J Neurosurg  Volume 123 • July 2015

cliNical articleJ Neurosurg 123:52–58, 2015

abbreviatioNs OS = overall survival; RFS = recurrence-free survival.submitted June 5, 2014.  accepted September 25, 2014iNclude wheN citiNg Published online March 27, 2015; DOI: 10.3171/2014.9.JNS141256.disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Extended orbital exenteration for sinonasal malignancy with orbital apex extension: surgical technique and clinical analysistakashi sugawara, md, phd,1 masaru aoyagi, md, phd,1 takahiro ogishima, md,1 Yoshihisa Kawano, md,1 masashi tamaki, md, phd,5 tomoyuki Yano, md,3 atsunobu tsunoda, md, phd,2 Kikuo ohno, md, phd,1 taketoshi maehara, md, phd,1 and seiji Kishimoto, md, phd2,4

Departments of 1Neurosurgery, 2Head and Neck Surgery, 3Plastic and Reconstructive Surgery, Tokyo Medical and Dental University, Tokyo; 4Department of Head and Neck Surgery, Kameda Medical Center, Chiba; and 5Department of Neurosurgery, Musashino Red Cross Hospital, Tokyo, Japan

obJect The majority of sinonasal malignancies present with advanced disease, and cure rates are generally poor. Surgical extirpation remains the mainstay of treatment. In cases of sinonasal malignancy with orbital apex extension, gross-total tumor resection requires orbital exenteration and bony skull base resection around the orbital apex to provide sufficient margins. In this retrospective study, the authors describe their surgical strategy in and technique for orbital ex-enteration with orbital apex resection in patients at Tokyo Medical and Dental University who had sinonasal malignancy with orbital apex extension. They also analyzed the clinical features of and the results in these patients.methods Between February 2001 and August 2012 at the authors’ institution, sinonasal malignancy with orbital apex extension was treated using craniofacial tumor resection with orbital exenteration including skull base bone around the orbital apex. The authors describe this technique and analyze the surgical indications, extent of resection, primary tumor location, outcome, pathological findings, and neoadjuvant and adjuvant therapies of the patients who underwent the technique.results The patients consisted of 12 men and 3 women with a mean age of 47.7 years (range 14–79 years). The longest postoperative follow-up was 9.5 years, and the shortest was 0.67 year (mean 3.0 years). Tumor originated at the ethmoid sinus in 6 patients (40%), maxillary sinus in 5 (33%), nasal cavity in 2 (13%), and orbital cavity and maxillary bone in 1 patient each (7%). Histological analysis of tumor specimens revealed squamous cell carcinoma in 9 patients (60%), rhabdomyosarcoma in 2 (13%), and small cell carcinoma, mucoepidermoid carcinoma, adenoid cystic carcinoma, and Ewing sarcoma in 1 patient each (7%). Two patients experienced recurrences at 1 and 5 months after treatment; these patients died at 5 and 10 months after surgery, respectively. Estimated 5-year recurrence-free survival (RFS) was 86.7%, and estimated 5-year overall survival (OS) was 86.2%; there was no perioperative mortality. None of the patients had new neurological deficits as a result of the surgery, but 5 patients suffered infectious complications from the graft transplanted into the cavity after resection. There were no other perioperative complications.coNclusioNs These authors are the first to describe a technique for extended orbital exenteration with orbital apex skull base resection. The technique provided sufficient margins for gross-total resection of the sinonasal malignancy with orbital apex extension. The estimated 5-year OS and RFS rates were high, and the perioperative complication rate was acceptably low, demonstrating the safety and efficacy of this technique.http://thejns.org/doi/abs/10.3171/2014.9.JNS141256KeY words craniofacial resection; orbital exenteration; sinonasal malignancies; surgical technique

52 ©AANS, 2015

extended orbital exenteration for sinonasal malignancy

The majority of sinonasal malignancies present with advanced disease, and cure rates are generally poor because early diagnosis is difficult. Surgical extir-

pation is the mainstay of treatment. Orbital involvement in sinonasal malignancies, especially of the orbital apex, is associated with a significant reduction in survival.2,10 In cases of sinonasal malignancy with orbital apex exten-sion, tumor resection with orbital exenteration and bone resection of the skull base around the orbital apex is re-quired to provide sufficient resection margins.

