Effect of distance from target on hypopituitarism after stereotactic radiosurgery for pituitary adenomas

Delayed hypopituitarism is the most common complication after stereotactic radiosurgery (SRS) for pituitary adenomas. The aim of this study was to investigate the relationship between the distance from the hypothalamic-pituitary axis to the treatment target and anterior pituitary function preservation after SRS. Between 2007 and 2020, consecutive adult patients who underwent single-session SRS for non-functioning or hormone-secreting pituitary adenomas with ≥ 6 months of follow-up were included. Distance measurements between hypothalamic-pituitary axis structures and the SRS target volume were quantified on MRI. The primary outcome was anterior pituitary function preservation. Outcomes were compared using multivariable regression and area under the receiver operator characteristic curve (AUROC) analyses. The study cohort comprised 224 patients, who were categorized by preservation (n = 168) and no preservation (n = 56) of anterior pituitary function after SRS. The mean and median clinical follow-up durations were 53.7 (38.0) and 46.0 (17.0–75.0) months, respectively. Independent predictors of anterior pituitary function preservation were a greater distance between the center of the pituitary gland and center of the SRS target [OR 1.101 (1.000–1.213), p = 0.050], and a shorter clinical follow-up duration [OR 0.985 (0.977–0.993), p < 0.0001]. The adjusted AUROC for the distance from the center of the pituitary gland and center of the SRS target in predicting anterior pituitary function preservation was 0.595. The sensitivity, specificity, positive predictive value and negative predictive value in predicting anterior pituitary function preservation at the optimal cut-off distance of 15 mm were 30.0, 88.0, 89.9 and 26.2%, respectively. Greater distance between the normal pituitary gland and the SRS target is associated with anterior pituitary function preservation and increasing this distance should be a goal of adenoma resection. Larger prospective, multi-center studies are necessary to corroborate this finding and establish the effects of distance on hypopituitarism after SRS for pituitary adenomas.


Introduction
Pituitary adenomas comprise 10-20% of all primary intracranial tumors, and they are classified into functioning and nonfunctioning adenomas based upon the presence and absence of endocrine secretory activity, respectively. [1] Resection is often the first-line treatment for functioning adenomas (FA) and larger nonfunctioning adenomas (NFA) with optic nerve compression [2]. Recurrence or progression occurs in 10-50% of pituitary adenomas after surgical resection [2,3]. Stereotactic radiosurgery (SRS) is an important adjuvant and salvage treatment for patients with subtotal resection or recurrence [4,5]. It has also been used as an upfront treatment for select patients with high surgical risk [4].
SRS leverages its ability to administer focused, highdose radiation in a single session while sparing nearby critical neurovascular structures of unwanted radiation [5]. However, anterior pituitary insufficiency still occurs in approximately 18-50% of patients after SRS for pituitary adenomas [6][7][8]. It is the most common intermediate to late complication, with a time-to-onset ranging from 2 months to 14 years after treatment [7,8]. Hypopituitarism is associated with increased mortality and significant health-related costs [9,10]. Higher treatment dose and differences in SRS dose intensities along the hypothalamic-pituitary (HP) axis have been identified as risk factors for post-SRS hypopituitarism [6,11]. We hypothesize that the distance between the radiosurgical target and the HP axis may be associated with SRS-induced endocrine complications. The aim of this retrospective cohort study was to determine the effects of distance between HP axis neuroanatomical structures and the SRS target volume on anterior pituitary function preservation after SRS for pituitary adenomas.

Study design
Between October 2007 and October 2020, consecutive adult patients who underwent single-session SRS at the University of Virginia Health System for treatment of pituitary adenomas were reviewed. The study was approved by the Institutional Review Board (IRB) committee at our institution and follows the guidelines set forth by the strengthening the reporting of observational studies in epidemiology (STROBE) statement [12]. Data were obtained retrospectively according to a standardized collection protocol, which included a medical chart abstraction and review of SRS planning MRI data.

