Introduction

Craniopharyngiomas constitute between 2.5 and 4 % of all benign intracranial tumors. The tumors originate from the remnants of the hypophyseal pharyngeal duct or squamous epithelial rests [11, 18]. Craniopharyngiomas are most commonly located in the sellar and suprasellar areas, close to vital structures such as the hypothalamus and thalamus [11]. The tumors are locally aggressive, with a high rate of recurrence and regrowth. Therefore, craniopharyngiomas are considered semimalignant tumors and are recognized as challenging to treat despite their benign microscopic appearance [33, 36]. Generally, complete surgical removal of craniopharyngiomas is believed to achieve the most satisfactory results and has a low recurrence rate. However, complete resection is not always feasible because of the deep location of the tumors or their close proximity to vital neurovascular structures [28]. Moreover, despite radical excision, most studies have reported recurrence rates between 5 and 57 % [7, 37]. Treatment of recurrent or residual craniopharyngioma is important to achieve a good prognosis and long-term tumor control. Depending on tumor location and extent, various transcranial approaches have traditionally been used to treat craniopharyngioma, including the bifrontal, subfrontal and pterional approaches. Repeat surgery is more technically difficult than the first surgery and occasionally results in increased morbidity and mortality because of factors including tumor adherence at the surgical scar and brain retraction [30, 38]. The transsphenoidal approach (TSA) has also been performed using a microscopic or endoscopic device, though classically the technique has been limited to predominantly intrasellar or intrasuprasellar subdiaphragmatic lesions [13, 39]. Recently, the development and application of extended TSA makes it practicable to remove recurrent or residual craniopharyngiomas [3, 17]. We present a series of 12 cases in which patients were surgically treated using extended endoscopic endonasal TSA and suggest that the technique is feasible for treatment of recurrent or residual craniopharyngioma.

Materials and methods

Patients

After the approval for the retrospective analysis by the Institutional Review Board, the electronic medical records were reviewed from a total of 12 adult patients with recurrent or residual craniopharyngioma who had undergone surgery using extended endoscopic endonasal TSA by a single surgeon at our institution between August 2010 and June 2013. The patients included six males and six females between the ages of 25 and 62. All patients were treated using surgical intervention upon their initial diagnosis; the mean number of surgeries prior to the extended endoscopic endonasal TSA was 1.3. Of the 12 patients, 2 had undergone 2 previous surgeries, 1 had undergone 3 previous surgeries, and the remaining patients had undergone 1 surgery. The type of previous surgery was the transcranial approach in nine patients and endoscopic TSA in two patients, and the remaining patient had undergone two previous surgeries including both the transcranial approach and endoscopic TSA. After extended TSA, the mean follow-up period was 15.8 months, with a range of 4 to 32 months. Four patients also underwent adjuvant radiotherapy. Of the two patients for whom subtotal resection was achieved, one patient was not eligible for adjuvant treatment because of previous radiotherapy.

Clinical data for each patient were obtained by reviewing relevant medical records. Clinical information, including the extent of resection, stalk preservation, endocrinological and visual outcome, and postoperative complications were analyzed. Visual outcome was evaluated using a visual impairment score (VIS) according to the German Ophthalmological Society guidelines developed by Fahlbusch [8]. The VIS was calculated by translating the scores from specific tables based on visual acuity and visual field defects. Pre- and postoperative scores were added, and the sum values ranged from 0 (best) to 100 (worst). All 12 patients were given endocrine assessments both pre- and postoperatively. Complete serum pituitary hormone levels were evaluated, and the results were interpreted by an endocrinologist. Based on the endocrinological assessments, nine patients suffered from preoperative pituitary hormone dysfunctions. One patient had hypogonadism, and the others had preoperative panhypopituitarism of the anterior pituitary gland, with or without DI. Three patients had normal preoperative pituitary hormone function.

Neuroradiological assessment was performed using magnetic resonance imaging (MRI) both preoperatively and within 2 days following the operation. Preoperatively, MRI was used to classify the lesion location and type (predominantly cystic, solid or mixed). In total, there were five cystic, three solid and four mixed lesions. The degree of tumor resection was confirmed by assessing both intraoperative findings and postoperative MRI. Gross total resection was defined as the absence of residual tumor visible based on these assessments [7]. Subtotal resection was defined as any small residual tumor confirmed intraoperatively or postoperatively.