In this study, we describe our surgical strategy and technique for orbital exenteration with orbital apex resec-tion. We also analyze the clinical features of and the re-sults in patients who underwent this technique.

methodsAll patients in this case series provided informed con-

sent for inclusion of their clinical data in this paper. Be-tween February 2001 and August 2012 at our institution, 65 patients with sinonasal malignancy underwent cranio-facial resection performed by a multispecialty skull base team consisting of neurosurgeons, head and neck surgeons, and plastic surgeons. Of these 65 patients, 15 had sinonasal malignancy with orbital apex extension. These patients un-derwent craniofacial tumor resection with orbital exentera-tion and bony resection of the orbital apex. We reviewed the surgical indications, extent of resection, primary tumor location, outcome, pathological findings, and neoadjuvant and adjuvant therapies for these 15 patients.

Computed tomography and MRI were performed pre-operatively in all patients to determine tumor size, tumor location, and relationship of each tumor to adjacent tis-sues.

surgical strategy Indications for attempted gross-total resection of these

lesions were as follows: no metastasis to other organs, in-vasion of the intraorbital tissue around the orbital apex, and absence of invasion into the cavernous sinus or dura mater. A thorough evaluation of preoperative imaging was required to determine if orbital apex resection was required. This cancer invasion stage corresponds to the T4bN0M0 stage (International Union Against Cancer Staging System, 7th edition) without dura mater invasion.

surgical techniqueA lumbar spinal drain was inserted for CSF drainage

after general anesthesia had been induced.

Step 1: Observing From the Nasal CavityIntranasal endoscopic inspection of the tumor was first

performed to visually determine the extent of nasal cavity invasion and to visualize the resection borders anticipated on review of CT and MRI (Fig. 1).

Step 2: Craniotomy and Exposing the Frontal FossaA semi-coronal skin incision, bifrontal craniotomy,

and temporal craniotomy on the affected side were made, and the supraorbital bar was removed. Modifications of

the skin incision and supraorbital bar removal were done based on the variation in tumor location and size. The anterior cranial fossa was exposed to the end of the ju-gum sphenoidale, exposing the optic sheath bilaterally (Fig. 2A). Opening the dura and subsequent sacrifice of the bilateral olfactory nerves were performed to achieve adequate exposure of the frontal cranial fossa to the pos-terior edge of the jugum sphenoidale. The dural defect was repaired with temporal fascia as described previously.1,9

Step 3: Exposing the Middle FossaDolenc’s approach was first described in 1985;3 this

technique is used for exposure of the optic nerve, inter-nal carotid artery, and ophthalmic artery. The dura of the middle fossa was peeled from the skull base to expose the middle meningeal artery. This artery was coagulated and cut at the foramen spinosum. The superior orbital fissure and optic canal were unroofed with a high-speed drill and micro-rongeur. Epidural dissection of the lateral wall of the cavernous sinus was started by dissection at the menin-go-orbital band to the third branch of the trigeminal nerve. The bleeding was controlled by gently packing with oxi-dized cellulose (Surgicel, Ethicon) soaked in fibrin glue. The dural dissection was extended to the anterior clinoid process. The anterior clinoid process was then removed to expose the optic sheath widely (Figs. 2B and C and 3).

The anterior wall of the foramen rotundum was drilled, and the second branch of the trigeminal nerve was ex-posed to prepare for ligation and transection of the nerve together with the dura mater. The falciform ligament was cut, and the optic sheath was longitudinally opened to expose the internal carotid artery and ophthalmic artery (Fig. 4A). The optic nerve was then cut, followed by li-gation and cutting of the ophthalmic artery (Fig. 4B and C). The oculomotor nerve, trochlear nerve, first branch of the trigeminal nerve, and abducens nerve were transected together with the dura mater at the transition between the dura mater and the periorbita near the superior orbital fis-sure while taking care not to damage the internal carotid artery (Fig. 5A). After this transection of cranial nerves, the dura mater of the middle fossa floor is peeled further posteriorly to expose the superior and lateral wall of the

Fig. 1. a: Sagittal view of 3D CT bone image with tumor (green). Red line signifies the resection line.  b: Coronal view of the T2-weighted MR image. c: Endoscopic view in the left nasal cavity. T = tumor. Figure is available in color online only.

