Nonsecreting pituitary adenomas (NSA) become symptomatic by compression of surrounding structures causing visual or endocrinological dysfunction. Especially in cases of invasion of the cavernous sinus, complete microsurgical tumor removal might be associated with elevated surgery-related morbidity and, as a consequence, in many cases parasellar tumor remains [34]. For the management of residual tumor, therapeutic options consist of observation, external beam irradiation, fractionated stereotactic radiotherapy, and radiosurgery [1, 2, 3, 4, 5, 6, 8, 11, 14, 15, 18, 19, 21, 22, 23, 24, 25, 30, 31, 32, 33, 36, 37, 38, 39, 46, 47, 48, 51, 52]. The aim of this retrospective study was to investigate the value of linear accelerator-based radiosurgery (LINAC-RS) for the treatment of NSA, with regard to its long-term tumor control and safety.

Patients and methods

Patient selection and follow-up

Between 1992 and August 2008, 65 patients suffering from progressive NSA were treated with LINAC-RS. Patient treatment and follow-up were conducted according to a prospective protocol [48, 50]. Patients were included suffered from NSA, tumor recurrence, or progressive residual tumor, definable tumor, maximum dimension of 35 mm and a minimal distance of 1–2 mm to the anterior optic system.

Patients were followed with regard to their ophthalmological, endocrinological, and radiological (CT/MRI) course. The first complete follow-up examination was performed 6 months after LINAC-RS. Afterwards radiological follow-up was performed at 12-month intervals, whereas ophthalmological and endocrinological follow-up examinations were performed every 6 months.

Criteria for treatment-related ophthalmological side effects were impairment of vision or visual field. Every intake of new hormonal medication related to one of these axes as well as pathological hormone deficits were interpreted as radiation-induced hypothalamopituitary dysfunction.

Tumor size was determined by measuring the maximum tumor diameter in at least two dimensions. According to the greatest tumor dimension, we defined a ‘‘response’’ as a reduction of ≥ 25%, ‘‘progression’’ as an increase ≥ 25% compared to baseline diameter, and “stable disease” for all other cases. Complete response was documented if MRI studies no longer displayed a signal specific for tumor [1, 28].

For the assessment of radiation-induced hypothalamopituitary dysfunction, the calculation of radiation doses was performed using dose–volume histograms in cases with a definable pituitary gland and stalk and at least partial function of the anterior pituitary lobe before LINAC-RS. Peak doses in the tumor and the median dose were evaluated.

LINAC-RS

The detailed components of our treatment system are described elsewhere [26, 27, 41, 48, 50]. Before 1996, the tumor contour was outlined on stereotactic CT images. Since 1996, all patients received MRI 1–3 days prior to LINAC-RS, which was integrated into intraoperative stereotactic CT [13, 42]. Treatment planning was performed using specialized software (STP 3.3 and 3.5, Howmedica Leibinger, Freiburg, Germany).

On the day of treatment, the patient’s head was fixated in a modified Riechert–Mundinger stereotactic frame [40]. To obtain contrast enhancement of the tumor and to visualize blood vessels for landmark correlation, 100 ml and 40–80 ml of contrast medium were applied approximately 30 min prior to and directly before CT scanning, respectively.

The dose applied to the anterior visual pathways was limited to 9 Gy and the volume of healthy brain tissue exposed to a minimum dose of 10 Gy has been restricted to 10 ml according to an in-house risk analysis introduced in 1996 [49].

In 45 of 61 (73.7%) patients, dose application was performed with circular collimators fitted to an adapted linear accelerator (Philips SL 75/20 at 9 MV or Elekta Sli25 at 6 MV) using STP 3.5 (until February 1996 STP2, Leibinger, Freiburg, Germany) [45]. An arching beam technique was used with 10 table positions. For individual treatment planning, this standard technique can be modified in terms of diameter, table position, number of table angles, ranges of gantry rotation, beam weight, irradiation dose, and number of isocenters [44].

In 16 of 61 (26.2%) patients, a micro-multileaf collimator (MMLC) with 1.5 mm leaf width (Siemens, Heidelberg, Germany) was used. Because the MMLC technique is able to treat complex-shaped tumors, this technique has been used predominantly since in 2001. A focal irradiation technique with 14–20 static fields was applied. Each beam was shaped according to the beam’s eye view of the tumor and structures at risk. Treatment planning was performed using Virtuoso 3.0.3 software (Howmedica Leibinger, Freiburg, Germany). A representative treatment plan is illustrated in Fig. 1.

Fig. 1
figure 1

Dose distribution for LINAC-RS using MMLC technique. Yellow chiasm and optic nerve, blue pituitary gland and stalk, red tumor, green line 50% isodose consistent with 8 Gy in this case

Full size image

Statistical analysis

For statistical analysis SPSS 17.0 for Windows (SPSS Inc., Chicago, IL, USA) was used. Actuarial local tumor control rate, rate of neurological deficits and new deficits in hypothalamopituitary function as well as actuarial survival were determined using the Kaplan–Meier method [20].

