Is it possible to avoid hypopituitarism after irradiation of pituitary adenomas by the Leksell gamma knife?

Transcription

1 European Journal of Endocrinology (2011) ISSN CLINICAL STUDY Is it possible to avoid hypopituitarism after irradiation of pituitary adenomas by the Leksell gamma knife? Josef Marek, Jana Ježková, Václav Hána, Michal Kršek, L ubomíra Bandúrová, Ladislav Pecen 1, Vilibald Vladyka 2 and Roman Liščák 2 Third Department of Medicine, First Medical Faculty, Charles University, U nemocnice 1, Praha 2, Czech Republic, 1 Institute of Informatics of the Czech Academy of Science, Pod Vodárenskou věží 2, Praha 8, Czech Republic and 2 Department of Stereotactic and Radiation Neurosurgery, Hospital Na Homolce, Roentgenova 2, Praha 5, Czech Republic (Correspondence should be addressed to J Ježková; fjjezek@cmail.cz) Abstract Objective: Radiation therapy is one of the treatment options for pituitary adenomas. The most common side effect associated with Leksell gamma knife (LGK) irradiation is the development of hypopituitarism. The aim of this study was to verify that hypopituitarism does not develop if the maximum mean dose to pituitary is kept under 15 Gy and to evaluate the influence of maximum distal infundibulum dose on the development of hypopituitarism. Design and methods: We followed the incidence of hypopituitarism in 85 patients irradiated with LGK in The patients were divided in two subgroups: the first subgroup followed prospectively (45 patients), irradiated with a mean dose to pituitary!15 Gy; the second subgroup followed retrospectively and prospectively (40 patients), irradiated with a mean dose to pituitary O15 Gy. Serum TSH, free thyroxine, testosterone or 17b-oestradiol, IGF1, prolactin and cortisol levels were evaluated before and every 6 months after LGK irradiation. Results: Hypopituitarism after LGK irradiation developed only in 1 out of 45 (2.2%) patients irradiated with a mean dose to pituitary!15 Gy, in contrast to 72.5% patients irradiated with a mean dose to pituitary O15 Gy. The radiation dose to the distal infundibulum was found as an independent factor of hypopituitarism with calculated maximum safe dose of 17 Gy. Conclusion: Keeping the mean radiation dose to pituitary under 15 Gy and the dose to the distal infundibulum under 17 Gy prevents the development of hypopituitarism following LGK irradiation. European Journal of Endocrinology Introduction Treatment of pituitary adenomas is a complex issue, involving neurosurgical, pharmacological and radiation treatment modalities. The success rates of neurosurgical treatment vary from 10 to 90% depending on the type of adenoma (1 6). Medical treatment often has to be continued lifelong and does not always control the growth and hormonal secretions of adenomas. Radiation therapy is typically used when both previous methods have failed. Conventional fractionated radiotherapy has achieved good results but only after a long latency and with considerable post-radiation morbidity. The main adverse effect is hypopituitarism, which is reported in 50 80% of patients followed up 10 years after conventional irradiation (7 9). Other less frequent side effects of conventional radiotherapy are optic neuropathy, radionecrosis, radiation-induced cerebral tumours and vascular injury (10 14). Radiosurgery primarily involves stereotactic irradiation with the Leksell gamma knife (LGK), or, alternatively, with a stereotactic linear accelerator. The hormonal normalization after LGK treatment takes several years but is considered to be faster than with conventional radiotherapy. Stereotactic radiosurgery with the LGK is supposed to decrease the incidence of side effects because of its highly consistent dose delivery, which is associated with less irradiation outside the targeted volume. Optic neuropathy is reported in 1% of patients who were irradiated by the LGK (15, 16), and the risk of damage to cranial nerves in the cavernous sinus is!1%. Moreover, these cranial neuropathies are often temporary (16). Radionecrosis in the hypothalamic and temporal regions related to LGK irradiation was described in!1% of patients as well. However, the majority of these patients underwent conventional fractionated radiation prior to LGK irradiation (16). As far as vascular injury is concerned, symptomatic carotid artery stenosis was reported in only two cases (17). Up until now, radiation-induced neoplasm has not been reported. However, LGK irradiation was reported not to eliminate the incidence of hypopituitarism. Published q 2011 European Society of Endocrinology DOI: /EJE Online version via

2 170 J Marek and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 data vary widely, ranging from 0 to 40% (18 20). Similarly, among our first patients treated with the LGK, 38.2% of them developed hypopituitarism (21). Consequently, in our previously published study, we analysed factors leading to this unfavourable outcome and evaluated the radiation tolerance of functioning sellar structures after LGK irradiation (22). This retrospective study demonstrated that the mean dose to the pituitary gland surrounding the adenoma is the most important factor causing the development of hypopituitarism. The radiation dose of 15 Gy was determined as the maximum safe limit of the mean radiation dose to the pituitary gland. Another factor, which influenced the development of hypopituitarism, was the maximal dose to the distal infundibulum. However, the full characterization of this relationship was not defined. Pretreatment factors such as the previous neurosurgery, previous partial pituitary deficiency and the tumour volume were demonstrated to play less important role in the development of hypopituitarism. On the basis of our previous report, the aim of this study was to verify that hypopituitarism does not develop if the maximum mean dose to the healthy pituitary is kept under 15 Gy and to evaluate if the maximum dose to distal infundibulum influences the development of hypopituitarism in a prospectively