Patient characteristics and preoperative clinical data are reported in Table 1.

Table 1 Summary of patient data
Full size table

Surgical procedures

All surgeries were performed using the endoscopic endonasal trans-tubercular and trans-sellar approaches. An image-guided neuronavigation system was used to confirm anatomical details and to remove the skull base bone. The details of the operative technique and process were similar to those previously published in the literature [10, 19]. After opening the dura mater, we dissected a plane between the tumor capsule and arachnoid membrane to protect the optic apparatus, which is supplied by perforator vessels, such as the superior hypophyseal artery. Next, the pituitary stalk was identified by dissecting the tissue between the pituitary gland and tumor capsule; if the pituitary stalk was confirmed to have obvious tumor invasion, it was sacrificed. Sharp dissection was performed to protect the optic nerve, tract and hypothalamus. Liliequist’s membrane was preserved if its integrity was confirmed. The reconstruction of the skull base bone was performed using various methods, including a nasoseptal pedicle flap [24].

Results

Total resection was accomplished in ten patients (83.3 %) and subtotal resection in two patients (16.7 %). In one case, severe adherence of the tumor to the posterior communicating artery prevented total resection; in the other, the arachnoid membrane was disrupted because of prior surgery, and there were severe adherences to the internal carotid artery and posterior communicating artery. The resection rate was calculated by comparing pre- and postoperative tumor volume. The overall resection rate was 97.8 %. Among the subtotal resection cases, the average resection rate was 87 %.

In three of the cases for which total resection was achieved, patients underwent adjuvant treatment because of rapid recurrence after a previous surgery. The extent of resection and information related to post-TSA adjuvant treatment are summarized in Table 2.

Table 2 The extent of resection and adjuvant treatment
Full size table

Postoperative endocrinopathy results had a relationship with whether stalk preservation was achieved. The stalk was preserved in six patients; in three patients, the stalk was resected because of intraoperative discovery of tumor invasion. The stalks of the remaining three patients had been resected during a previous surgery; those patients all suffered from preoperative panhypopituitarism. The three patients who received stalk resection showed postoperative endocrine dysfunction. In two of those three patients, the pituitary stalk was preserved only anatomically, and those two patients showed preoperative panhypopituitarism. New onset endocrinopathy occurred in one patient. Of the patients with normal preoperative pituitary function, no one was diagnosed with permanent postoperative endocrinopathy; these patients had suffered from transient DI for several months.

The preoperative mean total VIS was 28.6; the postoperative value was 29.6. Of the 12 patients, 3 had improved, 2 had deteriorated, and the others had unchanged total VIS. Among the improved patients, the visual field had normalized in two patients, and visual acuity had improved in one patient. However, two patients suffered decreases in visual acuity; one patient’s visual acuity VIS increased from 2 to 10, the other’s from 4 to 35. Neither patient complained of significant visual changes immediately after surgery, but showed signs of deterioration some days later. Both patients showed some improvement in visual acuity during follow-up.

Postoperative infection occurred in one patient, who suffered procedure-related fungal ventriculitis. The patient had preoperative panhypopituitarism and had undergone three previous operations (the most of all patients included in the study). The patient was also the only one to have undergone adjuvant radiotherapy before the extended TSA. There were no postoperative infections among the patients with normal preoperative pituitary function. We cannot entirely rule out the relationship between postoperative infection and the number of previous surgeries, preoperative hormonal status and prior adjuvant radiotherapy. There were no mortalities among the patients in the study, and no postoperative CSF leakage. The postoperative complications as well as endocrinological and visual outcomes are shown in Table 3.

