Fax: (011) 49-6221-56-5353
Achievement of long-term local control in patients with craniopharyngiomas using high precision stereotactic radiotherapy
Article first published online: 27 APR 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 11, pages 2308–2314, 1 June 2007
How to Cite
Combs, S. E., Thilmann, C., Huber, P. E., Hoess, A., Debus, J. and Schulz-Ertner, D. (2007), Achievement of long-term local control in patients with craniopharyngiomas using high precision stereotactic radiotherapy. Cancer, 109: 2308–2314. doi: 10.1002/cncr.22703
- Issue published online: 18 MAY 2007
- Article first published online: 27 APR 2007
- Manuscript Accepted: 14 FEB 2007
- Manuscript Revised: 5 FEB 2007
- Manuscript Received: 17 NOV 2006
- Rathke pouch;
- fractionated stereotactic radiotherapy;
- local control;
The long-term outcome in patients with craniopharyngiomas treated with fractionated stereotactic radiotherapy (FSRT) was evaluated.
A total of 40 patients with craniopharyngiomas were treated between May 1989 and July 2006 with FSRT. Most patients were treated for tumor progression after surgery. A median target dose of 52.2 grays (Gy) (range, 50.4–56 Gy) was applied in a median conventional fractionation of 5 × 1.8 Gy per week. Follow-up examinations included thorough clinical assessment as well as contrast-enhanced magnetic resonance imaging scans.
After a median follow-up of 98 months (range, 3–326 months), local control was 100% at both 5 years and 10 years. Overall survival rates at 5 years and 10 years were 97% and 89%, respectively. A complete response was observed in 4 patients and partial responses were noted in 25 patients. Eleven patients presented with stable disease during follow-up. Acute toxicity was mild in all patients. Long-term toxicity included enlargement of cysts requiring drainage 3 months after FSRT. No visual impairment, radionecrosis, or development of secondary malignancies were observed.
The long-term outcome of FSRT for craniopharyngiomas is excellent with regard to local control as well as treatment-related side effects. Cancer 2007. © 2007 American Cancer Society.
Craniopharyngiomas are rare benign tumors that are known to arise from remnants of the Rathke pouch. They are epithelial neoplasms that are usually located in the sellar and suprasellar regions. Craniopharyngiomas account for 1% to 2% of all intracranial neoplasms, but they are commonly diagnosed in children and adolescents.1 In general, the overall incidence, including patients of all ages, is 0.13 per 10,000 neoplasms per year; in the U.S. about 338 new cases are reported to be observed annually.2, 3 Craniopharyngiomas can compress surrounding structures and therefore lead to significant clinical impairment, including visual deficits, hydrocephalus, and endocrine abnormalities.4, 5
Both surgery and radiotherapy (RT) have been implemented effectively in the treatment of craniopharyngiomas,6, 7 with, however, distinct risk profiles for treatment-related side effects.8 Neurosurgical resection is an effective treatment option when a total resection can be achieved without major side effects.9 Partial resection can be implemented to alleviate clinical symptoms such as increased intracranial pressure or visual deficits due to compression of the optic system; moreover, limited surgery can help to reduce tumor volume or delay the need for RT. In the past, surgery was often chosen as the primary treatment, although RT alone is known to provide long-term disease control7, 8, 10, 11; the goal was to prevent deficits in endocrine or cognitive function. However, recent studies have shown that such deficits are present at the time of diagnosis or after surgery, and cannot be attributed fully to RT.12, 13
With modern techniques of RT such as fractionated stereotactic radiotherapy (FSRT) or stereotactic radiosurgery (SRS), it is possible to deliver the high local doses required for tumor control while sparing surrounding organs at risk, such as the optic chiasm and the brainstem. Therefore, the risk for treatment-related side effects can be minimized. This is extremely important in patients with an excellent outcome after RT with a normal life expectancy. Previously, we reported that FSRT is safe and effective in patients with craniopharyngiomas.14 In the current study, we retrospectively evaluated the long-term outcome of patients with craniopharyngiomas treated with FSRT at the University Hospital of Heidelberg and update our treatment results with special regard to treatment-related side effects as well as treatment outcome.