J Neurosurg  Volume 123 • July 2015 53

t. sugawara et al.

sphenoid sinus (Fig. 5B and C). The opened dura mater around the optic sheath was tightly repaired with the tem-poral fascia before opening the sphenoid sinus. Illustration of this widely exposed middle fossa is depicted in Fig. 6

Step 4: Resecting the Tumor With a Safety MarginThe line of resection at the frontal and middle fossa

skull bases included more than 5 mm of margin and was cut using a high-speed drill (Fig. 7A and B). In cases of sphenoid sinus tumor invasion, the posterior border in-cluded the foramen ovale so that the resection would be posterior enough to include the sphenoid sinus (Fig. 8A1–A3). In such cases, the root of the pterygoid process be-tween the foramina rotundum and ovale was drilled off to create a corridor to the posterior part of the sphenoid sinus. It was sometimes difficult to determine adequate margins from the frontal and middle fossae view. In these cases, endoscopic lighting from the nasal cavity transmit-ting through the ethmoid and sphenoid sinus wall served as a guide (Fig. 7B and D). If tumor did not invade the sphenoid sinus, the lateral wall of the sphenoid sinus was drilled off at the foramen rotundum to set the posterior boundary for excision in the anterior part of the sphenoid sinus (Fig. 8B1–B3). In our series, we defined ample mar-gins at the skull base as a sufficient amount of tissue left surrounding tumor to prevent exposure of the tumor dur-ing resection.

A palpebral conjunctiva incision was made (Fig. 9A), and the facial skin and subcutaneous tissue were peeled from the maxillary bone in a downward direction. The maxillary bone was cut using a bone saw without exposing the tumor (Fig. 9B and C). The tumor along with orbital content was then resected en bloc with a margin (Fig. 9D).

Step 5: Reconstructing the Defect After Tumor Resection and Making the Ocular Prosthesis Bed

Anterolateral thigh free flap or rectus abdominis myo-cutaneous free flap was harvested and transplanted to the cavity after resection (Fig. 10B). If feasible, the ocular prosthesis bed was made at the same time (Fig. 10C).

statistical methodsEstimated overall survival (OS) rate and recurrence-

free survival (RFS) rate were calculated using the Kaplan-Meier method. “Recurrence free” was defined as no evi-

Fig. 2. a: Exposure of the frontal fossa to the end of the jugum sphenoi-dale. b: Superior orbital fissure (SOF) and optic canal were unroofed, and the meningo-orbital band (MOB) and optic sheath were exposed. The anterior clinoid process (arrow) was removed.  c: The middle fossa was exposed to the lateral wall of the cavernous sinus. N.II = optic nerve sheath; V1 = ophthalmic branch of trigeminal nerve; V2 = maxillary branch of trigeminal nerve; V3 = mandibular branch of trigeminal nerve. Figure is available in color online only.

Fig. 3. Illustration of the exposed lateral wall of cavernous sinus and op-tic sheath. ON = optic nerve. Copyright Davinci Medical Illustration Of-fice. Published with permission. Figure is available in color online only. 

Fig. 4. a: Optic sheath and distal dural ring were opened, and the inter-nal carotid artery (IC), optic nerve (ON), and ophthalmic artery (arrow) were exposed.  b: The ON was cut.  c: The ophthalmic artery was cut. Figure is available in color online only.

J Neurosurg  Volume 123 • July 201554

extended orbital exenteration for sinonasal malignancy

dence of tumor recurrence on imaging studies such as CT, MRI, or FDG PET.

