Outcomes following endoscopic, expanded endonasal resection of suprasellar craniopharyngiomas: a case series

Transcription

1 See the Editorials and the Response in this issue, pp 1 5. J Neurosurg 109:6 16, 2008 Outcomes following endoscopic, expanded endonasal resection of suprasellar craniopharyngiomas: a case series PAUL A. GARDNER, M.D., 1 AMIN B. KASSAM, M.D., 1 4 CARL H. SNYDERMAN, M.D., 1,3 RICARDO L. CARRAU, M.D., 1,4 ARLAN H. MINTZ, M.D., 1 STEVEN GRAHOVAC, M.D., 5,6 AND SUSAN STEFKO, M.D. 7 Departments of 1 Neurological Surgery, 4 Otolaryngology, 5 Radiology (Division of Neuroradiology), and 7 Ophthalmology; 2 Minimally Invasive endoneurosurgery Center; and 3 Center for Cranial Base Surgery, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania; and 6 Department of Radiology, Christiana Hospital, Newark, Delaware T Object. Craniopharyngiomas are challenging tumors that most frequently occur in the sellar or suprasellar regions. Microscopic transsphenoidal resections with various extensions and variations have been performed with good results. The addition of the endoscope as well as the further expansion of the standard and extended transsphenoidal approaches has not been well evaluated for the treatment of this pathological entity. Methods. The authors performed a retrospective review of all patients who underwent a purely endoscopic, expanded endonasal approach (EEA) for the resection of craniopharyngiomas at their institution between June 1999 and February Endocrine and ophthalmological outcomes, extent of resection, and complications were evaluated. Results. Sixteen patients underwent endoscopic EEA for the resection of craniopharyngiomas. Five patients (31%) presented with recurrent disease. Complete resection was planned in 11 of the 16 patients. Three elderly patients with vision loss underwent planned debulking, 1 patient with vision loss and a moderate-sized tumor had express wishes for debulking, and 1 patient had a separate, third ventricular nodule that was not resected. Of those in whom complete resection was planned, 91% underwent near-total (2/11) or gross-total (8/11) resection. No patient who underwent gross-total resection suffered a recurrence. The mean follow-up period was 34 months. Of the 14 patients who presented with vision loss, 93% had improvement or complete recovery and 1 patient s condition remained stable. No patient experienced visual worsening. Eighteen percent of patients (without preexisting hypopituitarism) developed panhypopituitarism and 8% developed permanent diabetes insipidus. There were no cases of new obesity. The postoperative cerebrospinal fluid leak rate was 58%. All leaks were resolved, and there were no cases of bacterial meningitis. There was 1 vascular injury (posterior cerebral artery perforator branch) resulting in the only new neurological deficit. No patient died. Conclusions. Endoscopic EEA for the resection of craniopharyngiomas provides acceptable results and holds the potential to improve outcomes. (DOI: /JNS/2008/109/7/0006) KEY WORDS craniopharyngioma endonasal resection endoscopic resection outcome HE microscopic, transsphenoidal approach for the resection of certain craniopharyngiomas is well established and generally accepted, 6,13 15,17,20 24,29,30 and there are several studies comparing this approach with a transcranial approach. 21,24,30 Although the transsphenoidal approach is traditionally limited to predominantly sellar lesions with secondary suprasellar extension, 23 the development and application of the extended approach have provided greater access to suprasellar lesions. 17,20,25 Despite this modification, outcomes from the transsphenoidal approach for the resection of suprasellar tumors with adherence to critical surrounding structures have led to more frequent Abbreviations used in this paper: CSF = cerebrospinal fluid; DI = diabetes insipidus; EEA = expanded endonasal approach; GH = growth hormone; GKS = Gamma Knife surgery; ICP = intracranial pressure; PCA = posterior cerebral artery; TSH = thyroid-stimulating hormone; VP = ventriculoperitoneal. recurrences and poorer overall outcomes compared with purely sellar tumors. 17 Some of the limitations of this approach, as previously described, may be secondary to the use of the microscope, 2 distance from the surgical target, the small portal of entry to the surgical field (that is, the nostril), and the need for a speculum. Attempting to remedy these issues, in 1979 Halves and Bushe 10 recognized the facilitatory role of the endoscope when accessing suprasellar lesions. To date, however, the endoscope has played a limited role in most studies of craniopharyngioma, and there are few studies 7 in which the outcomes following endoscopic, endonasal craniopharyngioma resection have been evaluated. We present a series of 16 cases in which we undertook a purely endoscopic EEA for the resection of craniopharyngiomas. Short-term ophthalmological and endocrine outcomes along with extent of resection were examined. This report is meant to be an early evaluation of the technique s value in the management of these formidable lesions. By 6 J. Neurosurg. / Volume 109 / July 2008