MATERIALS AND METHODS
Between May 1989 and July 2006, 40 patients with craniopharyngiomas underwent RT using FSRT. The median age of the patients was 41 years (range, 6–70 years). Six children aged <18 years were treated with FSRT for craniopharyngiomas and are included in this analysis. The male:female ratio was 2:3.
All patients had undergone at least 1 surgical intervention allowing for pathologic confirmation of a craniopharyngioma in all patients. Most patients were treated with FSRT for tumor progression rather than electively after surgery. FSRT was applied adjuvantly after surgery in 12 of 40 patients (30%), and in 70% of all patients (n = 28) FSRT was performed for the management of tumor progression or recurrence and/or progressive growth of the cysts after surgery. The neurosurgical resection was complete in 7 patients after primary diagnosis, and FSRT was performed at a later timepoint for tumor progression. A subtotal resection was performed in 23 patients and a biopsy with drainage of the tumor cyst was conducted in 10 of 40 patients (25%). Eight patients presented with reservoirs before radiotherapy and 3 patients presented with ventriculoperitoneal shunts.
The main symptoms observed at the initiation of FSRT was loss of visual acuity and visual field defects, as well as endocrine dysfunction, often associated with major obesity; however, 10 patients (25%) were free of tumor-associated symptoms.
For treatment planning, patients were immobilized using an individually manufactured precision head mask made of Scotch cast, which was attached to a stereotactic base frame for computed tomography (CT) and magnetic resonance imaging (MRI) with a slice thickness of 3 mm, as described previously, with an overall geometric accuracy of 1 to 2 mm.14 Before 1993, CT-based target volume definition was performed; thereafter, all patients received CT and MRI in mask fixation for treatment planning. The target volume was defined, after stereotactically guided image fusion, on each slice of the three-dimensional data cube. The contrast-enhancing solid lesion depicted on T1-weighted MRI scans was defined as the macroscopic (gross) target volume (GTV), as well as all cystic components of the tumor and the wall of the cyst, visible in T2-weighted MRI scans. All patients presented with macroscopically visible tumor at the time of initiation of FSRT.
We considered the clinical target volume (CTV) to be identical to the GTV for craniopharyngiomas; the planning target volume (PTV) consisted of the GTV, adding a safety margin of 2 mm to account for possible patient misalignment. In patients treated before 1993, CT-based target volume definition was conducted applying an identical target volume concept. All patients had macroscopic tumor at the time of FSRT. The median PTV treated was 20.7 mL (range, 5.2–139 mL).
Treatment planning was performed using the 3-dimensional treatment planning system Voxelplan (dkfz, Heidelberg; also available as Virtuoso, Leibinger, Germany, or STP); the beams' eye view was employed for field optimization.
A median dose of 52.2 grays (Gy) (range, 50.4–56 Gy) was delivered in median single fractions of 1.8 Gy (range, 1.8–2 Gy), 5 times per week, using 3–5 noncoplanar isocentric fields, irregularly shaped by a midsize multileaf collimator with a leaf thickness of 5 mm at isocenter. The target doses were prescribed to the isocenter, the 90% isodose encompassed the target volume. With respect to the localization of craniopharyngiomas, the eyes, optic nerves, chiasm, and brainstem were of special concern during treatment planning. All patients were treated with a linear accelerator dedicated to stereotactic RT (Siemens, Erlangen, Germany) with 6 megaelectron volts (MeV) or 15 MeV. A typical treatment plan and dose distribution is depicted in Figure 1.
MRI or CT imaging was performed once during the treatment in an asymptomatic patient; in cases in which new symptoms occurred, imaging was scheduled more frequently as needed.