Table 3 Postoperative complications with endocrinologic and visual outcomes
Full size table

Discussion

Tumor recurrence is a common event in the management of craniopharyngioma. However, there is some debate over the optimal treatment for recurrent or residual craniopharyngioma. After initial surgical intervention, recurrence rates between 9 % and 51 % have been reported [1, 14, 32]. Craniopharyngioma recurrence is considered a significant factor related to long-term mortality. Many studies have confirmed that the extent of resection is the most important predictor of recurrence [6, 7, 26, 32]. Reported recurrence rates are between 6 % and 20 % in cases of GTR and 10 % to 75 % in cases of STR. In GTR cases, the mean period to tumor recurrence was between 20 and 96 months; in the STR group, the value was between 3.5 and 7.1 months [1, 4, 6, 12, 27]. Total resection should be the initial goal when treating craniopharyngioma. However, total resection is not always possible because of the tumor’s deep location or the involvement of neurovascular structures. Adjuvant radiotherapy, therefore, is employed as a combination therapy in patients with subtotal tumor resection. In cases of subtotal resection, it is generally accepted that postoperative radiotherapy reduces recurrence rates [30]. Therefore, some authors have emphasized the effectiveness of radiotherapy for treating recurrent or residual craniopharyngioma; a good long-term outcome, similar to that of gross total resection, has been reported [9, 14, 16, 29, 31]. If, after first line treatment using these modalities, tumor recurrence or progression of the residual tumor is observed, a second radiotherapy treatment is not possible because of the risk of mass effect-induced neuropathy. Moreover, if previous radiotherapy has not been administered as a first line treatment, surgical resection (when possible) for recurrent or residual craniopharyngioma reportedly results in good outcomes [27]. Surgical management of recurrent or residual craniopharyngioma is considered the appropriate choice.

Repeat transcranial surgery is recognized as more difficult than primary surgery. Many reports have indicated that repeat surgery results in lower total resection rates and higher postoperative mortality rates [2, 7, 35]. These results are due to the technical difficulties associated with repeat surgery; it is difficult to distinguish the natural cleavage plane in the presence of arachnoidal scars from prior surgery and the loss of a gliotic reaction between the tumor and the surrounding normal structures. Brain manipulation and severe adherence to surrounding vital organs, including the perforator vessels, optic apparatus, hypothalamus and thalamus, also make repeat surgery more challenging, resulting in higher morbidity and mortality.

Recently, treatment series analyzing craniopharyngioma using a pure endoscopic endonasal TSA have been reported [5, 10, 22, 23]. In the present study, we used a purely extended endoscopic endonasal TSA to remove only recurrent or residual craniopharyngioma. This helped avoid the risk of transcranial route revision via a virgin route. Furthermore, the technique is widely known to have advantages, such as providing a more direct view of the neurovascular structures of the suprasellar area without requiring brain retraction.

The most important factor related to the long-term outcome in craniopharyngioma patients is the extent of tumor resection. In our series, gross total resection was obtained in approximately 83 % of cases, without mortality. The overall resection rate was 87 % in cases of subtotal resection. Kitano and Taneda reported a total tumor removal rate of 85.7 % when treating recurrent craniopharyngioma using extended TSA; our data are consistent with those results [21]. A close correlation between the extent of resection and visual outcome has been reported in some studies [20, 34]. Patients who underwent subtotal resection were at risk of visual disturbance occurring during long-term follow-up [20]. In our study, there was no significant difference between pre- and postoperative visual impairment scores, except in one case.

It is generally known that postoperative CSF leakage occurs less frequently when using transcranial approaches. However, recently developed reconstruction techniques have reduced CSF leakage rates significantly, to approximately 0 % to 4 % [15, 24, 25]. There was a little difference in the reconstruction method adapted for each patient. The skull base of almost all patients was reconstructed using the hydroxyapatite cement paste after an autologous or artificial patch graft, which was covered by a nasoseptal pedicle flap. A button graft with fascia lata was applied in one patient, which was also followed by a nasoseptal pedicle flap. There were no cases of postoperative CSF leakage in our series, and only one patient suffered a postoperative infection.

The limitations of this study are its small number of patients and a short follow-up period. Larger studies with more extensive follow-up will help to establish the optimal role of extended TSA for recurrent or residual craniopharyngioma. Furthermore, the correlation between tumor size, location, tumor nature and individual anatomical features of the suprasellar area should be analyzed.

Conclusion

The extended endoscopic endonasal transsphenoidal approach is a safe and effective treatment option for recurrent or residual craniopharyngiomas that were initially treated surgically.