Patient identification and selection
The inclusion criteria for this study were: (1) age ≥ 18 years, (2) available SRS planning MRI data, (3) presence of a nonfunctioning or hormone-secreting pituitary adenoma, and (4) ≥ 6 months of clinical neuroendocrine follow-up. Patients were excluded for the following criteria: (1) whole-sellar SRS, (2) absence of a discrete adenoma, normal pituitary gland or stalk on the SRS planning MRI and (3) history of cranial fractionated radiation therapy before the first SRS treatment. Patients underwent treatment consistent with our institutional standard, which included a comprehensive neuroendocrine assessment comprising clinical evaluation and laboratory testing of adrenocorticotropic hormone (ACTH), cortisol, luteinizing hormone (LH), follicle-stimulating hormone (FSH), insulin-like growth factor-1 (IGF-1), prolactin, thyroid-stimulating hormone (TSH), free thyroxine (T4), 24-h urinary free cortisol (patients with Cushing's disease), growth hormone (patients with acromegaly), testosterone (males) and estradiol (females < 50 years) prior to the initial SRS. Pre-treatment hypopituitarism was defined as anterior pituitary hormone deficiency below the expected age and gender specific reference range and a formal diagnosis by the treating endocrinologist, and/or requirement of hormone replacement therapy. The exception to this was growth hormone deficiency, which was defined as a positive result from provocative testing, or a formal diagnosis by the treating endocrinologist. Patients with subjective visual disturbance or signs of optic apparatus compression were also referred for a formal neuroophthalmological assessment. Routine imaging and neuroendocrine assessments were performed at 6-month intervals for the first year, and then yearly until 5 years after SRS. After 5 years, follow-up was performed at 1 to 3-year intervals.

SRS technique
The SRS technique for treatment of patients with pituitary adenomas at our institution has been previously described [11]. In brief, a high-resolution MRI was obtained before and after contrast administration, combined with fat suppression, at 1 mm slice thickness. A Leksell stereotactic Gamma Knife frame was placed under monitored anesthesia care. Following frame placement, a CT was obtained and merged with the high-resolution MRI. Dose planning was performed using the GammaPlan software (Elekta AB, Sweden) by a multidisciplinary team comprising a neurosurgeon, radiation oncologist, and medical physicist. SRS was performed using the Gamma Knife unit (Elekta AB, Sweden), and the specific model used (including models U, B, C, Perfexion, Icon) varied by the year.

Clinical data
Baseline demographic data included age and sex. Medical history data included history of pituitary adenoma resection, history of hypopituitarism, history of central diabetes insipidus, history of visual deficit, presence of a functioning adenoma and duration of clinical neuroendocrine followup. SRS planning data included margin dose, treatment volume, and isodose. Hypopituitarism was defined as anterior pituitary hormone deficiency, or requirement of hormone replacement therapy due to pituitary insufficiency. Visual deficit was defined as subjective visual disturbance, evidence of visual field loss, or ocular motility disturbance on formal neuroophthalmological evaluation.

Neuroimaging review
From the post-contrast fat suppressed dose planning MRI, a single two-dimensional coronal slice with a clearly identifiable pituitary adenoma, pituitary gland and pituitary stalk was selected. The treatment plan was outlined, and distance measurements were made by connecting points between the following structures: (1) edge of the pituitary gland to the prescription isodose line, (2) center of the pituitary gland to the center target within the prescription isodose line, (3) edge of the pituitary gland to the center target within the prescription isodose line, (4) center of the pituitary gland to the prescription isodose line, (5) edge of the pituitary stalk to the prescription isodose line, (6) center of the pituitary stalk to the center target within the prescription isodose line, (7) edge of the pituitary stalk to the center target within the prescription isodose line and (8) center of the pituitary stalk to the prescription isodose line. The MRI was further reviewed for a two-dimensional coronal slice where the adenoma and the hypothalamus were clearly identifiable. Additional distance measurements were made by connecting points between the following structures: (9) edge of the hypothalamus to the prescription isodose line, (10) center of the hypothalamus to the center target within the prescription isodose line, (11) edge of the hypothalamus to the center target within the prescription isodose line and (12) center of the hypothalamus to the prescription isodose line (Fig. 1). All measurements were performed by individuals blinded to clinical information.

Outcomes
The primary outcome was the preservation of pre-SRS anterior pituitary function. No preservation of pre-SRS anterior pituitary function was defined as a new anterior pituitary hormone deficiency below the expected age and gender specific reference range and a formal diagnosis by the treating endocrinologist, or requirement of a new hormone replacement therapy after the initial SRS. The exception to this was growth hormone deficiency, which was defined as a positive result from provocative testing, or a formal diagnosis by the treating endocrinologist. Patients who underwent repeat resection, SRS or adjuvant cranial fractionated radiation therapy were censored at the time of additional treatment.

Statistical analysis
Statistical analyses were performed using Stata (version 15.0, College Station, TX). Univariable logistic regression analyses, dichotomized by the primary outcome, compared baseline demographic, medical history, treatment and imaging data. A multivariable logistic regression analysis, initially including all potential predictor variables, with stepwise backward elimination of non-contributory variables (p > 0.10), was performed to identify independent predictors of the primary outcome. Statistical models were assessed for goodness-of-fit using the Hosmer-Lemeshow test. Covariates were tested for multicollinearity using tolerance and variance inflation factor, and collinear variables were removed from the final model. An area under the receiver operator characteristic curve (AUROC) analysis was performed to assess the discrimination of independent variables included in the multivariable logistic regression model. Cut-off points for distance measurements between HP axis neuroanatomical structures and the SRS target volume were calculated using the Youden's Index (YI) [13]. Accuracy was assessed using sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Statistical significance was defined as p < 0.05, and all tests were two-tailed. Missing data were not imputed.