All patients were seen 6 weeks after the completion of RT; thereafter, follow-up visits were scheduled in 3-month intervals during the first year after RT, then at 6-month intervals. After 5 years, follow-up visits were set up yearly. All follow-up examinations included complete physical examination and MRI scans of the brain, as well as endocrinologic, ophthalmologic, and neurologic examinations. Data collection was completed by contacting the patients themselves, their home physicians, or the attending medial specialists caring for the patients.
For radiographic outcome, we defined a complete response (CR) as the absence of any tumor residuals on MRI scans; a partial response (PR) was defined as a >50% tumor reduction; and an objective response (OR) was classified as a response of 25% to 50% of the tumor. Less than 25% tumor reduction without any signs of tumor progression was determined to be stable disease (SD). For response analysis, the solid area as well the cystic components were included.
Functional outcome after FSRT was determined according to the improvement of impairment of visual acuity and endocrine function. We assessed all acute and long-term treatment-related side effects according to the Common Toxicity Criteria for Adverse Events (CTCAE) (version 3.0) (available at: http://www.fda.gov/cder/cancer/toxicityframe.htm; accessed February 5, 2007) and the Radiation Therapy Oncology Group/European Organization for Research and Treatment of Cancer (RTOG/EORTC) criteria, respectively.
Actuarial local control as well as overall survival rates were calculated from the onset of RT using the Kaplan-Meier method.15 In the Kaplan-Meier curve, patients alive at the time of the analysis were censored and marked as ticks on the curve. Deaths are depicted as circles in the curve.
In 39 patients, treatment could be completed without interruptions for ≥4 days. In 1 patient, headache, nausea/vomiting, anisocoria, and fatigue developed after 2 fractions due to a hydrocephalus that developed after a massive cyst enlargement. Therefore, the patient received cyst drainage with a Rickham reservoir, which was replaced after a few days with a ventriculoperitoneal shunt. Thereafter, RT was used after new CT and MRI scans were acquired postoperatively and a new treatment plan was calculated; treatment could then be completed without further complications.
The median follow-up time was 98 months (range, 2–326 months).
One 51-year-old female patient developed shunt insufficiency during follow-up, requiring neurosurgical intervention 49 months after FSRT. At this time, no signs of tumor progression could be observed on MRI scans. Thereafter, the patient was in such bad general condition that no further MRI evaluations were performed. Therefore, this patient was censored after 49 months and she died 88 months after RT. A second patient, a 29-year-old female, also developed shunt insufficiency 3 months after RT; during a neurosurgical resection the shunt as well as the Rickham reservoir were replaced and the patient died of perioperative complications. However, the craniopharyngioma did not progress after FSRT. A 25-year-old female patient treated with FSRT for progressive craniopharyngioma developed nausea, vomiting, and focal seizures due to shunt insufficiency 21 months after RT; at the time of last follow-up, after replacement of the ventriculoperitoneal shunt, the patient was doing well with stable disease after undergoing FSRT.
Twelve patients were treated with FSRT after subtotal resection, cyst aspiration, or biopsy of the tumor. In 28 patients, FSRT was performed for tumor progression, with a median time interval between the primary diagnosis and FSRT of 21 months (range, 1–256 months). Within the group of 6 children treated, the median time from surgery to FSRT was 38 months (range, 7–106 months).
The median actuarial progression-free survival calculated from the initiation of FSRT was 100% after 10 years. Of these patients, 4 patients (10%) achieved CR, a PR was observed in 25 patients (62.5%), and in 11 patients (27.5%) an objective response or stable disease was documented during follow-up.
The overall actuarial survival rates were 97% and 89%, respectively, after 5 years and 10 years (Fig. 2); 37 of 40 patients were alive at the time of last follow-up.
