[68 Ga]-DOTATATE PET/MR-based evaluation of physiologic somatostatin receptor 2 expression in the adult pituitary gland as a function of age and sex in a prospective cohort

The pituitary gland has the fourth highest physiologic avidity of [68 Ga]-DOTATATE. In order to guide our understanding of [68 Ga]-DOTATATE PET in clinical contexts, accurate characterization of the normal pituitary gland is first required. This study aimed to characterize the normal pituitary gland using dedicated brain [68 Ga]-DOTATATE PET/MRI as a function of age and sex. A total of 95 patients with a normal pituitary gland underwent brain [68 Ga]-DOTATATE PET examinations for the purpose of diagnosing CNS SSTR2 positive tumors (mean age: 58.9, 73% female). Maximum SUV of the pituitary gland was obtained in each patient. SUV of superior sagittal sinus was obtained to calculate normalized SUV score (SUVR) of the gland. The anatomic size of the gland was collected as maximum sagittal height (MSH). Correlations with age and sex were analyzed. The mean SUV and SUVR of the pituitary gland were 17.6 (range: 7–59.5, SD = 7.1) and 13.8 (range: 3.3–52.6, SD = 7.2), respectively. Older females had significantly higher SUV of the pituitary gland compared to younger females. When stratified by age and sex, both older and younger females had significantly higher pituitary SUV than older males. SUVR did not differ significantly by age or sex. MSH of the pituitary gland in younger females was significantly greater than in younger males at all age cutoffs. This study provides an empiric profiling of the physiological [68 Ga]-DOTATATE avidity of the pituitary gland. The findings suggest that SUV may vary by age and sex and can help guide the use of [68 Ga]-DOTATATE PET/MRI in clinical and research settings. Future studies can build on these findings to investigate further the relationship between pituitary biology and demographic factors.


Introduction
Somatostatin regulates hormone secretion and cell proliferation in the pituitary gland through the five somatostatin receptor subtypes (SSTR1-5) expressed on all pituitary cell types, with a high proportion of SSTR5, lower levels of SSTR2, and very low levels of SSTR3 and SSTR1 found in the normal pituitary gland [1,2]. Translating the differential expression patterns of SSTRs between normal and pathologic pituitary cells in the clinical settings, somatotroph adenomas can be treated with somatostatin analogs such as octreotide and lantreotide, which preferentially bind to SSTR2 and SSTR5, while nonfunctional pituitary adenomas (NFPA) express predominantly SSTR3 and SSTR2 (with expression levels of the latter demonstrating some heterogeneity across published studies) and are less responsive to somatostatin analogs [3][4][5][6][7][8][9][10][11][12][13]. Radiographically, the PET radiotracer [68 Ga]-DOTATATE targets SSTR2 with high affinity and as such, the normal pituitary gland expresses the fourth highest physiologic avidity of [68 Ga]-DOTATATE in the body after the spleen, kidneys, and adrenal glands [14,15].
[68 Ga]-DOTATATE PET/CT is widely used for the diagnosis and management of gastroenteropancreatic neuroendocrine tumors (GEPNET), typically using a skull base to thigh acquisition protocol with a large field of view [16]. Intracranially, to better characterize the pituitary gland and to diagnose and treat SSTR-positive intracranial tumors, a dedicated brain PET/CT or PET/MRI acquisition protocol is critical [17,18]. We have developed a dedicated [68 Ga]-DOTATATE brain PET/MRI acquisition protocol, which we have applied to evaluation of SSTR2positive neoplasms in the brain and skull base, including meningioma, esthesioneuroblastoma, and paraganglioma [18][19][20][21]. Prior studies have demonstrated that with high accuracy, [68 Ga]-DOTATATE PET allows for quantifiable, clinically relevant measurements of cellular SSTR2 expression in the target regions, with published approaches including reporting absolute standartized uptake values (SUV) and SUV ratio (SUVR) normalized to the superior sagittal sinus (SSS) [17,18,21].
[68 Ga]-DOTATATE PET, with its specificity towards SSTR2, has potential to aid in evaluating pituitary adenomas (pitNET) by helping differentiate between normal tissue and different sellar or parasellar tumor types and optimize patient selection for targeted therapy with somatostatin analogs [22]. In order to clinically utilize areas of differential avidity of [68 Ga]-DOTATATE PET in the context of pathologies in the sella turcica, accurate characterization of the normal pituitary gland is first required. While normative data for the anatomic pituitary gland size have previously been studied, no normative functional PET data without concurrent pituitary pathologies currently exists in the literature [23][24][25][26][27][28][29][30][31]. Our study aimed to evaluate physiological SSTR2 expression in the normal pituitary gland using dedicated brain [68 Ga]-DOTATATE PET/MRI in a clinical cohort of patients without known pituitary pathology stratified by demographic factors such as age and sex.