prolonged study. Subjects and methods We followed a group of 85 patients (54 females and 31 males, aged years, mean age 45.14G16.13 years) with pituitary adenomas, treated with the LGK at the Department of Stereotactic and Radiation Neurosurgery at the Na Homolce Hospital in Prague, Czech Republic. The follow-up period was months (meangs.d., 100.9G39.9 months; median 84 months). LGK irradiation was indicated as I) secondary therapy after incomplete surgery; II) primary therapy i) in patients contraindicated for surgery or ii) in patients refusing surgery. The group of 85 patients was divided in two subgroups according to the mean dose of radiation to the pituitary tissue surrounding the adenoma. The first subgroup (group I) consisted of 45 patients who were irradiated with a mean dose to the pituitary of!15 Gy. The second subgroup (group II) consisted of 40 patients who were irradiated with a mean dose to the pituitary of more than 15 Gy. The patients underwent LGK irradiation between 1993 and For the retrospective part of the study, all patients were included, in whom the radiation doses to the pituitary and pituitary stalk were measured (all patients from group II and six patients from group I), and for the prospective part of the study, all new patients with pituitary adenomas irradiated between September 2001 and September 2003 (39 patients from group I). In group I, there were 28 females and 17 males aged years (mean 48.6G15.9 years); 21 patients with acromegaly, 10 patients with prolactinomas, 5 patients with Cushing s disease and 9 patients with functionless adenomas. Twenty-two (48.9%) patients underwent neurosurgical operations prior to LGK irradiation. None of the patients underwent the fractionated radiotherapy prior to LGK irradiation. The follow-up period in group I ranged from 60 to 192 months (meangs.d., 77.8G23.7 months; median 73 months). In group II, there were 26 females and 14 males aged years (mean 41.2G15.6 years); 23 patients with acromegaly, 8 patients with prolactinomas, 6 patients with Cushing s disease, 1 patient with Nelson s syndrome and 2 patients with functionless adenomas. Twenty-five (62.5%) patients underwent neurosurgical operations prior to LGK irradiation. None of the patients underwent the fractionated radiotherapy prior to LGK irradiation. The follow-up period in group II ranged from 60 to 192 months (meangs.d., 126.9G38.6 months; median 135 months). None of the patients in both groups had surgery or radiation therapy after LGK radiation. Radiosurgery was performed using an LGK model B (Elekta Instrument AB, Stockholm, Sweden). Stereotactic imaging was performed by magnetic resonance imaging (MRI; Magnetom Expert 1T, Siemens, Erlangen, Germany), using classical native turbo spin-echo (TSE) sequences in T2-weighted (4100/99) and T1-weighted (500/15) axial and coronary cuts of 3 mm thickness, before and after contrast. In some patients, images were obtained by dynamic TSE T1-weighted (395/12) sequence. The accuracy of our MRI targeting was evaluated in a separate study (23). Irradiation was designed using GammaPlan 5.11 (Elekta Instrument AB). However, in the first year of LGK therapy (1993), a different planning system was used (Kula, Elekta Instrument AB). Pituitary radiosurgery planning is a complex procedure because a highly conformal dose plan is needed to spare any remaining normal pituitary gland as well as the optic pathway. The examples of pituitary radiosurgery planning are showed in Figs 1 and 2. Of the 85 patients, 6 were treated according to the planning system Kula, by which it was impossible to calculate the accurate dose to all specific structures. In the case of these six patients, the radiosurgery planning was later reconstructed into the system GammaPlan, which enabled accurate calculation. Data pertaining to the radiosurgical target volume and radiation doses to the pituitary adenoma, healthy pituitary and distal infundibulum are summarized in Table 1. Between both groups of patients, there was no significant difference in the radiosurgical target volume (non-significant P levels in Wilcoxon and ANOVA tests (majority of distribution of results also shows moderate violation of normality assumptions; therefore, both these tests were used)). There were found significant differences in radiation doses to the pituitary adenoma

3 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 R 8 Gy 8 Gy (both functionless and secretory adenomas). Both maximal (Wilcoxon P!0.004 and ANOVA P!0.013) and margin (Wilcoxon P!0.022 and ANOVA P!0.012) doses were higher in the group II in comparison with the group I. When comparing the patients only with hormone-secreting adenomas, there were found no significant differences in the radiosurgical target volume (group I 1636 mm 3 versus group II 1655 mm 3 ) and margin dose (group I 33.4 Gy versus group II 35.5 Gy) between both groups of patients. The difference was only in maximal dose (group I 64.5 Gy versus group II 67.8 Gy; Wilcoxon P!0.036 and ANOVA P!0.19 (NS)). The dose to the optic pathway was kept!8 Gy, and the dose to the brain stem was kept!14 Gy. The regular hormonal follow-up was carried out at a single centre (Third Department of Medicine, First School of Medicine, Charles University, Prague, Czech Republic). The tests were carried out every 6 months in order to establish the normalization of hormonal overproduction and the possible pituitary deficits after irradiation. To evaluate the success of treatment of the different pituitary adenomas, the following criteria of hormonal normalization were followed: i) in acromegaly patients, normal IGF1 according to sex and age; ii) in prolactinoma patients, prolactin (PRL)!619 miu/l in non-pregnant women, PRL!430 miu/l in postmenopausal women and PRL!375 miu/l in men; iii) in S 35 Gy Figure 1 Fifty years old women with Cushing s disease. The target tumour volume was 1600 mm 3. The tumour received a margin dose of 35 Gy at the 50% isodose line. The maximum dose to the optic pathway was 7.6 Gy. The volume