resultsA summary of characteristics for 15 patients is listed

in Table 1. The patients consisted of 12 men and 3 wom-en, with a mean age 47.7 years (range 14–79 years). The longest postoperative follow-up was 9.5 years, the short-est was 0.67 year, and the mean was 3.0 years. Extended orbital exenteration surgery was the first procedure for tu-

mor resection in 11 patients. In the 4 remaining patients, the surgery indexed for this report was the second sur-gery for 1 patient and the third, the fourth, and the fifth surgery for 1 patient each. The pathology of the 2 lesions that required fourth and fifth reoperations was rhabdo-myosarcoma. Ethmoid sinus was the most common pri-mary tumor location. Tumor originated from the ethmoid sinus in 6 (40%) of 15 patients, maxillary sinus in 5 (33%), nasal cavity in 2 (13%), and orbital cavity and maxillary bone in 1 patient each (7%). Histological analysis of tu-mor specimens revealed squamous cell carcinoma in 9 patients (60%), rhabdomyosarcoma in 2 (13%), and small cell carcinoma, mucoepidermoid carcinoma, adenoid cys-tic carcinoma, and Ewing sarcoma pathologies in 1 patient each (7%). Thirteen patients received neoadjuvant chemo-therapy; 2 patients did not, and their pathologies consisted of mucoepidermoid carcinoma and adenoid cystic carci-noma. Patients with squamous cell carcinoma and rhab-domyosarcoma (11 patients total) underwent preoperative radiation therapy, and the remaining patients with differ-ent pathologies (4) did not receive preoperative radiation.

Only 2 patients (Cases 2 and 8) experienced early re-currence at 5 and 1 months; these patients died at 10 and 5 months after surgery, respectively. Case 2 was a 79-year-old man with small cell carcinoma, and Case 8 was a 60-year-old man with squamous cell carcinoma. The patient in Case 2 had tumor invasion of the dura mater and brain that was observed intraoperatively but was not seen on preoperative CT or MRI studies. In this series of advanced-stage patients, estimated 5-year RFS was 86.7% and estimated 5-year OS was 86.2% (Fig. 11). In addition, there was no perioperative mortality. All of the patients without recurrence at 8 months remained disease free. Univariate analysis of factors, including age (p = 0.11), histology (p = 0.20), extent of resection (with or without the foramen ovale; p = 0.20), and intracranial invasion (p = 0.37) as the outcome predictors, revealed no significance.

Five patients suffered infectious complication around the graft transplanted to the cavity after resection. This infection originated from the ocular prosthesis bed in all 5 of these patients. This infection resolved with irrigation of

Fig. 5. a: The maxillary branch of the trigeminal nerve (V2) was cut, and the dura mater was cut at the transition between the dura mater and periorbita together with the cranial nerves of oculomotor nerve, trochlear nerve, V1, and abducens nerve.  b: Dura mater was peeled further posteriorly.  c: Middle fossa was totally exposed, and the superior and lateral wall of the sphenoid sinus and the tuberculum sellae (arrow) were well visualized. Figure is available in color online only.

Fig. 6. Illustration of widely exposed middle fossa after cutting of V1, V2, and ON. Gray line signifies the line of resection. Copyright Davinci Medical Illustration Office. Published with permission. Figure is available in color online only. 

Fig. 7. a and b: Resection line (green line) at the frontal fossa. c and d: Resection line (green line) at the frontal and middle fossae. Panels B and D showed endoscopic lighting from the nasal cavity transmitting through the sphenoid sinus wall was the landmark for deciding the re-section line. Arrow indicates the tuberculum sellae. Figure is available in color online only.

J Neurosurg  Volume 123 • July 2015 55

t. sugawara et al.

the abscess in 3 patients, and 1 patient had a small incision made at the infection site prior to irrigation. One patient needed to undergo removal of the infected skull bone flap and irrigation of the abscess 1 month after surgery. No patient suffered neurological or any other complication at-tributable to the surgical procedure.

discussionCancers of the nasal and paranasal sinuses are rare

and reported to be only 3%–5% of all head-and-neck can-cers.2,4 The asymptomatic growth of these tumors into the air-filled nasal and ethmoidal sinus spaces makes their early diagnosis difficult. Hence, many patients are admit-ted to the hospital with an advanced stage of the disease. Craniofacial surgery for this type of sinonasal malignan-cy is difficult to perform adequately and requires a team consisting of neurosurgeons, head and neck surgeons, and plastic surgeons.

Five-year OS for patients with sinonasal malignancy who have undergone anterior craniofacial resection has been reported to range between 40% and 58%.1,2,5–8,10 Five-year RFS has been reported as 24.4%–52.8%.1,2,5,7 The perioperative death rate has been reported as 3.6%–4.7%.2,7–9 These data indicate that anterior craniofacial surgery is still a challenging operation.