The log-rank test was performed to compare patient groups comprising censored patients in order to determine prognostic factors for endocrinological function. For evaluation of the course of endocrinological function, patients with complete insufficiency of the anterior lobe and with diabetes insipidus were excluded. Patients were compared concerning pre-existing conditions (age, sex, microsurgical operations, tumor volume, tumor location, hypothalamopituitary deficits) and treatment-associated factors (dose, radiation application technique, dose to pituitary gland and stalk). A p value ≤ 0.05 was considered statistically significant.

Results

Four patients were lost to follow-up. Sixty-one individuals with a minimum follow-up of 12 months were included in this retrospective analysis. For details of patients’ characteristics see Tab. 1.

Tab. 1 Patients’ characteristics and treatments before LINAC-RS
Full size table

The median therapeutic dose applied to the tumor margin was 13.0 Gy, minimum 10 Gy, and maximum 20 Gy. The dosage characteristics referring to the two collimation techniques used are displayed in Tab. 2.

Tab. 2 Dosage characteristics for therapeutic dose (TD) and maximum dose (MD) of the tumor
Full size table

Follow-up

Mean follow-up after LINAC-RS was 90 months (median 83, range 15–186 months); 36 months after LINAC-RS 1 patient (female, 60 years) died of reasons unrelated to pituitary disease and treatment. None of the patients developed radiation induced tumors during the follow-up period.

Tumor

At the end of follow-up, 60 patients (98.3%) showed local tumor control on MRI, including 24 patients (40.0%) with partial tumor remission (Fig. 2). Tumor progression was observed in 1 male patient treated on the basis of CT in 1995 suffering from an invasively growing adenoma with intra-, para-, supra-, and retrosellar extension. After LINAC-RS, none of the patients experienced tumor bleeding, necrosis, ischemia or pituitary apoplexy.

Fig. 2
figure 2

Coronary T1-weighted MRI with contrast enhancement prior to (left) and 54 months after (right) LINAC-RS (radiation dose at tumor surface 12 Gy, maximum dose 16 Gy)

Full size image

Hypothalamopituitary function

Four (9.8%) of 41 patients at risk experienced new endocrinological deficits after a mean of 44.5 months (median 54 months, range 10–60 months). The cumulative risk for radiation-induced hypothalamopituitary dysfunction was 12.1% after 60 and 120 months (Fig. 3).

Fig. 3
figure 3

Cumulative risk for radiation-induced damage of pituitary function according to Kaplan–Meier method. Ticks represent censored patients (n = 41, 90.2% censored, 4 events, cumulative probability 0.02 after 36 months, 0.12 after 142 months)

Full size image

Except for female gender (p = 0.043) and maximum dose to the pituitary gland less than 16 Gy (p = 0.035), there were no favorable prognostic factors for preservation of endocrinological function. Extrasellar tumor extension showed a tendency towards a positive prognostic factor (p = 0.064) for preservation of endocrinological function as well as no patient without pituitary deficiency prior to LINAC-RS suffered radiation-induced hormonal deficit (p = 0.107; Tab. 3).

Tab. 3 Prognostic factors for endocrinological function for NSA after LINAC-RS
Full size table

Neurology

Cranial nerves

None of the patients developed a new deficit of visual field or vision after application of a maximum of punctually 9 Gy to the anterior optic system.

Seizures

In one early case (1.7%) of those patients included in this study in 1995, a seizure disorder occurred 11 months after treatment. The therapeutic isodose was 18 Gy with a maximum dose of 57 Gy (maximum peak dose in study) leading to radiation-induced temporary contrast enhancement with edema in the right temporal lobe. Under anticonvulsant medication, seizures ceased and further follow-up revealed resolution of the radiation-induced changes. Since the limitation of the volume of peritumoral brain tissue exposed to 10 Gy to 10 ml, no seizure disorder has occurred [49].

Overall outcome

After a mean follow-up time of 90 months (median 83 months, range 15–186 months) treatment success with local tumor control and absent radiation-induced side effects could be demonstrated in 90.2% of cases. Six patients (9.8%) suffered aversive events (4 hypothalamopituitary dysfunction, 1 tumor progression, 1 seizure disorder). The Kaplan–Meier method revealed a cumulative probability for absence of aversive events of 87% after 141 months (Fig. 4).

Fig. 4
figure 4

Event-free survival after LINAC-RS for NSA. Cumulative risk for unfavorable events calculated with the Kaplan–Meier method. Ticks represent censored patients (n = 61, 90.2% censored, 6 events, cumulative probability 0.04 after 36 months and 0.13 after 161 months)

Full size image

Discussion

In the present study with the longest overall follow-up of a radiosurgery series, local tumor control was achieved in 98% of patients comparable to other radiosurgery series using relatively low therapeutic and maximum doses (Tab. 4). The only patient experiencing tumor progression after LINAC-RS had been treated in the pre-MRI era with a reduced dose (therapeutic dose 10 Gy) due to a large atypical adenoma [36, 38]. With today’s microsurgical options and imaging, a comparable patient would surely not be treated radiosurgically.