of hypophysis (*) was mm 3, the minimum dose to hypophysis was 3 Gy, the maximum dose was 29.1 Gy, the mean dose was 7.8G4.2 Gy and the integral dose was 1.6 mj. patients with Cushing s disease, an 0800 h plasma cortisol and 24-h free urinary cortisol in the normal range, and either suppressibility of plasma cortisol after an overnight dexamethasone (1 mg) suppression test with 0800 h cortisol level below 84 nmol/l or the restitution of circadian variability of plasma cortisol levels. The activity of the pituitary thyroid axis was determined by measuring peripheral levels of free thyroxine (ft 4 ). In order to exclude a diagnosis of incident primary hypothyroidism, TSH level was measured. Activity of the pituitary adrenocortical axis was monitored by measuring the levels of morning plasma cortisol (between 0800 and 0900 h). Cortisol levels O450 nmol/l at this time excluded adrenal insufficiency, while a level!150 nmol/l was diagnostic of hypocortisolism. In all patients with morning plasma cortisol levels between 150 and 450 nmol/l, the measurements of cortisol levels in response to insulininduced hypoglycaemia ( IU/kg insulin human rapid as i.v. bolus; samples drawn at 0, 30, 60 and 90 min) were performed. If cortisol level rose to 500 nmol/l, hypocortisolism was excluded. The activity of the pituitary gonadal axis was measured in men by determining plasma testosterone levels and in women by measuring serum levels of 17b-oestradiol, and assessed by the presence of regular menstrual bleeding. R 8 Gy Hypopituitarism after LGK irradiation Gy 8 Gy Figure 2 Seventy-two years old men with the functionless pituitary adenoma, who underwent one previous neurosurgery operation before LGK irradiation. The target tumour volume was 2800 mm 3. The tumour received a margin dose of 15 Gy at the 50% isodose line. The maximum dose to the optic pathway was 8.0 Gy. The volume of hypophysis (*) was mm 3, the minimum dose to the hypophysis was 6.4 Gy, the maximum dose was 19.0 Gy, the mean dose was 12.3G1.9 Gy and the integral dose was 1.5 mj. S

4 172 J Marek and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 Table 1 Radiosurgical target volume and radiation doses. Range Average Median Group I: the mean dose to the pituitary!15 Gy Target tumour volume (cm 3 ) Radiation doses (Gy) Adenoma maximum dose Adenoma margin dose Mean dose to the pituitary Dose to the distal infundibulum Group II: the mean dose to the pituitary O15 Gy Target tumour volume (cm 3 ) Radiation doses (Gy) Adenoma maximum dose Adenoma margin dose Mean dose to the pituitary Dose to the distal infundibulum The gonadotrophic deficiency was not evaluated in 28 women older than 45 years, because of the impossibility to decide whether the deficiency is caused by irradiation or whether it is the consequence of the natural menopause, and in 21 men over 50 years, because in them the testosterone levels can irregularly decrease as the consequence of andropause. As IGF1 level is not a reliable criterion for diagnosis of GH deficiency (24) and the determination of GH levels following provocative testing was not carried out, the activity of somatotropic axis was not evaluated. The development of diabetes insipidus was not observed after LGK irradiation. All hormonal levels were measured by commercial kits: GH and IGF1 by IRMA kits (Immunotech, Marseille, France); PRL by chemiluminescence analysis (ADVIA Centaur Bayer); plasma cortisol and urinary free cortisol by RIA kit (Immunotech); testosterone and oestradiol by chemiluminescence analysis (Architect Abbott); TSH and ft 4 by chemiluminescence analysis (Centaur Bayer). Statistical analysis Descriptive statistics for numerical data (meangs.d., median) and categorical data (absolute and relative frequencies) are presented. Inferential statistical analysis was done by means of the Cox proportional hazard regression model and logistic regression model; comparison of groups was done by means of ANOVA and Wilcoxon Rank Sum tests. Some of the variables were logarithmically transformed to normalize distribution of results for the purpose of statistical testing. For an estimation of time to an event (development of hypopituitarism in total and in individual axes), the product limit method (also called the Kaplan Meier method) was used; this is one variant of a nonparametric estimation. Mutual relationship between numerical parameters was investigated via correlation analysis done by means of Spearman rank correlation coefficient. Results Hypopituitarism after LGK irradiation did not develop in 97.8% (95% confidence interval (CI) (93.5; 100.0)) of patients in group I: the patients irradiated with a mean dose to the pituitary of!15 Gy. In contrast, hypopituitarism after LGK irradiation did not develop in only 27.5% (95% CI (13.6; 41.3)) of patients in group II: the patients irradiated with a mean dose to the pituitary of O15 Gy. In group I, the pituitary thyroid axis was evaluated in 36 patients and the pituitary adrenocortical axis in 44 patients who received no corresponding replacement therapies before irradiation. Neither thyroid nor adrenocortical deficiency developed after LGK irradiation. The pituitary gonadal axis was evaluated in only 13 patients, who were not replaced by sex hormones and fulfilled age criterion (women up to 45 years and men up to 50 years of age). Gonadal deficiency was developed by one patient (7.7%) 12 months after irradiation (Table 2). This patient had undergone two previous pituitary surgeries and had already had central hypothyroidism before he was irradiated. Table 2 Development of hypopituitarism in individual axes. Axis Number of axes with replacement therapy Number of axes without replacement therapy Number (%) of patients developing a new deficiency after irradiation Mean dose to the pituitary!15 Gy Thyroid (0%) Adrenal (0%) Gonadal a (7.7%) Mean dose to the pituitary O15 Gy Thyroid (58.3%) Adrenal (40.6%) Gonadal a (53.3%) a Gonadal axis was evaluated in women only up to 45 years and in men only up to 50 years of age.