2 Outcomes following EEA for resection of craniopharyngiomas comparing these results with those previously published for other approaches, we hope to better judge both approaches in terms of their respective advantages and disadvantages. Although authors of many past studies have assessed craniopharyngioma resection via a transsphenoidal route, 2,17, 23,24,29 few 7 have used a purely endoscopic technique and most are biased toward the resection of tumors with minimal supra- or parasellar extension. Our series consists of tumors located above and/or with extension superior and lateral to the sella, superior and posterior to the chiasm, and even extending into the third ventricle. Methods After obtaining institutional review board approval, we reviewed the medical records and imaging studies of all patients with the histopathological diagnosis of craniopharyngioma treated via an EEA between June 1999 and February No patient was excluded or lost to follow-up, resulting in a series of 16 patients (Table 1). Patients ages ranged from 36 to 80 years (mean age 55 years). There was a male/ female ratio of 5:3. Five patients (31%) presented to us with recurrent disease. One patient had undergone prior transsphenoidal resection, 4 had prior craniotomies (1 of whom also underwent prior stereotactic cyst drainage with intracystic brachytherapy), and 1 had undergone radiosurgery. In addition, 1 patient underwent a previous transsphenoidal procedure for a pituitary adenoma (histopathology verified). Other complicating conditions included obesity, chronic obstructive pulmonary disease, and pregnancy. Most patients (14 [87.5%] of 16) presented with progressive visual deficit as the primary indication for surgery (Table 2). We report ophthalmological and endocrine outcomes, as well as extent of resection and complications. The follow-up period for this study ranged from 2 to 82 months (mean follow-up 34 months). Operative Technique J. Neurosurg. / Volume 109 / July 2008 TABLE 1 Characteristics in patients who underwent an endoscopic EEA for craniopharyngioma Case Age (yrs), Date of Op Tumor No. Sex (mo/day/yr) Comment Type* 1 42, M 06/08/99 none I 2 36, F 04/12/02 presented 14 wks pregnant I 3 46, F 10/21/02 previous stereotactic cyst drainage, IIIb intracystic brachytherapy & eyebrow craniotomy 4 43, M 11/05/02 none IIIa & b 5 50, M 05/08/03 recurrent tumor (previous bicoro- II nal craniotomy) 6 58, M 07/16/03 obese/severe COPD IIIb 7 69, M 08/20/03 transsphenoidal adenoma resec- II tion 7/17/ , M 08/25/03 recurrent tumor (previous bicoro- I nal craniotomy) 9 56, F 01/26/04 none IIIb 10 51, F 06/21/04 none IIIa & b 11 55, F 07/02/04 recurrent tumor II 12 52, F 09/01/04 presented w/ increased thirst II 13 79, M 12/06/04 none IIIa 14 80, M 01/10/05 prior craniotomy SRS; ataxia w/ IIIb recurrence (pontine compression) 15 48, M 08/08/05 panhypopituitarism, gland II completely calcified 16 79, M 02/20/06 none IIIa & b * Type I denotes preinfundibular; Type II, transinfundibular; and Type III, retroinfundibular (a, rostral extension [interpeduncular cistern]; b, caudal extension [prepontine cistern]). Abbreviations: COPD = chronic obstructive pulmonary disease; SRS = stereotactic radiosurgery. We detail the operative technique elsewhere. 19 As such, only a brief overview will be provided here. The operations were all performed by 2 surgeons (A.B.K. or P.A.G. and C.H.S. or R.L.C.), a neurosurgeon, and an otolaryngologist. A binarial, purely endoscopic, image-guided approach was used. The patient was placed supine, the head fixed in pins, the neck in a neutral or slightly extended position (10 15, depending on the anterior/superior extent of the tumor), and the head turned 5 10 toward the surgeon to allow comfortable hand position and improve access. First, oxymetazoline-soaked pledgets were placed into the nares for several minutes, prior to preparation. Then, the imageguided system was calibrated with the patient. We preferred to use MR imaging, but we used CT angiography in cases in which the patient was unable to undergo MR imaging or the bone anatomy or involved vasculature was critical for resection. Next, the abdomen (for fat graft harvest) and midface were prepared with Betadine-soaked sponges. The patient was then draped, leaving only the nose and prepared abdomen exposed. Imaging guidance was once more confirmed, using external landmarks. Entering the right nostril, the middle turbinate was either lateralized or removed under endoscopic guidance and the anterior wall of the sphenoid was entered. The contralateral nostril was then entered and the sphenoid similarly approached, with lateralization of the middle turbinate. Next, the critical step of partial removal of the posterior internasal septum (~ 1 cm) was performed. This permitted introduction of instrumentation from the contralateral naris without obscuring the endoscopic view. After the posterior wall of the sphenoid sinus was exposed, the image guidance was once more introduced to confirm bone landmarks as well as tumor location. This step is important, as the normal sellar anatomy is often greatly distorted by tumor. The bone over the tumor s dural covering was next drilled and completely removed using angled rongeurs. Through this process, the surrounding anatomy was defined, with special attention given to the location of the cavernous sinuses, carotid arteries, and optic nerves. The bone over the superior intercavernous sinus was removed next (if necessary) using a combination of drill and rongeur. At this point, the critical portion of the sellar bone removal, over the medial opticocarotid recess, was carried out as described elsewhere. 19 This variable indentation on the sphenoid sinus side of the lateral tuberculum represents the medial confluence of the optic nerve, carotid artery, and anterior skull base, providing a keyhole of sorts for this access. When necessary, the superior intercavernous sinus was carefully and completely coagulated using bipolar cautery, prior to its division. This additional rostral exposure often provided the starting point for tumor surgery and/or dural entry, especially when approaching a supra- or retrochiasmatic lesion. In suprasellar lesions, the dura was opened in a cruciate manner, and the tumor was located if not immediately ap- 7