Suarez et al. reported that survival for patients with sinonasal tumors treated using craniofacial resection was 40% at 5 years and that the clinical outcome for these pa-tients with Stage T2 and T4 disease was almost the same.10 These authors also showed that 5-year survival was sig-nificantly affected by tumor histological findings, with 5-year survival rates of 71% in patients with esthesioneu-roblastomas, 65% in those with squamous cell carcinoma, 31% in those with adenocarcinoma, 17% in those with

undifferentiated carcinoma, and 0% in those with mela-noma. Patel et al. showed that the histology of the primary tumor, the extent of intracranial extension, and the status of the surgical margins were significant independent pre-dictors of RFS.7

The poor prognosis associated with malignant tumors of the paranasal sinuses is mainly a consequence of local recurrences in the skull base. Some reports have shown that orbital involvement significantly affects survival, par-ticularly if the orbital apex is involved.1,7,10 These studies have suggested that orbital apex involvement positively correlates with higher recurrence rates and shorter surviv-al. Therefore, we believe that en bloc resection with mar-gins, especially at the skull base around the orbital apex, is necessary to avoid local recurrences and contributed to our high estimated OS and RFS rates. Suarez also re-ported that involvement of the lateral wall of the sphenoid

Fig. 8. Case 12. Axial T1-weighted Gd-enhanced MR images (a1 and a2) showing tumor invasion of the sphenoid sinus. The resection line was made to include the foramen ovale (a3). Case 11. Axial T1-weighted Gd-enhanced MR images (b1 and b2) showing no tumor invasion of the sphenoid sinus. The resection line was made just anterior to the fora-men ovale (b3). Figure is available in color online only.

Fig. 9. a: Palpebral conjunctiva incision was made.  b: Facial skin and subcutaneous tissue were peeled off, and the lateral resection line of the maxillary bone was cut using a bone saw.  c: Medial side of the resec-tion line was cut using a bone saw and flat chisel.  d: Tumor with orbital content was resected en bloc without tumor exposure. Figure is avail-able in color online only.

Fig. 10. a: Defect after tumor resection below the frontal fossa.  b: Anterolateral thigh free flap (asterisk) was transplanted to the defect after resection. c: The ocular prosthesis bed was made (arrow). Figure is available in color online only.

J Neurosurg  Volume 123 • July 201556

extended orbital exenteration for sinonasal malignancy

sinus was one of the factors affecting survival;10 therefore, it is also important to make a sufficient resection margin at the deepest part in the sphenoid sinus to prevent recur-rence, as we did in this study.

In our series, the estimated 5-year OS for patients with sinonasal malignancy who had undergone extended or-bital exenteration was 86.2%, and estimated 5-year RFS was 86.7%. These survival rates are better than those in previous reports.2,6–8,10 Furthermore, our 5-year OS rate was better than the 70.3% 5-year OS in the 50 sinonasal malignancy patients (mean age 49.1 years) with no inva-sion of the orbital apex who also underwent craniofacial resection in the same time period at our institution. The surgical strategy for patients without orbital apex invasion was as follows: When the tumor did not invade orbital bone, the periorbita and orbital contents were preserved, and the medial side of orbital bone was resected with the tumor. When the tumor invaded orbital bone, the orbital contents, excluding the orbital apex, were resected with the tumor. Histological analysis of specimens from these 50 patients without orbital apex resection revealed olfac-tory neuroblastoma in 15 patients (30%), squamous cell carcinoma in 13 (26%), adenoid cystic carcinoma in 6 (12%), rhabdomyosarcoma in 4 (8%), adenocarcinoma in 2 (4%), osteosarcoma in 2 (4%), and other pathologies in 8 (16%). Patients with olfactory neuroblastoma have better outcomes than patients with other sinonasal malignancies. When we excluded the 15 patients with olfactory neuro-blastoma from these 50 patients, 5-year OS decreased to 58.7%. Nonetheless, this rate is still better than previously reported rates but worse than the rate for our 15 cases with orbital apex tumor extension who had undergone extended orbital exenteration. Although these results were initially surprising, we believe that our orbital apex resection al-lowed for more complete resection of tumor from the sphenoid sinus. Tumor resection from the sphenoid sinus is one of the most technically difficult parts of the excision in most cases but is also one of the most crucial steps in preventing recurrence. Total tumor resection with ample safety margins became feasible using our surgical tech-nique of extended orbital apex resection, providing wide exposure of the posterior part of the sphenoid sinus. There was no perioperative mortality in our patient series, and the postoperative infectious complication rate was 33.3%. All of these infections involved the lacrimal gland around the ocular prosthesis bed. However, these infections were easily controlled in most cases. Irrigation alone cleared the infection in 4 patients, and only 1 patient required an addi-tional operation such as removal of the infected bone flap. There were no neurological and systemic complications. This perioperative complication rate is within an accept-able limit for this type of surgery. There were no factors that independently predicted outcome. However, the num-ber of patients was too small (only 2 of 15 patients died) to draw any unequivocal conclusions.