5 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 Hypopituitarism after LGK irradiation 173 Table 3 Development of hypopituitarism related to previous pituitary deficiency before LGK irradiation. Number (%) of patients Number (%) of patients developing a new deficiency after irradiation Number (%) of patients without a new deficiency after irradiation Mean dose to the pituitary!15 Gy Patients receiving replacement therapy Patients without replacement therapy Mean dose to the pituitary O15 Gy Patients receiving replacement therapy Patients without replacement therapy 11 (24.4%) 1 (9.1%) 10 (90.9%) 34 (75.6%) 0 (0%) 34 (100%) 15 (37.5%) 14 (93.3%) 1 (6.7%) 25 (62.5%) 15 (60%) 10 (40%) There were no statistically significant differences between both groups in the number of patients with and without replacement therapy. In group II, 29 patients (72.5%) developed hypopituitarism in at least one axis months (mean GS.D., 41.9G30.6 months; median 33 months) after irradiation. The pituitary thyroid axis was evaluated in 36 patients, who were not replaced with T 4 after irradiation. Twenty-one patients (58.3%) experienced subnormal serum ft 4 levels following LGK irradiation. The pituitary adrenal axis was studied in 32 patients, who were not replaced by hydrocortisone after irradiation. Thirteen of them (40.6%) developed cortisol deficiency after irradiation. The pituitary gonadal axis was evaluated in 15 patients, who were not replaced by sex hormones and fulfilled age criterion. Eight of these patients (53.3%) developed sex hormone deficiency (Table 2). In summary, new pituitary deficiency developed at least one pituitary axis in 93.3% of patients already receiving replacement therapy to another axis before LGK irradiation and in 60% of patients without replacement therapy before LGK irradiation (Table 3). The development of hypopituitarism was observed in 88.0% of patients with previous neurosurgery before LGK irradiation and in 46.7% of patients without previous neurosurgery before LGK irradiation (Table 4). The mean dose to the pituitary (dose!15 vs O15 Gy) was statistically significant for the development of hypopituitarism in a group of 47 patients with previous neurosurgery (P!0.0001) and in a group of 38 patients without previous neurosurgery (P!0.0003). Similarly, the mean dose to the pituitary (dose!15 vs O15 Gy) was statistically significant for the development of hypopituitarism in a group of 26 patients receiving replacement therapy before LGK irradiation (P!0.0001) and in a group of 59 patients without replacement therapy (P!0.0001). Percentage of patients without the development of hypopituitarism according to the Kaplan Meier curve in group II (the patients irradiated with a mean dose to the pituitary of more than 15 Gy) is shown in Fig. 3. The hypopituitarism developed only in one out of all patients irradiated with a mean dose to the pituitary of!15 Gy during the follow-up of 5 years and in 5, 40 and 55% of patients irradiated with a mean dose to the pituitary of more than 15 Gy within 1, 3 and 5 years after LGK irradiation. Table 4 Development of hypopituitarism related to previous neurosurgical treatment before LGK irradiation. Number (%) of patients Number (%) of patients developing a new deficiency after irradiation Number (%) of patients without a new deficiency after irradiation Mean dose to the pituitary!15 Gy Patients with previous neurosurgery Patients without previous neurosurgery Mean dose to the pituitary O15 Gy Patients with previous neurosurgery Patients without previous neurosurgery 22 (48.9%) 1 (4.5%) 21 (95.5%) 23 (51.1%) 0 (0%) 23 (100%) 25 (62.5%) 22 (88.0%) 3 (12.0%) 15 (37.5%) 7 (46.7%) 8 (53.3%) There were no statistically significant differences between both groups in the number of patients with and without previous neurosurgery.

6 174 J Marek and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 Percentage of patients without hypopituitarism Time in years Figure 3 Percentage of patients without the development of hypopituitarism (Kaplan Meier estimate) in group II: the patients irradiated with a mean dose to the pituitary of more than 15 Gy. Between group I and group II, the significant differences were demonstrated in the mean dose to the pituitary (Wilcoxon P! and ANOVA P!0.0001) and the dose to the distal infundibulum (Wilcoxon P! and ANOVA P!0.0001). According to the univariate and multivariate analyses, both the mean dose to the pituitary and the dose to the distal infundibulum are the most important and independent factors for the development of hypopituitarism. Statistical significance of other factors has not been proved (Table 5). Correlation analysis according to Spearmen rank correlation coefficient demonstrated that the dose to the distal infundibulum correlates with the mean dose to the pituitary; this correlation is more expressed in patients irradiated with a mean dose to the pituitary of more than 15 Gy (r s 0.63) than in patients irradiated with a mean dose to the pituitary of!15 Gy (r s 0.35). The threshold level of the mean dose to the pituitary and the maximum dose to the distal infundibulum for the development of hypopituitarism was evaluated according to the Cox model. The threshold levels of 18.2 Gy for the mean dose to the pituitary and 16.8 Gy for the maximum dose to the distal infundibulum were found. When comparing both groups of patients according to the success of the treatment, hormonal normalization was achieved in 42.8% of acromegaly patients, 50% of prolactinoma patients, 80% of patients with Cushing s disease in the group I