3 P. A. Gardner et al. parent. An extracapsular dissection allowed the surgeon to define tumor as well as key surrounding structures (for example, the optic chiasm and perforating vessels, pituitary gland and stalk, and major arteries). The tumor capsule was then entered, and debulking was performed. Finally, the extracapsular dissection was completed, and the tumor was removed from any adherent normal structures by using sharp dissection. After a final inspection, the dural closure was begun. This reconstruction step has undergone considerable evolution during our experience and continues to be evaluated for improvement. By the end of the series, a lumbar drain was used only in cases in which the third ventricle was entered or in communication as a result of the resection. Throughout this series, some type of onlay graft, followed by fat graft has been used. At the end of the series, this has progressed to an inlay of DuraGen (Integra Lifesciences Corp.), followed by an acellular dermal onlay graft. The sphenoid sinus was ultimately packed with fat, which was covered with Surgicel (Ethicon) and followed by nasal stents and either pledgets or a Foley catheter balloon as a buttress. Since then, we have begun using a vascularized nasal septal mucosal flap as a final layer over the onlay graft with promising results. 9 Ophthalmological Results TABLE 2 Ophthalmological results* Case No. Preop Deficit/Presenting Symptom Postop Result 1 chiasmal compression improved 2 progressive chiasmal syndrome resolved 3 progressive, severe visual loss unchanged/stabilized 4 mild mod chiasmal compression resolved 5 lt rt optic neuropathy improved 6 6-mo history decreased visual acuity resolved (mild mod afferent visual defect) 7 recurrent chiasmal compression delayed postop deficit, improved w/ VP shunt compared w/ preop 8 chiasmal compression improved 9 lt rt superior temporal quadrantanopia improved 10 progressive vision loss resolved 11 profound vision loss improved 12 vision loss resolved 13 vision loss improved 14 no vision loss/brainstem compression no change 15 no vision loss/headache no change 16 vision loss resolved * mod = moderate. Results All 16 patients underwent a detailed neuroophthalmological examination before and after surgery (Table 2). The conditions of 2 patients without preoperative visual deficits were unchanged postoperatively. The remaining 14 patients had progressive visual deficits preoperatively. Six (43%) of these patients had complete resolution of their visual defect (Fig. 1), and 7 (50%) had improvement but not complete resolution (Fig. 2). One patient s deficit remained stable. It was difficult to determine changes in this patient s vision related to the endoscopic procedure because her vision was extremely poor preoperatively, due in part to unrelated retinal disease. No patient had worsening of vision from optic apparatus damage following surgery. One patient had visual worsening 4 days after surgery secondary to hydrocephalus. His vision improved following the placement of a VP shunt. Endocrine Results All 16 patients underwent endocrine evaluation both preand postoperatively with complete serum pituitary hormone panels interpreted by an endocrinologist. Five patients suffered from panhypopituitarism (anterior gland) preoperatively (1 had DI), 1 had DI with other deficits, and 1 had isolated DI (Table 3). Only 1 (8%) of 12 patients who underwent long-term follow-up suffered permanent DI. Four others developed temporary DI in the postoperative period that resolved at last follow-up without further treatment. Of the 11 patients without preoperative panhypopituitarism, 2 developed new-onset panhypopituitarism following surgery (18%). One patient had a new, single axis deficit that required TSH replacement. One patient s hormone dysfunction (GH and TSH) resolved following tumor resection. The 7 remaining patients had no change in the anterior gland. When evaluated by tumor classification, preoperative anterior gland panhypopituitarism or significant dysfunction (Case 12) seemed to be more common in primary (nonrecurrent) Type II tumors (2 of 3, compared with 0 of 2 Type I and 2 of 6 Type III). All patients with Type I tumors suffered from panhypopituitarism postoperatively (1 of these patients had already been suffering from panhypopituitarism). Not surprisingly, the only patient with preoperative DI had a Type II tumor. In addition, permanent DI was more common, although not universal, in postoperative (whether treated with EEA or prior surgery) Type II tumors (3 of 5 [1 new after EEA] vs 1 of 3 [new] Type I and 0 of 8 Type III). Extent of Resection and Recurrence Extent of resection was determined using pre- and postoperative volumetric analysis of MR images (Table 4). Fat saturation was used during postoperative contrast-enhanced imaging to remove any confusion between enhancing tumor and fat graft. The comparison was performed by an independent neuroradiologist (S.G.). Evaluation of the series without consideration of goal of surgery has shown that 12 (75%) of the 16 patients underwent at least near-total resection. Two of these had a small amount of residual enhancement within the stalk (Fig. 3) and 2 had 5% residual tumor, for a gross-total resection rate of 50%. However, complete resection was planned for only 11 of the 16 patients. Of these 11 patients, 8 (73%) underwent gross-total resection (Fig. 4) and 2 other patients (Cases 9 and 12) had only residual enhancement within the stalk (Fig. 3), resulting in 91% (10/11) with gross-total or near-complete ( 95%) resection. The 2 patients with residual stalk enhancement underwent intentional stalksparing surgery during which tumor was dissected from the stalk, but the stalk was not sacrificed, thus preserving pituitary function. One of these patients suffered regrowth, and the other s condition has remained stable (treated with radiosurgery postoperatively). One patient underwent 8 J. Neurosurg. / Volume 109 / July 2008