One 14-year-old child died of organ dysfunction 106 months after FSRT. In this patient, organ damage developed after a complete hypoparathyroidism developed, before any therapeutic intervention, at the age of 2 years. After FSRT, the patient's craniopharyngioma showed CR without any signs of tumor progression. Another patient developed shunt insufficiency 3 months after FSRT that required shunt replacement. During surgery the patient suffered from an intracranial hemorrhage and died during the postoperative course in intensive care. The third patient died at the age of 59 years, 7 years after FSRT, without tumor progression.
With regard to direct adjuvant FSRT as opposed to FSRT for tumor progression, no differences in local control as well as overall survival could be observed.
Visual acuity improved after FSRT in 5 patients; no patient showed new impairment of vision during follow-up after FSRT.
In all patients, endocrinologic examinations to evaluate hormonal function were performed on a regular basis before RT and during follow-up to determine endocrinopathies. Obesity was defined as a significant subjective weight gain since primary diagnosis.
At the time of presentation for FSRT, 8 patients presented with intact hormone function. In 4 patients, elevated levels of prolactin could be observed, of which 1 patient's prolactin level normalized after FSRT. Nine of 40 patients (22.5%) presented with panhypopituitarism that developed postoperatively.
In all, 50% of patients developed partial hypopituitarism after surgical resection of the tumor. After FSRT, only 1 patient with partial hypopituitarism developed panhypopituitarism, and 1 patient developed new deficiency of 1 hormone. However, before surgical resection of the craniopharyngioma, no hormonal deficiencies were present.
Only 1 patient suffered from seizures before and after FSRT; obesity was observed in 4 patients, and 5 suffered from cranial nerve deficits before FSRT.
Formal neuropsychological assessment was not performed on a regular basis in this series. Thirty-eight patients did not demonstrate any decline in neuropsychologic status during follow-up. In 2 patients, major cognitive dysfunction developed, as described previously.14 Six children aged <18 years were treated in this analysis. One 8-year-old presented with fatigue, panhypopituitarism, and major visual deficits before RT, after 9 surgical procedures for recurrent craniopharyngioma were performed. During follow-up, he was doing well and attending special education classes. However, 8.8 years after FSRT, he died in multiorgan failure due to hormonal deficits.
We did not observe any severe side effects such as radionecrosis or secondary malignancies.
The treatment of craniopharyngiomas is still discussed controversially. For the majority of physicians, the principal curative approach is still considered to be surgery; however, this is not always possible with moderate, acceptable toxicity.16, 17 Craniopharyngiomas usually grow in tight connection with adjacent anatomical structures such as the hypothalamus, optic chiasm, and optic nerves, as well as arteries of the circle of Willis and the cerebral parenchyma itself. Complete resection has been attempted since the 1950s, however, with an initial mortality rate of 20%.18, 19 Over the years, neurosurgical techniques have improved with the implementation of microsurgery; however, a macroscopic total resection still remains possible in only approximately 8% to 78% of all patients,20–22 with higher rates of side effects noted in the pediatric population.22 Complete neurosurgical resections can commonly be performed in anatomic locations lying ventrally of the optic chiasm or purely intrasellar and with a tumor dimension <3 cm.23 In larger tumors, or in tumors with a macroscopic cystic appearance as well as localizations behind the optic chiasms, total resection is limited by the high risk of surgery-related toxicity.
For incompletely excised craniopharyngiomas or after a biopsy only has been performed, RT has been recommended in the past.24, 25 Comparable results to complete neurosurgical resection have been achieved.26–30 However, each treatment option is associated with a distinct risk profile; surgery is commonly associated with acute side effects that can be classified as ‘all or nothing,’31 whereas RT-induced side effects may occur over a long time-course after RT.
Commonly, external radiation using conventional techniques has been recommended, with total doses ranging between 45 Gy and 55 Gy; recurrence rates of 15% to 60% have been reported,32–34 with the majority ranging between 20% and 30%.