and in 65.2% of acromegaly patients, 37.5% of prolactinoma patients and 83.3% of patients with Cushing s disease in the group II at the time of 5 years after LGK irradiation. All differences in hormonal normalization between both groups were not found to be significant (non-significant P levels in c 2 - test). In the group of acromegaly patients, the effect of LGK irradiation depended on the baseline hormonal activity of adenoma; the less active adenomas were normalized earlier (Wilcoxon P! and Kruskal Wallis P!0.0047). Growth of the adenoma after LGK irradiation was not observed in any of the groups. Discussion Radiation therapy is one of the treatment options for pituitary adenomas, which is used especially when neurosurgical treatment and pharmacological treatment have failed. Radiation treatment (both conventional fractionated irradiation and radiosurgery) is reported to be associated with the development of side effects, especially with the development of hypopituitarism. With regard to the development of hypopituitarism after conventional radiotherapy, the total absorbed dose to the hypothalamo pituitary axis is the major factor determining the risk and speed of radiation-induced hypopituitarism (25). The incidence of hypopituitarism increases with time after irradiation (13, 25). It has been observed that the somatotrophic axis is the most radiosensitive one, followed by the gonadal, adrenocortical and TSH axes. Furthermore, the more the pituitary function is disturbed prior to conventional radiotherapy, the greater the incidence of hypopituitarism. The incidence of hypopituitarism varies in different published studies and is reported to rise up to 80% 10 years after conventional radiotherapy (7, 9). A high number of pituitary deficiencies are reported even in recent studies of fractionated stereotactic radiotherapy (FSRT). Schalin-Jäntti et al. (26) referred the development of new hypopituitarism in 40% of patients during a mean follow-up of 5.25 years; Roug et al. (27) reported the development of hypopituitarism in 29% of patients with a median of 48 months after FSRT. Data regarding the incidence of hypopituitarism following LGK irradiation differ in published studies as well (Table 6). Possible explanations for the different incidence of hypopituitarism following LGK irradiation include different lengths of follow-up after irradiation, different sizes of adenomas, previous damage of pituitary and above all different radiation doses. Our patients with pituitary adenomas were treated with an LGK model B (Elekta Instrument AB). It is an Table 5 Prognostic factors for the development of hypopituitarism. Prognostic factors Multivariate Cox regression model (stepwise selection) P value Univariate Cox regression model P value Tumour volume NS NS Adenoma maximum dose NS NS Adenoma margin dose NS NS Mean dose to the pituitary ! Dose to the distal ! infundibulum Previous neurosurgery NS NS Replacement therapy NS NS NS, non-significant.

7 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 Hypopituitarism after LGK irradiation 175 Table 6 Overview of the development of hypopituitarism after LGK irradiation. Study Number of Target tumour patients NFA a /SA b volume (mm 3 ) Maximal dose (Gy) Margin dose (Gy) Follow-up (months) Time to the development of hypopituitarism (months) Percentage of patients developing hypopituitarism (29) 25 0/25 (GH, prolactin and ACTH) Mean 1340G830 ( ) Mean 28G6 (12 35) Mean 20 (6 36) Range (30) 16 0/16 (GH) Mean 50 Mean 25 Mean 16.8 (6 31.2) 0 (31) 65 22/43 (GH, prolactin and ACTH) Mean 48.5 Mean 25.4 (15 36) Mean 25.5 (3 54) 1.5 (32) 37 0/37 (prolactin, GH and ACTH) Mean 900 ( ) Mean 54.8 (35 80) Mean 28.7 (18 40) Mean (33) 73 31/42 (prolactin, GH and ACTH) Mean 4400 ( ) Mean 31.6 ( ) Mean 28.9G (34) 43 0/43 (ACTH) Mean 47 (12 60) Median 20 (3.6 30) Median 44 (18 113) Mean 21 (8 35) 16 (28) 17 0/17 (GH) Mean 25 (20 40) Mean (20) 92 61/31 (prolactin, GH and ACTH) Mean 3800 ( ) Mean 15 ( ) Mean 55.2G28.8 Median (19) 30 30/0 Median 1700 ( ) Median 29.1 ( ) Median 16.0 ( ) Median 57.7 ( ) 10 (18) 42 42/0 Mean 32G8 (20 70) Mean 16.2G4 (10 34) Mean 31.2 (6 102) 0 (17) 43 0/43 (prolactin, GH and ACTH) Median 4300 ( ) Median 40 (30 60) Median 20 ( ) Median 36 (12 108) 16 (35) 33 33/0 Median 5000 ( ) Median 36 (30 40) Median 16 (12 40) Median 43 (16 106) Median 24 (19 48) 28 (36) 30 0/30 (GH) Median 1430 ( ) Median 20 (15 35) Median 46 (9 96) Range (37) 78 56/22 (prolactin, GH and ACTH) Median 2300 ( ) Median 30 (20 32) Median 15 (14 16) Median 36 (9 100) 4 (38) 54 54/0 Mean 2300 ( ) Mean 33.2G0.7 (24 42) Mean 16.6G0.4 (12 21) Mean 41.1G (39) 82 0/82 (GH) Median 25 (12 40) Mean 49.5 (6 108) Range (21) 96 0/96 (GH) Median 1350 ( ) Median 70 (26 80) Median 35 (10 42) Median 54 (12 120) Median 30 (6 72) 38.2 (40) 37 0/37 (GH) Mean 49.5 (25 70) Mean 29.3 ( ) Range (41) /0 Mean 4800 ( ) 41.5 (10 70) Mean 18.5 (5 25) Mean 47.9 (6 127) Mean 26 (8 107) 19.7 (42) 28 0/28 (PRL) Median 3400 ( ) Mean 43.1 ( ) Mean 18.9 (0.3 25) Median 52 (15 122) Mean 44 (33 51) 29 (43) 46 0/46 (GH) Median 3300 ( ) Median 43.5 (30 60) Median 21.5 ( ) Median 63 (22 168) Median 32 (12 120) 33 (44) 90 0/90 (ACTH) Mean 49 (18 60) Mean 23 (8 30) Median 45 (12 132) Mean 16 (4 36) 22 (45) 62 62/0 Median 4000 ( ) Median 34.5 Median 16 (11 20) Median 64 (23 161) Median (46) 40 0/40 (ACTH) Median 521 ( ) Median 29.5 (15 40) Median 48 (12 120) Range (47) 35 0/35 (PRL) Median 980 ( ) Median 70 (40 80) Median 34 (20 49) Median 66 (18 138) Median 54 (31 103) 14.3 a Non-functional adenomas. b Hormone-secreting adenomas.