4 Outcomes following EEA for resection of craniopharyngiomas FIG. 1. Humphrey visual field tests. A: Preoperative study demonstrating a dense temporal defect. B: Postoperative visual field testing showing resolution of visual deficit. an unintentional, subtotal (70%) resection. This case was aborted due to arterial (PCA branch) injury. All patients with primary (nonrecurrent) tumors underwent gross- or near-total resection. Six (75%) of 8 patients with primary tumors underwent gross-total resection. The other 2 patients with primary tumors underwent stalk-sparing surgery. Five patients underwent intentional, incomplete resections/debulking. Three of these patients were 65 years old with large tumors causing visual compromise (Fig. 5). Two were recurrent tumors. One patient had a large, suprasellar solid portion as well as a large cyst that extended well into the third ventricle (similar to a Type IV, purely intraventricular lesion). He was a highly educated patient and expressed distinct wishes for a subtotal resection with preservation of pituitary function. The cyst was treated separately via stereotactic 32 P injection. All patients had relief of compressive symptoms, and residual solid tumor was treated with radiosurgery. One patient underwent complete resection of a suprasellar craniopharyngioma with the exception of a small, separate, enhancing nodule in the lateral wall of the third ventricle (Fig. 6). This area was treated using radiosurgery, but the nodule recurred considerably and was treated using repeated EEA. The decision to merely J. Neurosurg. / Volume 109 / July 2008 debulk these tumors rather than attempt complete resection was complex and depended on patient age and overall condition, patient wishes, potential impact of aggressive resection given tumor location (for example, hypothalamic or vascular involvement), and presenting symptoms among other factors. In this series we treated tumor remnants, whether left intentionally or not, using radiosurgery. No patient suffered recurrence in the setting of gross-total resection. Four patients with known residual tumor experienced regrowth. One of these patients, mentioned earlier, with only residual rim enhancement within the stalk, experienced regrowth of the remnant. This patient (Case 12) had a particularly aggressive tumor and underwent repeated endoscopic EEA twice, successfully treating visual symptoms, followed by stereotactic radiosurgery. The patient s tumor has responded to radiosurgery and was noted to be decreasing at last follow-up. Three other patients had regrowth of a known remnant. All 3 were recurrent prior to EEA. The patients in Cases 5 and 11 experienced regrowth and were treated endonasally, and the patient in Case 7 underwent a second resection via a craniotomy due to growth lateral to the middle cerebral artery. The advantage to the endonasal approach lies in treating tumors that are medial to neurovas- 9

5 P. A. Gardner et al. FIG. 2. Humphrey visual field tests. A: Preoperative study showing dense bilateral defects, greatest bitemporally. B: Postoperative study performed in the same patient demonstrating significant improvement compared with preoperative testing. cular structures, thus allowing resection without traversing these structures. This advantage was negated in this situation, making an endonasal approach impractical. This patient eventually died (3 years after EEA) secondary to complications associated with stereotactic cyst treatment. Postoperative Complications The most common complication in this series was postoperative CSF leak. There were 11 CSF leaks (58%) during 19 total operations (including 3 reoperations for recurrence/regrowth; 69% not including reoperations) (Table 5). Lumbar drain diversion in the immediate postoperative period is used in all cases requiring extensive arachnoid dissection or ventricular transgression. One leak resolved with lumbar drainage alone, and 10 leaks were managed with reexploration and repair. Long-term CSF diversion was required in 6 (40%) of 15 patients who had lumbar puncture proven chronic elevation in ICP. Another patient developed delayed visual decline postoperatively because of documented hydrocephalus. This condition resolved to the immediate postoperative, improved state after shunt treatment. The overall shunt placement rate was 7 (44%) of 16, and despite CSF leaks there were no cases of bacterial meningitis. Shunt treatment was not used as a primary modality for CSF leak repair. Rather, it was used as an adjuvant in those patients with recurrent leakage and documented (lumbar puncture) increase in ICP. Several tumors were closely associated with or attached to posterior and anterior circulation arteries. During one such case, a PCA (P 