyperprolactinemia, and clinicians have hesitated to perform biopsies that may cause iatrogenic hypopituitarism. Therefore, the diagnosis of pituitary stalk lesions has been based on not only pathologic examination but also clinical and radiologic evaluations, such as MRI.

Pituitary stalk lesions can be divided into the following 3 categories: congenital/developmental, inflammatory/infectious, and neoplastic²⁾. It is important for clinicians to discriminate neoplastic lesion from the other lesion in terms with treatment and prognosis.

A few studies of pituitary stalk lesions have been performed; however, most of the study subjects were adults with only a small number of children^3,4). Herein, we reviewed the clinical features, radiologic findings, and pathologic diagnoses of pituitary stalk lesions in children and adolescents treated at the Seoul National University Children's Hospital. Also, we investigated the difference of clinical and radiologic findings between the neoplastic and nonneoplastic lesions of pituitary stalk.

This study is a retrospective review of patients under 18 years of age with pituitary stalk lesions, diagnosed and treated at the Seoul National University Children's Hospital between January 1, 2000, and June 30, 2013. We performed a text search for head MRI reports by using 'pituitary stalk', 'infundibulum', and 'infundibular stalk', as keywords, and a total of 231 patients were found. Of these, 152 patients with a normal pituitary stalk, described as 'normal pituitary stalk' or 'normal infundibulum' in their MRI report, were excluded. In addition, 1 patient who had an MRI report only without clinical findings was excluded and 2 patients whose MRI abnormality was postoperative change were excluded. In the remaining 76 patients, we characterized the pituitary stalk lesions by using the initial brain MRI on which a stalk abnormality was identified (Fig. 1). All laboratory, pathological, and clinical evaluation, were reviewed. We compared the clinical, radiological, and laboratory findings of neoplastic group and nonneoplastic group.

In this study, we reviewed the etiology of pituitary stalk lesions in children and adolescents evaluated at the Seoul National University Children's Hospital. Previous studies of pituitary stalk lesions have largely focused on adults or a specific disease/clinical feature, such as LCH⁵⁾ or DI⁶⁾, or have contained few cases.

Turcu et al.³⁾ and Hamilton et al.⁴⁾ published large case series with a focus on pituitary stalk lesions. The study of Hamilton et al.⁴⁾ included 44 adults and 21 children and adolescents, and that of Turcu et al.³⁾ included 135 adults and 17 children and adolescents. In accordance with our study, neoplastic lesions predominated in these studies regardless of the classification of LCH as a tumorous condition by Hamilton et al.⁴⁾ and as an inflammatory disorder by Turcu et al.³⁾. Therefore, when a pituitary stalk lesion is discovered and believed to be a primary malignancy, it may be necessary to evaluate stalk tissue. However, because of the risks of hypopituitarism caused by a stalk biopsy, few patients receive a pathologic diagnosis based on tissue from the pituitary stalk. In the studies of Hamilton et al.⁴⁾ and Turcu et al.³⁾, 26 of 65 patients (40%) and 37 of 152 patients (24%), respectively, had a pathologic diagnosis with tissue samples from the pituitary stalk or other sites. In our study, 37 of 76 patients (48.7%) underwent biopsy or excision at the stalk or other site. In the study of Beni-Adani et al.⁷⁾, among 7 children with DI who had surgery for the pathologic diagnosis of a thickened pituitary stalk, none deteriorated neurologically or endocrinologically after the operation. The authors argued that, therefore, open biopsy should be the preferred diagnostic method in children with a thickened pituitary stalk.

In terms of endocrine disturbance in our study, half of the patients developed DI and a half had one or more adenohypophyseal hormone deficiencies. Secondary hypothyroidism was the most common, and secondary hypogonadism was the least common adenohypophyseal hormone deficiency. In the study of Turcu et al.³⁾, secondary hypogonadism was the most common, and secondary adrenal insufficiency was the least common pituitary hormonal deficiency. The hormonal test most frequently performed was for evaluation of thyroid function, and the least performed test was for evaluation of GH.

Congenital lesions were found most frequently in children in the series of Hamilton et al.⁴⁾. In the study of Turcu et al.³⁾, no specific category predominated in pediatric patients. In our study, neoplasm, including LCH, was the most common etiology of pituitary stalk lesions, accounting for 68.4%. Next, congenital lesions accounted for 21.1%. However, because Seoul National University Children's Hospital is a tertiary hospital where many children are treated for neoplasm, selection bias may have occurred. We classified LCH as a neoplasm because it was sometimes difficult to distinguish on MRI between an intracranial tumor, such as an intracranial germ cell tumor, and LCH. Three patients were diagnosed with acute lymphoblastic leukemia/lymphoma (1 patient with acute lymphoblastic leukemia, L1; 1 patient with precursor B-cell acute lymphoblastic leukemia; and 1 with diffuse large B-cell lymphoma). The patient with acute lymphoblastic leukemia, subtype L1, had normal pituitary hormone status. In the patient with precursor B-cell acute lymphoblastic leukemia, pituitary hormones were not evaluated except for those of the hypothalamo-neurohypophyseal tract and in the patient with diffuse large B-cell lymphoma, tests for GH gonadal axis were not done.

In our study, we found only 1 case of metastatic solid cancer, which was a retinoblastoma with brain metastasis. In the study of Turcu et al.³⁾, neurosarcoidosis was diagnosed in 11 patients and was the most common etiology of inflammatory lesions. In addition, in a previous study, stalk thickening and hypophyseal and hypothalamic infiltration were reported in patients with neurosarcoidosis⁸⁾. Nevertheless, we did not find any cases of neurosarcoidosis. Sarcoidosis manifests most commonly in the fourth decade of life^9,10,11). It occurs throughout the world in all races, and the highest annual incidence of 5 to 40 cases per 100,000 people has been observed in northern European countries^12,13). Childhood-onset sarcoidosis is rare and can occur as a part of systemic sarcoidosis or as an isolated presentation^9,14,15,16).

In the study of Turcu et al.³⁾, congenital lesions had a specific pattern of enhancement on MRI, but neither the pattern of MRI enhancement nor the thickness of the pituitary stalk lesion correlated with clinical features.

There are some limitations in our study. First, it was a retrospective review. As a result, uniform testing of pituitary hormonal status was not done. Second crucial limitation is the lack of diagnoses confirmed via tissue biopsy. Because biopsy from the stalk can cause deficiencies of pituitary hormones, diagnoses were based on clinical features and imaging. The unknown pituitary lesions could be diagnosed as a neoplasm later. Third, the number of patient was too small to apply the conclusions to general population in spite of relatively large pediatric cohort comparing with previous studies. In our study, the predominant neoplastic patient number was due to selection bias as mentioned above.

Despite all these limitations, our study showed that the etiology of pituitary stalk lesions in children and adolescents was variable and slightly different from that in adults. Neoplastic lesion was common and infectious/inflammatory lesion was rare. Male predominance was found especially in neoplastic group. On MRI, pituitary stalk thickness was related with neoplasm but enhancement of lesion was not. The most common presenting symptom and hormone abnormality of pituitary stalk lesion was DI, which was more prevalent in neoplastic stalk lesions. GH and ACTH deficiencies were more associated with nonneoplastic lesions.

In conclusion, the etiology of pituitary stalk lesions in children and adolescents was variable and slightly different from that in adults. A systemic evaluation of clinical symptoms, endocrine disturbances, and radiologic findings of pituitary stalk lesion, could help to suspect neoplastic lesions which need a pathologic evaluation such as biopsy or excision.