Study cohort
Of the 466 eligible patients, 242 were excluded from the present study (153 patients for whole-sellar SRS, 4 patients for the absence of a discrete pituitary adenoma, pituitary gland or pituitary stalk on the dose planning MRI, 73 patients for < 6 months of clinical neuroendocrine follow-up, 4 patients for the lack of accessible medical history data, and 8 patients for a history of cranial fractionated radiation therapy, Fig. 2). The remaining 224 patients (mean age 50.1 ± 15.2 years; 52.2% female) were categorized by the primary outcome of preservation (n = 168) and no preservation (n = 56) of pre-SRS anterior pituitary function. 54.5% (n = 122) of the patients had a history of hypopituitarism prior to SRS. The rate of new or worsening hypopituitarism after SRS was 25.0% (n = 56) at a mean and median time to onset of 30.5 (30.0) and 18.0 (6.5-38.5) months, respectively. The rate of new or worsening hypopituitarism after SRS requiring hormone replacement was 16.5% (n = 37). The affected axis was adrenal in 12 patients (21.4%), thyroid in 44 patients (78.6%), gonadal in 25 patients (44.6%) and somatic in 16 patients (28.6%). More than one axis  Table 1 compares the baseline characteristics between the preservation and no preservation of pre-SRS anterior pituitary function cohorts. Univariable predictors of anterior pituitary function preservation were a greater distance

Effects of distance on endocrine outcomes
The adjusted AUROC for the distance between the center of the pituitary gland and center of the target volume in predicting anterior pituitary function preservation was 0.595 (0.506-0.684) ( Table 3, Figure 5A depicts the dose plan utilized at the time of her treatment. The distance between the center of the gland and the center of the SRS target was 9.2 mm. Figure 5B depicts an alternative plan utilizing a steeper fall off in dose towards the pituitary gland and stalk, thereby better shielding these structures from unwanted radiation dose. This may be considered in patients undergoing SRS dose planning with less than 15 mm between the center of the pituitary gland and the center of the SRS target volume.