1 ) perforating vessel was injured. Copious irrigation with the aid of a ventriculostomy catheter inserted into the naris revealed the site of arterial avulsion, and a successful repair/closure was achieved using hemostatic agent (Syvek marine polymer [off-label use]) with gentle compression. The patient did suffer a small thalamic infarction from this arterial injury with resultant aphasia, but fortunately has recovered well with minimal residual aphasia. The overall vascular injury rate for all surgeries 8 was 5% (1 of 19). The rate of permanent neurological deficit is the same (same patient). No patient died during or as a result of surgery or during the 30-day postoperative period. As mentioned earlier, 1 patient died following a stereotactic cyst treatment 3 years after EEA. One patient died of complications of anticoagulation therapy (acute subdural hematoma, distant from the site of surgery) 2 months following surgery. Another patient died of an unrelated medical condition ~ 2 years after surgery. 10 J. Neurosurg. / Volume 109 / July 2008

6 Outcomes following EEA for resection of craniopharyngiomas TABLE 3 Endocrine results* Posterior Pituitary Function (DI) Anterior Pituitary Function Case Postop/ Permanent No. Preop Deficit Postop Deficit (long term) Preop DI Transient DI Postop DI 1 none panhypopitism (except ACTH) no yes yes 2 none panhypopitism no yes no 3 panhypopituitarism no change no no no 4 none none no yes no 5 none none yes NA NA 6 panhypopituitarism panhypopituitarism no no no 7 hypothyroidism & hypogonadism no change no no no 8 panhypopituitarism no change yes NA NA 9 GH & thyroid axis deficits normal no no no 10 IGF-I deficit & TSH elevation continued hypothyroidism, Hashimoto no yes no thyroiditis & papillary carcinoma diagnosed 11 hypothyroidism, borderline adrenal in- hypothyroid & adrenal insufficiency no no no sufficiency 12 hypothyroidism, adrenal insufficiency no change yes NA NA 13 hypogonadism, borderline adrenal in- 1-mo FU (hypothyroidism at discharge), no yes NA; 1-mo FU (DI sufficiency lost to FU at discharge) 14 panhypopituitarism no change no no no 15 panhypopituitarism no change no yes no 16 thyroid & testosterone deficiencies no change no no no * ACTH = adrenocorticotropic hormone; FU = follow-up; IGF-I = insulin-like growth factor I; NA = not applicable. The DI was present following a prior craniotomy. Discussion Although small, this series demonstrates the potential of the purely endoscopic EEA as another aid in the treatment of suprasellar craniopharyngiomas. The addition and sole use of the endoscope in the expanded, microscopic approaches described by others 17,20,25 allows improved visualization, which, in turn, allows the further expansion of the previously described approaches. TABLE 4 Extent of resection Vol (cm 3 ) Case % Resection FU No. Preop Postop (solid & cystic portion) (mos) *; cyst progression ; regrowth* ; regrowth* (residual rim enhance ment w/in stalk) ; regrowth (residual enhancement 90.7; regrowth 29.5 w/in stalk) * * * Complete resection not planned. Small separate enhancing nodule in the lateral wall of the third ventricle not resected. J. Neurosurg. / Volume 109 / July 2008 Craniopharyngiomas are notoriously difficult to cure, but complete resection has been shown to provide a survival benefit. 1,3 5,8,16,27,33,34 Complete resection rates reported in the literature range from 6 to 0%, 3,5,6,12,13,24,26,27,31,32,34 36 with the exception of the study by Yaşargil et al., 36 in which the authors reported a 90% gross-total resection rate (at the cost of a 16.7% mortality rate). Our small series of suprasellar tumors has a 73% gross-total resection rate for tumors intended for complete resection (50% for the overall series). The series size limits direct comparisons; however, thus far, the rates of resection compared with historical case series do seem to be equivalent or better. Even more difficult to quantify is the patient s quality of life following tumor resection. Although the aim in most cases is complete resection and cure, there is a greater potential for some decrement in the patient s function, whether neurological, ophthalmological, or endocrinological, following radical surgery. In our study, all patients were functional postoperatively. Indeed, only 1 patient suffered a decrement in function (the patient with a PCA branch infarct). This resulted in a significant aphasia that recovered over the course of a year. This patient is living independently at home. In this series, the extent of resection was evaluated using an emphasis on an intent-to-treat analysis; only those tumors in which complete resection was the goal based on clinical factors were used in the analysis. Han et al. 11 have previously reported on surgical cohorts using this type of intention-to-treat analysis. No patient who underwent a gross-total resection experienced recurrence. The regrowth rate for the series is 25% (9% for those intended for complete resection). It is worth noting that all patients in our series who suffered recurrence or regrowth were patients with Type II (transinfundibular) tumors (Table 4), which grow and presumably originate from within the infundibulum. These tumors are perhaps 11

7 P. A. Gardner et al. FIG. 3. Preoperative (left) and postoperative (right) coronal MR images showing suprasellar, Type II (infundibular) craniopharyngioma treated using a stalk-sparing resection. The enhancement seen under the chiasm has not progressed during follow-up (see text on recurrence). FIG. 4. Contrast-enhanced T1-weighted MR images. Preoperative sagittal (A) and coronal (B) images showing a complex craniopharyngioma with suprasellar extension. Postoperative sagittal (C) and coronal (D) images showing gross-total resection. 12 J. Neurosurg. / Volume 109 / July 2008