8 176 J Marek and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 optimal tool for radiosurgery in the sellar region thanks to the distribution of radioactive Cobalt-60 sources in space, which allows a steep decline of isodoses at the margin. In spite of this tool, we recorded a high incidence of hypopituitarism in our first patients treated by LGK. This fact led us to analyse factors inducing the development of hypopituitarism. In our previously published study (22), according to the statistical analysis (univariate and multivariate methods) of the potential risk factors responsible for radiation-induced hypopituitarism, it was found that the most important factor is the mean dose to the pituitary gland. When this dose did not exceed 15 Gy, no impairment of thyroid or gonadal function was observed, rising to 18 Gy for adrenocortical function. The cumulative risk of adrenocortical deficiency development was calculated at 85% after 90 months when the mean dose on the healthy pituitary exceeded 20 Gy. Similarly, when the dose was 17 Gy (or more) on healthy pituitary, the cumulative risk of hypogonadism was 60%, and of hypothyreosis was 85% (also calculated for the period of 90 months). On the basis of this analysis, a radiation dose of 15 Gy was determined as the maximum safe limit of the mean radiation dose to the pituitary gland. In our group of patients, hypopituitarism developed in only one (2.2%) patient among those irradiated with a mean dose to the pituitary of!15 Gy in contrast to 29 (72.5%) of the patients irradiated with a mean dose to the pituitary of more than 15 Gy. The follow-up period for patients irradiated with a mean dose to the pituitary of!15 Gy was shorter than the follow-up period for a group of patients irradiated with a mean dose to the pituitary of more than 15 Gy (follow-up median 73 vs 135 months). However, when comparing both subgroups of followed patients at the same time point of 5 years after LGK irradiation, it was found that hypopituitarism developed in only 1 (2.2%) patient among those irradiated with a mean dose to the pituitary of!15 Gy and in 22 (55.0%) patients among those irradiated with a mean dose to the pituitary of more than 15 Gy. In this study, for the entire group of 85 patients, the radiation dose of 18.2 Gy was calculated as the maximum safe limit of the mean radiation dose to the pituitary gland. As mentioned above, in our previously published study (22), it was found that hypothyreosis or hypogonadism did not develop, when the mean dose to the pituitary gland did not exceed 15 Gy, rising to 18 Gy for adrenocortical deficiency. With regard to the results of both observations, we recommend keeping the mean dose to pituitary gland under 15 Gy. Information about the mean dose to the pituitary is also given in the study by Ikeda et al. (28). In their study, 17 acromegalic patients irradiated with LGK after transsphenoidal surgery were followed. The dose to the pituitary did not exceed 10 Gy, and the mean follow-up was 4.9 years. Of 17 patients, 3 exhibited panhypopituitarism before LGK irradiation. Hypopituitarism was not observed in any of the 14 remaining patients. Data referring to the development of hypopituitarism related to previous partial pituitary deficiency and neurosurgery before LGK irradiation support the importance of keeping the mean radiation dose to the pituitary gland under 15 Gy. In patients receiving some replacement therapy before LGK irradiation, a new deficiency developed in only one of patients irradiated with a mean dose to the pituitary of!15 Gy in comparison with 93.3% of patients irradiated with a mean dose to the pituitary of O15 Gy. Similarly, in patients undergoing neurosurgery before LGK irradiation, a new deficiency developed in only one of patients irradiated with a mean dose to the pituitary of!15 Gy in comparison with 88.0% of patients irradiated with a mean dose to the pituitary of O15 Gy. In our previous study, it was confirmed that the maximum dose to the distal infundibulum plays a similarly important role in the development of hypopituitarism. Results proved that the maximum dose to the distal infundibulum is an independent factor for the development of hypopituitarism. This dose should be kept under 17 Gy. In the study published by Feigl et al. (20), it was demonstrated that the dose to the pituitary stalk was significantly higher in patients with deterioration of pituitary function as compared with those in whom function remained unchanged. When comparing the success of hormonal normalization, there were no significant differences between both groups of patients. In agreement with our previously published data relating to acromegaly patients, the effect of the LGK therapy depends on the baseline hormonal activity of the adenoma the less active adenomas are normalised earlier (21). Hypopituitarism is the most common side effect associated with LGK irradiation. Our study proves that keeping the mean radiation dose to the pituitary gland under 15 Gy and the infundibulum dose to levels!17 Gy prevents the development of hypopituitarism following LGK irradiation. It is known that hypopituitarism can develop many years after irradiation. We observed the results of a mean dose to the pituitary of up to 15 Gy and the maximum dose to the distal infundibulum of up 17 Gy for 5 years. Future long-term studies will be necessary to confirm that they remain maximum safe doses to avoid hypopituitarism. The absence of investigation of GH axis after LGK irradiation is another limitation of our study. In conclusion, to avoid hypopituitarism, the mean radiation dose to the pituitary tissue surrounding the adenoma should be!15 Gy, and the maximum dose to distal infundibulum should be!17 Gy. These limits should become a rule when irradiating pituitary adenomas, similar to the dose rules of 8 Gy for the optical tract or Gy for brain stem.