Patient population
Institutional review board approval with informed consent was obtained for this HIPAA compliant study. A total of 116 patients underwent brain [68 Ga]-DOTATATE PET examinations (109 PET/MR, 7 PET/CT) between August 2018 and February 2022 as part of a prospective clinical trial (Clini-calTrials.gov Identifier: NCT04081701) for the purpose of diagnosing CNS SSTR2 positive tumors, including meningioma, paraganglioma, esthesioneuroblastoma, adenoma, and schwannoma. Seven patients who were ineligible for PET/MRI underwent PET/CT and MRI separately. From this cohort, all patients who had normal pituitary glands based on whole-brain MRI appearance and clinical follow ups were included. Patients with a history of radiation treatment prior to imaging or with any imaging evidence of sellar or parasellar neoplasm were excluded. Additionally, in patients with multiple longitudinal [68 Ga]-DOTATATE PET/MRI examinations, only the earliest scan was included. Clinical chart review was performed to collect clinical and demographic characteristics of the study population, including age, sex, surgical history, and radiation treatment history. In our intention to capture sex differences in the pituitary gland, sex is defined as a biological sex-a multi-dimensional construct based on a cluster of anatomical and physiological traits, that include external genitalia, secondary sex characteristics, gonads, chromosomes and hormones [32]. Sex was collected as a binary variable-male and female. Chart review presumed that documented sex is equivalent to sex assigned at birth.

Image acquisition
PET/MRI was performed on the Biograph mMR scanner (Siemens Healthineers, Erlangen, Germany) in all cases except one patient who was scanned on the GE SIGNA PET/MR scanner (GE Healthcare, Milwaukee, WI). All PET data acquisitions started at 7 ± 3 min post-injection of 172.9 ± 18.4 MBq of [68 Ga]-DOTATATE. The PET data were continuously acquired in 3D List Mode for a total period of 50 min and then histogrammed to a single sinogram of a time frame of 7-57 ± 3 min post-injection. The acquisition time of 50 min was used in our cohort for the specific purpose of obtaining dynamic PET data of the tracer uptake over time, which was a focus of a different, unrelated study.
All PET images were reconstructed with the default Ordered Subsets Expectation Maximization reconstruction algorithms of the manufacturer with Point Spread Function modeling (OSEM-PSF) using three iterations and 21 (Biograph mMR) or 28 subsets (Signa). The resulting image matrix size was 344 × 344 × 127 (192 × 192 × 89) voxels with a voxel size of 2.086 × 2.086 × 2.031 mm (1.875 × 1.875 × 2.780) mm for Biograph mMR (Signa). During image reconstruction, the PET data were corrected for attenuation, scatter, randoms, normalization, dead-time, decay and frame duration using the default settings. For attenuation and scatter correction, the manufacturer's default method and settings for estimating the MR-based brain tissue attenuation map were employed.
MRI was performed according to institutional protocol, including pre-and postcontrast sagittal 3D T1 SPACE (TR/TE 600-700 ms/11-19 ms, 120 degree flip, 1 mm slice thickness) and postcontrast 3D T2 FLAIR (TR/TE 6300-8500 ms/394-446 ms, 120 degree flip, 1 mm slice thickness). MR-based attenuation correction was obtained according to manufacturer's standard-of-care specifications. For patients who underwent PET/CT and MRI separately, the CT image set of the PET/CT was subsequently registered to the postcontrast T1-weighted MR images using the rigid registration algorithm residing on Syngo.Via workstation (Siemens Healthineers, Erlangen, Germany) and the resulting transformation matrix was then applied to the PET image set to register it to the MRI images.