8 Outcomes following EEA for resection of craniopharyngiomas FIG. 5. Contrast-enhanced T1-weighted MR images. Preoperative sagittal (A) and coronal (B) images showing a giant, complex craniopharyngioma, with extension into the third and lateral ventricles in an elderly man with vision loss. Postoperative sagittal (C) and coronal (D) images showing debulking of this giant tumor. Debulking was chosen due to the patient s advanced age and concern over potential hypothalamic and thalamic injury with aggressive resection. FIG. 6. Coronal contrast-enhanced T1-weighted MR images. Left: Preoperative image showing a suprasellar craniopharyngioma with a separate nodule (arrow) in the lateral wall of the third ventricle. Resection of the nodule was not planned as the floor of the third ventricle appeared intact on preoperative imaging. In retrospect and with further experience, perhaps this nodule would have been accessible. Right: Postoperative image obtained in the same patient showing resection of the suprasellar tumor, with the third ventricular nodular residual lesion (arrow). J. Neurosurg. / Volume 109 / July

9 P. A. Gardner et al. TABLE 5 Surgical complications* Case No. Complication Treatment Result 1 none NA NA 2 CSF leak reexploration resolution 3 CSF leak reexploration/vp shunt resolution 4 delayed CSF leak w/ HCP reexploration/vp shunt resolution 5 none NA NA 6 CSF leak w/ chemical meningitis, re- reexploration, awaiting reop resolution currence 7 postop HCP/cyst chiasm compression VP shunt improvement (delayed vision deficit) 8 none NA NA 9 PCA/P 1 injury (avulsion) w/ resultant therapy & brief rehab hospital stay, antico- recovered except trace dysarthria & partial lt thalamic CVA, DVT, de- agulants, reexploration twice/vp shunt facial droop, no sequelae, resolulayed CSF leak w/ HCP tion 10 CSF leak LD only resolution 11 CSF leak reexploration/vp shunt resolution 12 CSF leak, recurrence reexploration/vp shunt, resection twice resolution (1/05 & 5/05) w/o postop CSF leak, GKS 13 CSF leak, DVT reexploration/vp shunt, IVC filter, resolved, then died of SDH while anticoagulation taking anticoagulants 14 none NA NA 15 CSF leak reexploration/ld placed resolved 16 CSF leak reexploration resolved * CVA = cerebrovascular accident; DVT = deep venous thrombosis; HCP = hydrocephalus; IVC = inferior vena cava; LD = lumbar drain; rehab = rehabilitation; SDH = subdural hematoma. the most difficult to cure surgically (3 of the 4 in this study had been previously resected). In all of these cases, an attempt was made to preserve the stalk, with resultant preservation of existing endocrine function. 13 Although this was successful, in the future, it may be wiser to merely sacrifice the stalk in those cases in which there is no other remnant. Although certain to provide worse endocrine outcomes, in the long term, this practice may benefit the patient by reducing recurrence rates. 28 As an initial evaluation of the endoscopic EEA technique, this series shows it to be effective in chiasmal decompression. Although published studies have reported rates of postoperative visual decline ranging from 2 to 15%, 5,6,26,32,34,35 there were no such occurrences in this series. Furthermore, the likelihood of improvement in or complete recovery of vision was 92%. In terms of endocrinopathy, panhypopituitarism and DI are particularly common following the resection of craniopharyngiomas; rates in the literature have ranged from 24 to 66% and 43 to 79%, 6,31,34 36 respectively. In our series, 18% of patients developed panhypopituitarism and 8% developed permanent DI. Perhaps most surprising is the high rate of panhypopituitarism following resection of Type I tumors. This may be due to the fact that both Type I tumors without preoperative dysfunction were resected early (first) in the series. Therefore, it is difficult to know if this finding is a result of tumor location or a reflection of our learning curve. In contrast, Type III tumors had a very low rate of dysfunction (1 [20%] of 5). The aggressiveness of resection obviously has a significant effect on the aforementioned rates. This small series, however, appears to support the use of purely endoscopic techniques to provide a high rate of endocrine function preservation, while concurrently achieving respectable resections. This is especially true for Types I (preinfundibular) and III (retroinfundibular) tumors, where the stalk can be separated and potentially preserved. In this series, this is illustrated by Type III tumors, where there was a very low rate of postoperative pituitary dysfunction. Although there are only a few patients in whom complete resection with preservation of current hormonal status can be achieved (2 in this series, both with Type III tumors), we believe that an attempt at preservation of pituitary gland and stalk is reasonable and at times successful. Given the tumor type and locations, some degree of postoperative pituitary dysfunction is expected and preservation of function must be weighed against the recurrence risk. Once again, this is emphasized by Type II tumors, in which the tumor has often invaded the pituitary stalk. This problem illustrates the difficulty in dealing with such a pathological entity. Our experience with EEA resection of craniopharyngiomas has demonstrated the need for new techniques that prevent postoperative CSF leaks. Leakage rates were higher in this series than in our experience with other sellar and suprasellar lesions, and there are several reasons for this. One is the extensive, intentional violation of the anterior skull base dura that is necessary for lesion access. This larger opening is more difficult to effectively seal, and its proximity to critical neurovascular structures (carotid arteries, optic nerves, and so on) increases the challenge of obtaining a watertight repair. Of particular interest has been our realization that despite similar size exposures for other anterior skull base and parasellar pathological entities such as pituitary adenomas and meningiomas, craniopharyngiomas have a disproportionately higher incidence of CSF leakage following surgery. 