RT of tumors in close vicinity to risk structures is limited by the tolerance dose of these organs. Modern high-precision RT techniques allow the precise application of a high dose to a defined target while surrounding organs at risk can be spared. With FSRT, this physical advantage is accompanied by the radiobiologic advantage of fractionation with regard to treatment-related side effects, because the total dose is applied in a certain number of fractions.35, 36 Previous preliminary results from our institution demonstrated excellent local control rates, with 100% at 10 years and an overall survival of 83% at 10 years. Treatment-related side effects such as endocrine or visual deficiencies were extremely low.14 Only recently, Minniti et al.37 reported similar results in 39 patients treated with fractionated stereotactic conformal RT, with a local progression-free survival rate of 92% and an overall survival rate of 100% at 5 years.
In patients with benign tumors such as craniopharyngiomas, long-term outcome is extremely important because survival after RT is comparable to the normal life expectancy. Our long-term results after a median follow-up of 98 months emphasize that FSRT yields an extremely good treatment outcome with very low rates of associated side effects. With FSRT, it is possible to reduce the safety margins of 2 cm used in conventional RT to 2 mm, which leads to smaller PTVs and subsequently increased sparing of normal tissue. In combination with the radiobiologic effect of fractionation, this explains our low rates of radiation-induced side effects.
Another important issue is the timing of RT. It is currently still discussed controversially whether RT should be applied early in the time course or should be reserved for tumor recurrence. Regine et al.29 reported 78% survival rate at 20 years in children treated at the time of first diagnosis as opposed to 25% in children treated for recurrence. However, this could not be observed in the adult subgroup. Jose et al.11 reported that RT administered in a dose of 50 Gy for tumor recurrence yielded a 77% survival rate at 10 years. Habrand et al.38 also observed a better outcome in patients treated immediately as opposed to receiving RT for disease recurrence. A recent study published by Pemberton et al.39 found RT to be effective for craniopharyngiomas adjuvantly as well as at recurrence. Our long-term experience as reported in the current analysis supports the idea that FSRT is an effective means to achieve long-term control without severe side effects leading to an impairment of quality of life, with no difference in local control reported in patients treated directly after neurosurgical intervention after primary diagnosis as opposed to patients treated for tumor recurrence during follow-up. No formal neuropsychologic assessment was performed in these patients; however, the observations made by the treating physicians could not document any neuropsychologic alteration in 38 of 40 patients.
Particle therapy such as proton RT offers a distinct physical dose advantage, resulting in high doses to a target volume and a steep dose fall-off around the treatment volume. However, it is known that proton RT does not offer significant advantages for the high-dose regions of the treatment plan; however, with proton RT, the low-dose to medium-dose regions can be minimized, potentially reducing the long-term treatment-related side effects associated with these RT doses.31, 40 Fitzek et al.31 recently reported on a group of 15 patients with craniopharyngiomas who were treated with a combination of photons and protons at the Harvard Cyclotron Laboratory, with a median proton component of 26.9 cobalt Gy equivalent (GyE), and a median total doses of 56.9 GyE. After a median observation time of 13.1 years after RT, the 10-year local control rate was 85%. The functional status of all patients who were alive at the time of last follow-up (11 of 15 patients) was unaltered compared with the pre-RT status. A main caution in applying proton RT for craniopharyngiomas is that cysts may form during treatment, and with these liquid-filled cystic structures the PTV is altered. With the steep dose gradient provided by protons, there is a high risk that the target volume might be missed during therapy if the tumor volume changes. In the current study, we considered a sufficient safety margin for target volume definition with respect to potential increase in cystic components. Moreover, MRI was performed once during treatment in asymptomatic patients, and more frequently as needed clinically in patients developing symptoms to evaluate potential alterations in cystic components.
In conclusion, high-precision RT can lead to excellent outcomes in patients with craniopharyngiomas, with high local control rates and the preservation of organ function. No serious long-term effects were observed, and control rates as well as overall survival are within the range reported with conventional RT.