Quantitative imaging analysis
All reconstructed PET images were initially displayed in quantitative units of becquerel (Bq)/mL. Then the absolute maximum SUV metric was calculated at every image voxel by dividing the respective Bq/ml pixel value with the ratio of the administered dose of the radiotracer, in units of Bq, over the subjects' body weight (in units of g) to remove the confounding effect of radiotracer dose and body weight when quantifying the [68 Ga]-DOTATATE uptake in every tissue. Regional absolute maximum SUV scores were subsequently extracted from a set of image pixels defining the volume-ofinterest (VOIs) in each PET image. Previous studies with [68 Ga]-DOTATATE PET have demonstrated high sensitivities and specificities in measuring cellular SSTR2 expression in the target regions with both standard SUV and normalized SUV (SUVR) [17,18,21]. In order to standardize the comparison of SUV of the pituitary glands across patients, we normalized the VOI-based pituitary SUV to the superior sagittal sinus (SSS) SUV as previously studied, which in the context of our brain imaging can be considered as a cranial blood pool reference given its lack of SSTR2 expression and its function as a cranial blood pool [18]. The VOIs were drawn for the target lesions, including the pituitary gland and SSS, and maximum SUV were reported as part of the routine clinical radiology report at our institution. Maximum SUV is referred to as SUV hereafter.
The PET/MR and PET/CT images were read by a fellowship trained neuroradiologist with additional board certification in nuclear medicine. The images were interpreted for the clinical purpose of diagnosing suspected CNS SSTR2 positive tumors, and the radiologists had access to the full patient information at the time of study interpretation. The anatomic delineation of the VOIs in the PET images was based on the coregistered sagittal 3D T1-weighted postcontrast MR images with respective axial and coronal reformats, which were drawn to include the entire pituitary gland as visualized on postcontrast T1 imaging, as determined by the interpreting neuroradiologist.

Maximum sagittal height analysis
In additional to the SUV based PET analyses, in order to evaluate the anatomic pituitary gland size as a function of age and sex and to compare our cohort to the existing literature of normative data of the pituitary gland, which only pertains to such anatomic measurements, the anatomic size of the pituitary gland in our cohort was collected as maximum sagittal height (MSH) [23]. MSH was measured in cm using the sagittal postcontrast 3D T1-weighted series and recording the largest craniocaudal dimension of the gland. For 7 patients who received [68 Ga]-DOTATATE PET/CT, MSH was not collected.

Statistical analysis
Mann Whitney test, two-way ANOVA, simple linear regression, and simple logistic regression were performed to study relationships between SUV/SUVR and age or sex, across the entire cohort and when the cohort was stratified by age, sex, and both. Age cutoffs of 55 and 60 were used to stratify the cohort by age for statistical analyses. The aforementioned age cutoffs were used to maintain balance of sample sizes when the cohort was stratified by age and to account for major hormonal changes after menopause in females. Mann Whitney test, two-way ANOVA, and simple linear regression were used to determine statistical significance between MSH and age or sex across the cohort and when stratified by age, sex, and both. Simple linear regression was performed to study relationships between MSH and SUV/SUVR across the cohort and when stratified by age, sex, and both. Graph-Pad Prism 9 was used to perform all statistical analyses.

3
P values below 0.05 were considered to indicate statistical significance.

Descriptive and correlative analysis of SUV
Across the cohort, mean SUV of the pituitary gland was 17.6 (range: 7-59.5, SD = 7.1). Mean SUV of the SSS and SUVR of the pituitary gland normalized to SUV of SSS were 1.4 (range: 0.5-3.1, SD = 0.5) and 13.8 (range: 3.3-52.6, SD = 7.2), respectively. Mean SUV of the pituitary gland in the cohort, stratified by sex and age, is shown in Table 2 at the two age cutoffs. SUV and SUVR of the pituitary gland did not differ significantly between males and females and between younger and older patients (Fig. 2). Within the same sex groups, at age cutoffs of 55 and 60, older females had a significantly higher SUV of the pituitary gland compared to younger females ( (Fig. 3). SUVR did not differ significantly when stratified by age, sex, and both. Two-way ANOVA revealed no statistically significant interactions between the effects of age and sex on both SUV and SUVR. Similarly, regression analyses did not demonstrate age or sex as statistically significant predictors of SUV or SUVR, both across the entire cohort and when stratified by age, sex, and both. Finally, logistical regression between SUV/SUVR and sex revealed no statistically significant relationship.

Maximum sagittal height (MSH)
In our cohort, MSH of the pituitary gland was significantly greater in younger females than in younger males at both age cutoffs (0.61 vs 0.54, p = 0.01; 0.61 vs. 0.54, p = 0.01, respectively). The difference in MSH between females and males was not observed in older groups and between all females and all males. A representative patient with a measurement of MSH is demonstrated in Fig. 5. Mean and SD of the MSH of the pituitary gland in groups stratified by age and sex are demonstrated in Table 3. MSH did not differ significantly when the cohort was stratified by age alone. Regression analysis revealed no correlation between MSH and age across the entire cohort and when stratified by age, sex and both. Regression analysis did not demonstrate SUV or SUVR as statistically significant predictors of MSH, both in the entire cohort and when stratified by age, sex, and both.