7,18 This increased rate may be related to a higher incidence of either transient or permanent hydrocephalus associated with craniopharyngiomas. 5,12,36 This is also reflected in the high 14 J. Neurosurg. / Volume 109 / July 2008

10 Outcomes following EEA for resection of craniopharyngiomas rate of permanent CSF diversion demonstrated in this series. Given the early and often ineffective reconstruction techniques used in this series, any increased ICP, whether transient or permanent, was detrimental. However, the most important differentiating factor for increased CSF leakage is likely that resection of craniopharyngiomas, more than other parasellar lesions, involves the violation of arachnoid cisterns and offers entry into the third ventricle. This violation inevitably creates a greater obstacle to defect closure. Further modification of closure techniques will hopefully provide reduction in the CSF leakage rate, as has been the case after changes instituted later in this series. Through constant modification, the CSF leakage rates decreased considerably over the course of this study (Fig. 7). From 1998 to 2004, the leakage rate was 69%. The following year it was reduced to 20% (including the reoperations for tumor regrowth in this group, not listed in Table 5), once suturing and balloon buttresses were used. The recent addition of a vascularized, nasal septal flap 9 to our reconstructions appears to hold tremendous promise. It is important to note that all CSF fistulas were eventually sealed without associated morbidity. This result was associated with a high permanent CSF diversion rate in this series. There were no cases of bacterial meningitis and no complications associated with reoperation or shunt placement, although the potential for serious complication clearly exists. With careful management and awareness of this issue, CSF leakage is not believed to be an insurmountable problem and does not negate the efficacy of this approach for craniopharyngioma. As for vascular injury, perhaps the most feared complication of endoscopic manipulation of neurovascular structures, our rate of injury was low (5%). Just as important was the ability to successfully manage an arterial injury during a purely endoscopic endonasal approach. Conclusions J. Neurosurg. / Volume 109 / July 2008 FIG. 7. Bar graph illustrating CSF leakage rates early in the series compared with the last year of the series. The time periods are misleading, as only 1 case of CSF leakage occurred prior to 2003, making this comparison largely one of and The incorporation of the transplanum modification into traditional transsphenoidal approaches, facilitated by the use of the endoscope, has provided access to many tumors with suprasellar or even retrochiasmatic extension, previously thought to be unresectable via a transsphenoidal route. Craniopharyngioma is one such tumor. In our opinion, the use of the endoscope as the sole means of visualization in the EEA provides surgeons improved visualization of a region inherently fraught with potential disaster and helps to expand the indications for transsphenoidal resection of sellar region tumors in general and craniopharyngiomas in particular. Comparison of our results with those previously described is best left to others. Hopefully, however, these data will help guide the future development of endoscopic techniques. This series illustrates the ability to access complex suprasellar, parasellar, and retrosellar craniopharyngiomas via an EEA, regardless of the degree of sellar expansion. Difficulties with preventing postoperative CSF leakage are highlighted by this series. Together with standard approaches and other technological advances such as stereotactic cyst therapies and radiosurgery, the endoscopic EEA may prove a valuable weapon in the armamentarium against this truly challenging tumor. Expansion of the series, both in terms of number of patients and period of follow-up is critical for complete and accurate evaluation of the approach and technique. References 1. Chakrabarti I, Amar AP, Couldwell W, Weiss MH: Long-term neurological, visual, and endocrine outcomes following transnasal resection of craniopharyngioma. J Neurosurg 102: , Couldwell WT, Weiss MH, Rabb C, Liu JK, Apfelbaum RI, Fukushima T: Variations on the standard transsphenoidal approach to the sellar region, with emphasis on the extended approaches and parasellar approaches: surgical experience in 105 cases. Neurosurgery 55: , De Vile CJ, Grant DB, Kendall BE, Neville BG, Stanhope R, Watkins KE, et al: Management of childhood craniopharyngioma: can the morbidity of radical surgery be predicted? J Neurosurg 85: 73 81, Dhellemmes P, Vinchon M: Radical resection for craniopharyngiomas