Tab. 4 Results of GKRS, LINAC-RS, CRS, stereotactically guided fractionated radiotherapy, and conventional radiotherapy in series with follow-up of at least 48 months and comparable studies
Full size table

Remarkably, in this series there was no correlation between tumor size and local control, which is in contrast to the findings of other centers [36, 38]. This is not easy to comprehend as those centers used higher therapeutic doses. We presume that there are additional factors affecting biological impact that currently cannot be quantified. In addition, every patient receiving a therapeutic dose of 12 Gy exhibited local tumor control in our series.

It is worth mentioning that all NSA treated with the MMLC technique using low therapeutic and maximum doses (Fig. 2) showed local tumor control at the end of follow-up. The MMLC technique provides conformal dose distribution and seems to be a useful tool in lesions with an irregular configuration like NSA. The use of the MMLC since 2001 is an additional step towards lower peak doses and more conformal and homogenous irradiation, which is desirable in benign lesions like pituitary adenomas.

In one series, 11 out of 41 patients (27%) developed new endocrinological deficits [38]. The actuarial risk after 5 years was 32% compared to 12% in this study. At least partially, this may be a consequence of the higher applied therapeutic and maximum doses of 16.3 Gy and 34.0 Gy (compared to 13.1 Gy and 28.8 Gy). On the other hand, another evaluation [25] kept the rate of new endocrinological deficits as low as 2% after applying a relatively high therapeutic mean dose of 20 Gy. The reasons for such inconsistent results will hopefully be identified on the basis of future evaluations.

A maximum of 16 Gy applied to the pituitary gland was a significant prognostic factor for preservation of endocrinological function in this series. This is reasonably consistent with the results of three GKRS series [15, 23, 47]. Vladyka et al. [47] found a mean dose of more than 17 Gy and 20 Gy to be a risk factor for impairment of the gonadotrophic and corticotrophic/thyrotrophic function, respectively. Because the critical doses are approximately in the same range, they strongly support efforts for dose reduction in recognizable pituitary tissue [23].

In this study, extrasellar tumor location had a tendency to be a favorable prognostic factor for preservation of pituitary function (Tab. 3). None of the patients with extrasellar tumor developed a new endocrinological deficit. Generally, one would expect that tumors with no intrasellar portion could be treated radiosurgically with a lower risk for endocrinological function because of radiation sparing and easier recognition of the pituitary gland and stalk, which the present evaluation could not confirm. Data remains inconclusive in this regard at the moment.

Generally, after macroscopic complete tumor removal additional treatment is required in up to 6% of NSA patients and in even 20–26% after incomplete removal with or without conventional radiotherapy [12, 16, 29]. In the light of several radiation-based treatment options and individual tumor location and size, tailored adjuvant treatment strategies are necessary.

In a prospective study [8], fractionated stereotactic radiation therapy led to hypopituitarism in 28.5% of 110 patients after 4 years. Another series [30] revealed impairment of visual acuity in 4 out of 60 patients (6.7%) and hypopituitarism in 3 of 63 patients (4.6%). After 5 years, tumor control was achieved in 93% [30]. Although the results for NSA are not reported separately, LINAC-RS seems to yield more desirable results especially concerning the visual system. Treatment results after radiosurgery, CyberKnife radiosurgery, stereotactically guided radiotherapy, and conventional fractionated radiotherapy of NSA are presented in Tab. 4. The first reports of CyberKnife treatment of pituitary adenomas are available [7, 19, 22]. Despite the few cases with short follow-up, it can be recognized that optimal radiation dose and application mode have to be elaborated.

Because of the higher rate of local tumor control and lower rate of hypopituitarism compared to conventional radiotherapy, LINAC-RS should be preferred in selectable tumors, although it has to be kept in mind that for the review of results of conventional radiation therapy, data of the last few decades were included [3] with many cases suffering tumors not selectable for radiosurgery. Radiation-induced hypopituitarism occurred in of 13–56% after conventional radiotherapy [3].

Conclusion

Our series regarding radiosurgical treatment for NSA is the one with the longest overall follow-up studies presented so far. We could demonstrate that LINAC-RS is a minimally invasive, safe, and effective treatment option for recurrent NSA or microsurgically inaccessible residual tumor. The present study shows that extrasellar growth, which in case of parasellar tumor is a serious problem for operation planning, is even a favorable growth pattern for safe LINAC-RS and preservation of pituitary function.

In cases where 12 Gy can be applied to the tumor surface with maximally punctual application of 9 Gy to the optic system and a dose to the pituitary gland not exceeding 16 Gy LINAC-RS is superior to external beam irradiation and fractionated stereotactic radiation therapy and therefore should be preferred.

The main focuses for the future should certainly be avoidance of visual complications and further reduction of hypopituitarism.