A limitation of this study is the short mean follow-up period of 3.0 years while calculating the estimated 5-year OS and RFS rates. However, 5 patients are alive and have been recurrence free for more than 50 months after tumor resection. The OS and RFS were not decreased after a few years postresection, as shown in Fig. 11. The use of histori-ta

ble

1. su

mm

ary o

f clin

ical

char

acte

ristic

s in

15 p

atie

nts w

ith si

nona

sal m

align

ancy

Case 

No.

Age (

yrs),

Se

xPrimary T

umor Site

Histo

logy

Prior

Trea

tmen

tAd

juvant T

herapy

Ocula

r Prosth

esis Be

d/Oc

ular P

rosth

esis

Comp

lication

Recurre

nce/

Outco

meFo

llow-Up

(mos)

158, M

Ethm

oid si

nus

Muc

oepid

ermo

id ca

rcino

ma

—CT

+RT (50 G

y)+/−

Infection

−/alive

114

279, M

Ethm

oid si

nus

Small

cell c

arcin

oma

CTCT

−/−—

+/dead

 10

334, F

Ethm

oid si

nus

Squa

mous

cell c

arcin

oma

SR+C

T+RT

 (35 G

y)—

+/+Infection

−/alive

 80

 477, M

Maxilla

ry sinus

Squa

mous

cell c

arcin

oma

CT+R

T (50 G

y)—

−/−Infection

−/alive

 35

 560, M

Maxilla

ry sinus

Squa

mous

cell c

arcin

oma

CT+R

T (50 G

y)—

−/−—

−/alive

 58

 621, M

Orbital cavity

Rhabdomy

osarcoma

SR×4

+CT+RT

 (48 G

y)CT

+/+—

−/alive

 53

745, M

Ethm

oid si

nus

Squa

mous

cell c

arcin

oma

CT+R

T (40 G

y)—

+/+—

−/alive

 54

 860, M

Nasal cavity

Squa

mous

cell c

arcin

oma

CT+R

T (65 G

y)CT

−/−—

+/dead

  5

 950, M

Ethm

oid si

nus

Adenoid

 cystic c

arcin

oma

SR×2

RT (5

0 Gy)

+/+—

−/alive

 34

1014, F

Nasal cavity

Rhabdomy

osarcoma

SR×3

+CT+RT

 (50 G

y)—

+/−—

−/alive

  8

1116, M

Maxilla

ry bo

neEw

ing sa

rcom

aCT

CT+R

T (56 G

y)+/−

—−/a

live

22

1231, F

Maxilla

ry sinus

Squa

mous

cell c

arcin

oma

CT+R

T (50 G

y)—

+/−Infection

−/alive

 26

1355, M

Maxilla

ry sinus

Squa

mous

cell c

arcin

oma

CT+R

T (50 G

y)—

+/+—

−/alive

 18

1462, M

Maxilla

ry sinus

Squa

mous

cell c

arcin

oma

CT+R

T (40 G

y)—

−/−—

−/alive

 20

1554, M

Ethm

oid si

nus

Squa

mous

cell c

arcin

oma

CT+R

T (70 G

y)—

+/−Infection

−/alive

17

CT = ch

emotherapy; R

T = radia

tion therapy; S

R = surgery; − = no; + = ye

s; — = no

t applicable.