in children: surgical technique and clinical results. J Pediatr Endocrinol Metab 19 (1 Suppl): , Duff JM, Meyer FB, Ilstrup DM, Laws ER, Schleck CD, Scheithauer BW: Long-term outcomes for surgically resected craniopharyngiomas. Neurosurgery 46: , Fahlbusch R, Honegger J, Paulus W, Huk W, Buchfelder M: Surgical treatment of craniopharyngiomas: experience with 168 patients. J Neurosurg 90: , Frank G, Pasquini E, Doglietto F, Mazzatenta D, Sciarretta V, Farneti G, et al: The endoscopic extended transsphenoidal approach for craniopharyngiomas. Neurosurgery 59 (1 Suppl):ONS75 ONS83, Gupta DK, Ojha BK, Sarkar C, Mahapatra AK, Sharma BS, Mehta VS: Recurrence in pediatric craniopharyngiomas: analysis of clinical and histological features. Childs Nerv Syst 22:50 55, Hadad G, Bassagasteguy L, Carrau RL, Mataza JC, Kassam A, Snyderman CH, et al: A novel reconstructive technique following endoscopic expanded endonasal approaches: vascular pedicle nasoseptal flap. Laryngoscope 116: , Halves E, Bushe KA: Transsphenoidal operation on craniopharyngiomas with extrasellar extensions. The advantage of the operating endoscope. Acta Neurochir Suppl (Wien) 28:362, Han PP, Ponce FA, Spetzler RF: Intention-to-treat analysis of Spetzler-Martin grades IV and V arteriovenous malformations: 15

11 P. A. Gardner et al. natural history and treatment paradigm. J Neurosurg 98:3 7, Hoffman HJ, DeSilva M, Humphreys RP, Drake JM, Smith ML, Blaser SI: Aggressive management of craniopharyngiomas in children. J Neurosurg 76:47 52, Honegger J, Buchfelder M, Fahlbusch R: Surgical treatment of craniopharyngiomas: endocrinological results. J Neurosurg 90: , Honegger J, Buchfelder M, Fahlbusch R, Daubler B, Dorr HG: Transsphenoidal microsurgery for craniopharyngioma. Surg Neurol 37: , Im SH, Wang KC, Kim SK, Chung YN, Kim HS, Lee CH, et al: Transsphenoidal microsurgery for pediatric craniopharyngioma: special considerations regarding indications and method. Pediatr Neurosurg 39:97 103, Isaac MA, Hahn SS, Kim JA, Bogart JA, Chung CT: Management of craniopharyngioma. Cancer J 7: , Kaptain GJ, Vincent DA, Sheehan JP, Laws ER Jr: Transsphenoidal approaches for the extracapsular resection of midline suprasellar and anterior cranial base lesions. Neurosurgery 49: , Kassam AB, Carrau RL, Snyderman CH, Gardner P, Mintz A: Evolution of reconstructive techniques following endoscopic expanded endonasal approaches. Neurosurg Focus 19(1):E8, Kassam AB, Gardner PA, Snyderman CH, Carrau RL, Mintz AH, Prevedello DM: Expanded endonasal approach, a fully endoscopic transnasal approach for the resection of midline suprasellar craniopharyngiomas: a new classification based on the infundibulum. J Neurosurg 108: , Kouri JG, Chen MY, Watson JC, Oldfield EH: Resection of suprasellar tumors by using a modified transsphenoidal approach. Report of four cases. J Neurosurg 92: , Laws ER Jr: Comment on: Surgical treatment of craniopharyngiomas: an evaluation of the transsphenoidal and pterional approaches. Neurosurgery 36:724, Laws ER Jr: Transsphenoidal microsurgery in the management of craniopharyngioma. J Neurosurg 52: , Laws ER Jr: Transsphenoidal removal of craniopharyngioma. Pediatr Neurosurg 21:57 63, Maira G, Anile C, Rossi GF, Colosimo C: Surgical treatment of craniopharyngiomas: an evaluation of the transsphenoidal and pterional approaches. Neurosurgery 36: , Mason RB, Nieman LK, Doppman JL, Oldfield EH: Selective excision of adenomas originating in or extending into the pituitary stalk with preservation of function. J Neurosurg 87: , Matson DD, Crigler JF Jr: Radical treatment of craniopharyngioma. Ann Surg 152: , McMurry FG, Hardy RW, Dohn DF, Sadar E, Gardner J: Long term results in the management of craniopharyngiomas. Neurosurgery 1: , Minamida Y, Mikami T, Hashi K, Houkin K: Surgical management of the recurrence and regrowth of craniopharyngiomas. J Neurosurg 103: , Norris JS, Pavaresh M, Afshar F: Primary transsphenoidal microsurgery in the treatment of craniopharyngiomas. Br J Neurosurg 12: , Patterson RH: Comment on: Surgical treatment of craniopharyngiomas: an evaluation of the transsphenoidal and pterional approaches. Neurosurgery 36:724, Shirane R, Su C, Kusaka Y, Jokura H, Yoshimoto T: Surgical outcomes in 31 patients with craniopharyngiomas extending outside the suprasellar cistern: an evaluation of the frontobasal approach. J Neurosurg 96: , Symon L: An approach to radical excision of craniopharyngioma by the temporal route. Pediatr Neurosurg 21 (1 Suppl):64 68, Tena-Suck ML, Salinas-Lara C, Arce-Arellano RI, Rembao-Bojórquez D, Morales-Espinosa D, Sotelo J, et al: Clinico-pathological and immunohistochemical characteristics associated to recurrence/regrowth of craniopharyngiomas. Clin Neurol Neurosurg 108: , Van Effenterre R, Boch AL: Craniopharyngioma in adults and children: a study of 122 surgical cases. J Neurosurg 97:3 11, Weiner HL, Wisoff JH, Rosenberg ME, Kupersmith MJ, Cohen H, Zagzag D, et al: Craniopharyngiomas: a clinicopathological analysis of factors predictive of recurrence and functional outcome. Neurosurgery 35: , Yaşargil MG, Circic M, Kis M, Siegenthaler G, Teddy PJ, Roth P: Total removal of craniopharyngiomas: approaches and long-term results in 144 patients. J Neurosurg 73:3 11, 1990 Manuscript submitted October 26, Accepted June 6, Sources of support: none reported. Address correspondence to: Amin Kassam, M.D., Department of Neurological Surgery, Suite B-400, Presbyterian University Hospital, 200 Lothrop Street, Pittsburgh, Pennsylvania kassamab@upmc.edu. 16 J. Neurosurg. / Volume 109 / July 2008