9 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 Declaration of interest The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported. Funding This work was supported by an institutional grant from the Ministry of Education of the Czech Republic (MSM ). References 1 Hardy J. Transsphenoidal microsurgery of prolactinomas. In Secretory Tumors of the Pituitary Gland, pp Eds PM Black, NT Zervas, EC Ridgway & JB Martin, New York: Raven Press, Fahlbusch R & Buchfelder M. Present status of neurosurgery in the treatment of prolactinomas. Neurosurgical Review (doi: /bf ) 3 Freda PU, Sharon L, Wardlaw SL & Post KD. Long-term endocrinological follow-up evaluation in 115 patients who underwent transsphenoidal surgery for acromegaly. Journal of Neurosurgery (doi: /jns ) 4 Swearingen B, Biller BMK, Barker FG, Katznelson L, Grinspoon S, Klibanski A & Zervas NT. Long-term mortality after transsphenoidal surgery for Cushing s disease. Annals of Internal Medicine Kreutzer J, Vance ML, Lopes MB & Laws ER Jr. Surgical management of GH- secreting pituitary adenomas: an outcome study using modern remission criteria. Journal of Clinical Endocrinology and Metabolism (doi: /jc ) 6 Jane JA Jr & Laws ER. The surgical management of pituitary adenomas in a series of 3,093 patients. Journal of the American College of Surgeons (doi: /s (01) ) 7 Barrande G, Pittino-Lungo M, Coste J, Ponvert D, Bertagna X, Luton JP & Bertherat J. Hormonal and metabolic effects of radiotherapy in acromegaly: long-term results in 128 patients followed in a single center. Journal of Clinical Endocrinology and Metabolism (doi: /jc ) 8 Biermasz NR, van Dulken H & Roelfsema F. Postoperative radiotherapy in acromegaly is effective in reducing GH concentration to safe levels. Clinical Endocrinology (doi: /j x) 9 Minniti G, Jaffrain-Rea M, Osti M, Esposito V, Santoro A, Solda F, Gargiulo P, Tamburrano G & Enrici RM. The long-term efficacy of conventional radiotherapy in patients with GH-secreting pituitary adenomas. Clinical Endocrinology (doi: /j x) 10 Flickinger JC, Nelson PB, Taylor FH & Robinson A. Incidence of cerebral infarction after radiotherapy for pituitary adenoma. Cancer (doi: / ( )63:12!2404::AID-CNCR O3.0.CO;2-3) 11 Fukamachi A, Wakao T & Akai J. Brain stem necrosis after irradiation of pituitary adenoma. Surgical Neurology (doi: / (82)90146-x) 12 Bowen J & Paulsen CA. Stroke after pituitary irradiation. Stroke Brada M, Rajan B, Traish D, Ashley S, Holmes-Sellors J & Uttley D. The long-term efficacy of conservative surgery and radiotherapy in the control of pituitary adenomas. Clinical Endocrinology (doi: /j tb02137.x) 14 Fisher BJ, Gaspar LE & Noone B. Radiation therapy of pituitary adenoma: delayed sequelae. Radiology Hypopituitarism after LGK irradiation Liščák R, Vladyka V, Marek J, Šimonová G & Vymazal J. Gamma knife radiosurgery for endocrine-inactive pituitary adenomas. Acta Neurochirurgica (doi: /s ) 16 Sheehan JP, Niranjan A, Sheehan JM, Jane JA, Laws ER, Kondziolka D, Flickinger J, Landolt AM, Loeffler JS & Lunsford LD. Stereotactic radiosurgery for pituitary adenomas: an intermediate review of its safety, efficacy, and role in the neurosurgical treatment armamentarium. Journal of Neurosurgery (doi: /jns ) 17 Pollock BE, Nippoldt TB, Stafford SL, Foote RL & Abboud ChF. Results of stereotactic radiosurgery in patients with hormoneproducing pituitary adenomas: factors associated with endocrine normalization. Journal of Neurosurgery (doi: /jns ) 18 Sheehan JP, Kondziolka D, Flickinger J & Lunsford LD. Radiosurgery for residual or recurrent nonfunctioning pituitary adenoma. Journal of Neurosurgery (Supplement 5) (doi: /jns supplement5.0408) 19 Wowra B & Stummer W. Efficacy of gamma knife radiosurgery for nonfunctioning pituitary adenomas: a quantitative follow up with magnetic resonance imaging-based volumetric analysis. Journal of Neurosurgery (Supplement 5) (doi: / jns supplement5.0429) 20 Feigl GCh, Bonelli ChM, Berghold A & Mokry M. Effects of gamma knife radiosurgery of pituitary adenomas on pituitary function. Journal of Neurosurgery (Supplement 5) Ježková J, Marek J, Hána V, Kršek M, Weiss V, Vladyka V, LiščákR, Vymazal J & Pecen L. Gamma knife radiosurgery for acromegaly long-term experience. Clinical Endocrinology (doi: /j x) 22 Vladyka V, Liščák R, Novotný J Jr, Marek J & Ježková J. Radiation tolerance of functioning pituitary tissue in gamma knife surgery for pituitary adenomas. Neurosurgery (doi: /01.NEU ) 23 Novotný J Jr, Novotný J, Vymazal J, Liščák R & Vladyka V. Assessment of the accuracy of stereotactic target localization using magnetic resonance imaging: phantom study. Journal of Radiosurgery (doi: /b:jora ) 24 Abs R, Bengtsson BA, Hernberg-Stahl E, Monson JP, Tauber JP, Wilton P & Wüster Ch. GH replacement in 1034 growth hormone deficient hypopituitary adults: demographic and clinical characteristics, dosing and safety. Clinical Endocrinology (doi: /j x) 25 Littley