Discussion
Here we report the normative data of SUV and size of the adult pituitary gland with [68 Ga]-DOTATATE PET/MR from a heterogenous cohort of 95 patients, stratified by age and sex. In our cohort, SUV/SUVR of the pituitary gland did not differ significantly between males and females and between younger and older groups. However, older females at age cutoffs of 55 and 60 had significantly higher pituitary     Table 2). The differences in SUV observed between groups stratified by age and sex generate hypotheses regarding the relationship between somatostatin receptor expression in the pituitary gland and demographic factors. The higher SUV observed in older females compared to younger females may reflect effects of menopause-related hormonal changes on somatostatin signaling. Pituitary size may increase in women during the fifth to sixth decades of life, potentially due to increased gonadotroph activity from loss of negative feedback by estrogen [25,30,31]. FSH upregulates SSTR2 expression in ovarian granulosa cells, suggesting a link between menopausal status and cellular SSTR2 expression [33]. The direct effect of estrogen on SSTR expression in pituitary cells is less clear, with in vitro studies showing both upregulation and downregulation [34][35][36][37]. The exact mechanisms of hormonal regulation of SSTR expression in pituitary cells are likely to be complex and receptor subtype specific [1].
When the cohort was stratified by both age and sex, both younger females and older females had significantly greater pituitary SUV than older males, suggesting that older males may express the lowest pituitary SUV. While pituitary SUV did not differ significantly between older males and younger males in our cohort, age-related pituitary atrophy may partially provide an explanation for such findings. Although andropause in males is less established than menopause in females, animal studies have shown minimal direct influence of androgens on SSTR expression on pituitary cells [38][39][40]. In general, pituitary gland reaches its maximum size in second to third decade of life in the setting of physiologic hypertrophy during puberty, followed by a gradual decline with age in both sexes [25,29,30]. While SUV or SUVR did not directly correlate with pituitary size in our cohort, one hypothesis is that age related pituitary atrophy may partially explain the lower pituitary SUV observed in older males, an effect that older females may be shielded from secondary to hormonal changes during menopause.
In pituitary adenomas, reported histologic SSTR expression is heterogeneous and varies between tumor subtypes. For instance, SSTR2 is relatively downregulated in NFPA compared to somatotropinomas [6]. SSTR3 is expressed in most pituitary adenomas including NFPA, while SSTR5 and SSTR2 are more specific to somatotroph, thyrotroph, corticotroph, lactotroph, and gonadotroph adenomas and are likely involved in the regulation of hormonal secretion [4]. The normal pituitary gland expresses high levels of SSTR5, lower levels of SSTR2, and very low levels of SSTR3 and SSTR1 [2]. The somatostatin analogs octreotide and lanreotide bind preferentially to SSTR2 and are used in the treatment of acromegaly, while they have shown a lack of efficacy in NFPA where the most predominant subtype is SSTR3, often followed by SSTR2, with SSTR2 expression demonstrating some heterogeneity across published studies [5][6][7][8][9][10][11][12][13].
Hence, [68 Ga]-DOTATATE PET, with its specificity towards SSTR2, is equipped to utilize the differential expression of SSTR by pituitary cells and tumors and offer clinically valuable information that complements anatomic imaging. While dynamic T1 weighted contrast enhanced MR  sequence is the standard imaging for pituitary adenomas, it lacks assessment of intrinsic permeability properties of the gland. In large adenomas, the normal pituitary gland can be structurally displaced which impairs visualization on anatomic imaging [22]. The novel functional imaging has potential to aid in differentiating between different types of pituitary adenomas and other non-pituitary sellar malignancies. Postoperative surveillance of residual pituitary adenomas and preoperative delineation of tumors from normal tissue may be enhanced to prevent often incomplete transsphenoidal resection or to preserve normal tissue and prevent surgical complications such as hypopituitarism, which affects up to 5% of patients post-surgery [22]. Prior studies have demonstrated efficacy of [68 Ga]-DOTATATE PET in differentiating primary and recurrent pituitary adenomas from normal pituitary tissue with significantly lower [68 Ga]-DOTATATE SUV than the normal pituitary gland, in agreement with the high SUV of the normal pituitary gland observed in our cohort [26,27]. Notably, [F18]-FDG PET avidity is higher in adenomas than normal pituitary tissue, and prior studies showed that the ratio of [F18]-FDG to [68 Ga]-DOTATATE may be the more accurate metric, suggesting the utility of dual tracer PET in the evaluation of pituitary adenomas [26,27]. Furthermore, [68 Ga]-DOTATATE PET has therapeutic potential to optimize patient selection for molecular targeted therapies using somatostatin analogs and peptide receptor radionuclide therapies (PRRT) such as Lu-177 DOTATATE, already approved for therapeutic use in SSTR positive neuroendocrine tumors (NETs). In demonstrating the baseline distribution of SSTR expression, our study benefits future diagnostic and therapeutic approaches targeting SSTR in pitNET by laying a basis for the expected range of normal pituitary DOTATATE avidity in the context of demographic characteristics.