Discussion
Resection is frequently the first-line treatment for symptomatic pituitary adenomas [14]. The prevalence of preoperative hypopituitarism in patients undergoing surgical resection ranges from 70 to 85%, and the incidence of new or worsening post-operative hypopituitarism ranges from 5 to 15% [15][16][17]. SRS offers adjuvant or salvage treatment for those with subtotal resection or recurrence, and an alternative to surgical resection for patients with high surgical risks [4,14]. The incidence of hypopituitarism after SRS, reported to range from 18 to 50%, is lower than the reported rates of 30-100% after fractionated radiotherapy [6-8, 11, 18-22]. However, pituitary insufficiency remains the most common adverse effect after SRS for pituitary adenomas [6,8]. Hypopituitarism is associated with significant morbidity, increased mortality, poorer quality of life, and greater health related costs [9,10,[23][24][25]. Given the frequent utilization of SRS as an adjuvant post-surgical treatment, identification of solutions to minimize the risk of new or worsening hypopituitarism after SRS for pituitary adenomas is critical.
In this study, we sought to identify relationships between HP axis to SRS target volume distances and anterior pituitary function preservation after SRS. We identified that a greater distance between the center of the pituitary gland and the center of the SRS target volume was an independent predictor of anterior pituitary function preservation. Based upon these findings, we performed an AUROC analysis to generate a predictive tool for anterior pituitary function preservation. A distance of 15 mm between the center of the gland and the center of the SRS target volume was associated with a specificity of 88% and PPV of 90% for anterior pituitary function preservation. Taken together, our observations suggest that distances of 15 mm or greater between the center of the gland and the center of the SRS target are associated with anterior pituitary function preservation, whereas distances of less than 15 mm may be associated with new or worsening hypopituitarism after SRS. Thus, a minimum 15 mm distance between the pituitary gland and any residual adenoma should be a goal of resections undertaken prior to SRS.
We also found an independent association between shorter clinical follow-up duration and anterior pituitary function preservation after SRS. This finding likely reflects a delay in the onset of SRS treatment-related hormonal adverse effects [6,8]. In our patient cohort, the median time to onset of new or worsening hypopituitarism was 18 months (IQR 6.5-36.5 months) after SRS. This was consistent with previous reports, which ranged between 21 and 120 months after SRS [8,11,22,[26][27][28]. The time to development of new endocrinopathy may also depend upon the affected hormone. In a cohort of 48 non-functioning adenoma patients treated with SRS, Gopalan et al. reported early onset of cortisol deficiency at a mean of 27 months, compared to gonadotropin deficiency, which occurred a mean of 120 months after SRS [22]. Thyroid hormone and growth hormone deficiency occurred at a mean of 40 and 71 months after SRS, respectively. While the majority new endocrinopathies develop within 5 years after SRS, late-onset hypopituitarism can occur in approximately 15% of cases [8]. In the context of the present study, these observations underscore the need for continued longitudinal endocrine follow-up in SRS-treated pituitary adenoma patients.
Treatment dose is a previously described risk factor for new hypopituitarism after SRS for pituitary adenomas [8,11,21,28,29]. Furthermore, higher doses to the pituitary gland or the pituitary stalk have been found to independently predict development of pituitary insufficiency after SRS [11,21,29]. Therefore, to mitigate these adverse radiation effects, Taussky et al. proposed the placement of a fat graft between the residual tumor and the normal gland at the time of surgery among patients expected to undergo adjuvant SRS or fractionated radiation therapy [14]. In their patient cohort, there were no new cases of pituitary insufficiency observed over a median follow-up period of 4 years, substantiating their hypothesis that a greater distance between the two anatomical structures can preserve anterior pituitary function after SRS. Our observation of an association between the distance from the center of the pituitary gland to the center of the SRS treatment target and new hypopituitarism may be related to the effects of radiation dose delivery to surrounding structures. Given that radiation falls off via an inverse square law, the dose fall off is proportional to the inverse of the square of the distance. Knowledge of this relationship between dose and distance should be utilized during pituitary adenoma resection to safely maximize the distance between the two structures among patients expected to undergo adjuvant SRS. In the present analysis, distance between the center of the gland and the center of the isodose line was a significant predictor of anterior pituitary function preservation, while margin dose was not. This suggests that distance may carry a greater importance than margin dose in SRS treatment planning. However, these preliminary findings should be interpreted with caution, and their validity and generalizability warrant further evaluation in a multicenter study. In patients with smaller distances between SRS target and normal pituitary gland, a steeper dose fall off towards the gland through the use of advanced radiosurgical planning techniques should be emphasized to preserve anterior pituitary function.

Limitations
There are a number of limitations to this study which must be acknowledged. Most importantly, our retrospective study design is subject to confirmation bias, in that variables were chosen based upon data availability and hypothesis generation. Importantly, we did not have data on residual gland or tumor volume were not readily available, which limits the interpretation of potential correlations between tumor size 1 3 and distance measurements, and between residual gland and distance measurements. Our results are contingent upon the accuracy of reported data, which was largely based upon medical chart abstraction. Due to the retrospective nature of our study, a diagnosis of hypopituitarism was made by the treating endocrinologist. Furthermore, the power of our analysis was limited by the relatively small number of patients who developed new or worsening hypopituitarism. A larger, multicenter study from the International Radiosurgery Research Foundation (IRRF) is underway to verify the findings of the current study. Measurements between the SRS treatment target and HP axis structures may be subject to restricted generalizability and measurement bias. Loss to follow-up may have contributed to selection bias, resulting in outcome differences. While this was adjusted for in our analyses, it is not possible to entirely account for the possibility of late-onset hypopituitarism in patients with insufficient follow-up. Additional limitations include reporting, recall and missing data biases, due to the sourcing of the study data from a single center. Finally, our AUROC analysis found the distance between the gland and the SRS target to be a weak predictor of anterior pituitary function preservation.

Conclusions
Greater distance between the center of the pituitary gland and the center of the SRS target volume appears to be associated with anterior pituitary function preservation after SRS for pituitary adenomas. A distance of 15 mm or more between the center of the pituitary gland and the center of the SRS target volume predicts anterior pituitary function preservation. This finding underscores the importance of dose and distance in SRS radiosurgical planning and the relevance of maximizing the distance between the pituitary gland and the residual adenoma during resection. Larger prospective, multi-center studies are necessary and underway to corroborate this finding and determine the effects of distance on hypopituitarism after SRS for pituitary adenomas.
Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis was performed by NI. Statistical guidance was provided by ZX and C-JC. The first draft of the manuscript was written by NI and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding None.
Data availability This study data is available from the corresponding author upon reasonable request.

Conflict of interest The authors declare no conflict of interest.
Ethical approval This article utilizes retrospective data and does not contain any studies with human participants performed by any of the authors.