MD, Shalet SM & Beardwell CG. Hypopituitarism following external beam radiotherapy for pituitary tumours in adults. Quarterly Journal of Medicine Schalin-Jäntti C, Valanne L, Tenhunen M, Setälä K, Paetau A, Sane T & Kouri M. Outcome of fractionated stereotactic radiotherapy in patients with pituitary adenomas resistant to conventional treatment: a 5.25-year follow-up study. Clinical Endocrinology Roug S, Rasmussen AK, Juhler M, Kosteljanetz M, Poulsgaard L, Heeboll H, Roed H & Feldt-Rasmussen U. Fractionated stereotactic radiotherapy in patients with acromegaly: an interim single-centre audit. European Journal of Endocrinology (doi: /eje ) 28 Ikeda H, Jokura H & Yoshimoto T. Transsphenoidal surgery and adjuvant gamma knife treatment for growth hormone-secreting pituitary adenoma. Journal of Neurosurgery (doi: /jns ) 29 Morange-Ramos I, Regis J, Dufour H, Andrieu JM, Grisoli F, Jaquet P & Peragut JC. Short-term endocrinological results after gamma knife surgery of pituitary adenoma. Stereotactic and Functional Neurosurgery (Supplement 1) (doi: / ) 30 Landolt AM, Haller D, Lomax N, Scheib S, Schubiger O, Siegfried J & Wellis G. Stereotactic radiosurgery for recurrent surgically

10 178 J Marek and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2011) 164 treated acromegaly: comparison with fractionated radiotherapy. Journal of Neurosurgery (doi: /jns ) 31 Lim YL, Leem W, Kim TS, Rhee BA & Kim GK. Four years experience in the treatment of pituitary adenomas with gamma knife radiosurgery. Stereotactic and Functional Neurosurgery (Supplement 1) (doi: / ) 32 Kim SH, Huh R, Chang JW, Park YG & Chung SS. Gamma knife radiosurgery for functioning pituitary adenomas. Stereotactic and Functional Neurosurgery (Supplement 1) (doi: / ) 33 Mokry M, Ramschak-Schwarzer S, Simbrunner J, Ganz JC & Pendl G. A six year experience with the postoperative radiosurgical management of pituitary adenomas. Stereotactic and Functional Neurosurgery (Supplement 1) (doi: / ) 34 Sheehan JM, Vance ML, Sheehan JP, Ellegala DB & Laws ER Jr. Radiosurgery for Cushing s disease after failed transsphenoidal surgery. Journal of Neurosurgery (doi: / jns ) 35 Pollock BE & Carpenter PC. Stereotactic radiosurgery as an alternative to fractionated radiotherapy for patients with recurrent or residual nonfunctioning pituitary adenomas. Neurosurgery (doi: /01.neu ) 36 Attanasio R, Epaminonda P, Motti E, Giugni E, Ventrella L, Cozzi R, Farabola M, Loli P, Beck-Peccoz P & Arosio M. Gamma-knife radiosurgery in acromegaly: a 4-year follow-up study. Journal of Clinical Endocrinology and Metabolism (doi: /jc ) 37 Petrovich Z, Yu C, Gianotta SL, Zee CS & Apuzzo ML. Gamma knife radiosurgery for pituitary adenoma: early results. Neurosurgery (doi: /01.neu ) 38 Losa M, Valle M, Mortini P, Franzin A, da Passano CF, Cenzato M, Bianchi S, Picozzi P & Giovanelli M. Gamma knife surgery for treatment of residual nonfunctioning pituitary adenomas after surgical debulking. Journal of Neurosurgery (doi: /jns ) 39 Castinetti F, Taieb D, Kuhn JM, Chanson P, Tamura M, Jaquet P, Conte-Devolx B, Régis J, Dufour H & Brue T. Outcome of gamma knife radiosurgery in 82 patients with acromegaly: correlation with initial hypersecretion. Journal of Clinical Endocrinology and Metabolism (doi: /jc ) 40 Landolt AM, Lomax N, Scheib SG & Girard J. Gamma knife surgery after fractionated radiotherapy for acromegaly. Journal of Neurosurgery (doi: /sup ) 41 Mingione V, Yen ChP, Vance ML, Steiner M, Sheehan J, Laws E & Steiner L. Gamma knife surgery in the treatment of nonsecretary pituitary macroadenoma. Journal of Neurosurgery (doi: /jns ) 42 Pouratian N, Sheehan J, Jagannathan J, Laws ER, Steiner L & Vance ML. Gamma knife radiosurgery for medically and surgically refractory prolactinomas. Neurosurgery (doi: /01.neu bd) 43 Pollock BE, Jacob JT, Brown PD & Nippoldt TB. Radiosurgery for growth hormone-producing pituitary adenomas: factors associated with biochemical remission. Journal of Neurosurgery (doi: /jns ) 44 Jagannathan J, Sheehan JP, Pouratian N, Laws ER, Steiner L & Vance ML. Gamma knife surgery for Cushing s disease. Journal of Neurosurgery (doi: /jns ) 45 Pollock BE, Cochran J, Natt N, Brown PD, Erickson D, Link MJ, Garces YI, Foote LR, Stafford SL & Schomberg PJ. Gamma knife radiosurgery for patients with nonfunctioning pituitary adenomas: results from a 15-year experience. International Journal of Radiation Oncology, Biology, Physics (doi: /j.ijrobp ) 46 Castinetti F, Nagai M, Dfour H, Kuhn JM, Morange I, Jaquet P, Conte-Devolx B, Regis J & Brue T. Gamma knife radiosurgery is a successful adjunctive treatment in Cushing s disease. European Journal of Endocrinology (doi: /eje ) 47 Ježková J, Hána V, Kršek M, Weiss V, Vladyka V, Liščák R, Vymazal J, Pecen L & Marek J. Use of the Leksell gamma knife in the treatment of prolactinoma patients. Clinical Endocrinology (doi: /j x) Received 13 October 2010 Accepted 11 November 2010