J Neurosurg  Volume 123 • July 2015 57

t. sugawara et al.

cal controls to compare outcome is also a limitation of this study. We intend to accumulate more cases and monitor them for a longer period.

conclusionsWe are the first to describe the surgical strategy and

procedures for extended orbital exenteration with skull base bone resection around the orbital apex having a suf-ficient resection margin for sinonasal malignancy with or-bital apex extension. The estimated 5-year OS and RFS were high, and the perioperative complication rate was acceptably low because it was possible to make sufficient resection margins using this procedure. We plan further characterization of the effectiveness of this technique with longer patient follow-ups and the accumulation of more case experience.

acknowledgmentWe thank Robert Ayer, MD (Department of Neurosurgery, Uni-

versity of California Irvine, Orange, California), for assisting with this publication.

references 1. Cantù G, Solero CL, Mariani L, Salvatori P, Mattavelli F,

Pizzi N, et al: Anterior craniofacial resection for malignant ethmoid tumors—a series of 91 patients. Head Neck 21:185–191, 1999

2. Cantu G, Solero CL, Miceli R, Mattana F, Riccio S, Colombo S, et al: Anterior craniofacial resection for malignant para-nasal tumors: a monoinstitutional experience of 366 cases. Head Neck 34:78–87, 2012

3. Dolenc VV: A combined epi- and subdural direct approach to carotid-ophthalmic artery aneurysms. J Neurosurg 62:667–672, 1985

4. Fukumitsu N, Okumura T, Mizumoto M, Oshiro Y, Hashi-moto T, Kanemoto A, et al: Outcome of T4 (International Union Against Cancer Staging System, 7th edition) or recur-rent nasal cavity and paranasal sinus carcinoma treated with proton beam. Int J Radiat Oncol Biol Phys 83:704–711, 2012

5. Ganly I, Patel SG, Singh B, Kraus DH, Bridger PG, Cantu G, et al: Craniofacial resection for malignant paranasal sinus tumors: report of an International Collaborative Study. Head Neck 27:575–584, 2005

6. McCaffrey TV, Olsen KD, Yohanan JM, Lewis JE, Ebersold MJ, Piepgras DG: Factors affecting survival of patients with tumors of the anterior skull base. Laryngoscope 104:940–945, 1994

7. Patel SG, Singh B, Polluri A, Bridger PG, Cantu G, Chees-man AD, et al: Craniofacial surgery for malignant skull base tumors: report of an international collaborative study. Can-cer 98:1179–1187, 2003

8. Shah JP, Kraus DH, Bilsky MH, Gutin PH, Harrison LH, Strong EW: Craniofacial resection for malignant tumors involving the anterior skull base. Arch Otolaryngol Head Neck Surg 123:1312–1317, 1997

9. Solero CL, DiMeco F, Sampath P, Mattavelli F, Pizzi N, Salvatori P, et al: Combined anterior craniofacial resection for tumors involving the cribriform plate: early postoperative complications and technical considerations. Neurosurgery 47:1296–1305, 2000

10. Suarez C, Llorente JL, Fernandez De Leon R, Maseda E, Lopez A: Prognostic factors in sinonasal tumors involving the anterior skull base. Head Neck 26:136–144, 2004

author contributionsConception and design: Sugawara, Aoyagi, Tamaki, Ohno, Maehara, Kishimoto. Acquisition of data: Sugawara, Ogishima, Kawano, Tamaki, Yano, Tsunoda. Analysis and interpretation of data: Sugawara. Drafting the article: Sugawara. Critically revis-ing the article: Sugawara, Aoyagi, Ohno, Maehara, Kishimoto. Reviewed submitted version of manuscript: Aoyagi, Ohno, Kishimoto. Statistical analysis: Sugawara. Administrative/techni-cal/material support: Aoyagi, Tsunoda, Kishimoto. Study supervi-sion: Sugawara, Aoyagi, Ohno, Maehara, Kishimoto.

correspondenceTakashi Sugawara, Department of Neurosurgery, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo, 113-8519 Japan. email: sugawara.nsrg@tmd.ac.jp.

Fig. 11. Kaplan-Meier curves of RFS (left) and OS (right).

J Neurosurg  Volume 123 • July 201558