Our analysis of MSH revealed a significantly larger pituitary gland in younger females compared to younger males. This finding, while only exclusive to the younger groups in our cohort, is in concordance with prior studies that found a significant sex difference in pituitary gland size, with females having a larger size on average than males [23][24][25]31]. In contrast, previous studies on the relationship between pituitary size and age have shown mixed results. One study with a cohort of less than 35 years of age demonstrated a positive correlation between pituitary size and age that was more robust in females [28]. On the other hand, another study with a cohort aged 21 to 82 showed age was inversely correlated with pituitary height, which was most pronounced in females [30]. In general, the pituitary gland reaches its maximum size in the second to third decade of life during puberty, followed by a gradual decline in both sexes with age [25,29]. Additionally, hormonal changes throughout one's life may also confound the observed relationship between MSH and age. For example, the pituitary gland transiently increases in size during pregnancy [41]. Nulliparous females were reported to have significantly larger pituitary glands than their multiparous counterparts [42]. Females using oral contraceptive pills experience significant reduction in pituitary volume than those who do not [43]. Menopausal females receiving estrogen replacement therapy had greater mean pituitary heights than those who did not [44].
There are several limitations to the study that must be acknowledged. While our cohort contained a non-homogenous sample, it was skewed towards females (73%) and towards older individuals (mean age = 58.9). Our cohort was also a subset from a larger pool of patients who underwent [68 Ga]-DOTATATE PET examinations for evaluation of their CNS SSTR2 positive tumors. No PET imaging was dedicated solely for the assessment of the pituitary gland. As a result, every patient in the cohort had a history of intracranial pathology, most predominantly meningioma. Furthermore, based on lack of pertinent clinical history of the majority of patients, there was paucity of data in the electronic medical record regarding FSH, menopause, TSH, oral contraception (OCP) use, BMI, and other forms of exogenous hormones, which precluded us from further investigation of the relationship between these variables and pituitary [68 Ga]-DOTATATE avidity. Pituitary gland volume may change with oral contraceptive use, pregnancy, nulliparity, hormone replacement therapy, obesity, and primary hypothyroidism [41][42][43][44][45][46][47]. Future studies in relation to the aforementioned variables will further enhance our understanding of [68 Ga]-DOTATATE PET in various clinical settings. Additionally, while the broad ranges and non-normal distribution of SUV/SUVR in our study may present a potential challenge to its clinical applicability, our study represents a stepping-stone for future studies focusing on the potential clinical utility of [68 Ga]-DOTATATE PET in pituitary neoplasms.
Lastly, in our intention to capture sex differences in pituitary gland, sex was collected as a binary variable: male and female, and chart review presumed that documented sex is unchanged from sex assigned at birth. Given sex is a multidimensional construct, our presumption of binary variable does not account for the entirety of the gender spectrum. It is also worth noting that to the best of our knowledge, our cohort had no transgender-identifying or gender diverse patients.

Conclusion
We present normative data of the SUV and size of the pituitary gland with [68 Ga]-DOTATATE PET/MR from a heterogenous cohort of 95 patients, stratified by age and sex. Our study provides a starting point for utilizing [68 Ga]-DOTATATE PET in clinical and research settings, potentially enhancing diagnosis and management of pathologies in the sella turcica and offering personalized approaches with molecular theragnostic agents. Our results generate interesting hypotheses relating pituitary biology to demographic factors, which require further corroboration with future studies both in vitro and in vivo. Future studies with larger, heterogenous populations, accounting for relevant clinical and demographic variables, and with dedicated pituitary imaging will further enhance our understanding of [68 Ga]-DOTATATE PET of the pituitary gland in clinical settings.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by SHK, SJCC, and JI. The first draft of the manuscript was written by SHK, SJCC, and JI and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data availability
The datasets analysed during the current study are available from the corresponding author on reasonable request.