- Email: [email protected]
Subcranial fronto-orbito-nasal approach for ethmoidal cancers
Subcranial Fronto-Orbito-Nasal Approach for Ethmoidal Cancers SURGICAL TECHNIQUES AND RESULTS F.X. Roux, M.D.,* R. Moussa, M.D.,* B. Devaux, M.D.,* F. Nataf, M.D.,* P. Page, M.D.,* O. Laccourreye, M.D.,† G. Schwaab, M.D.,‡ D. Brasnu, M.D.,† and J. Lacau Saint-Guily, M.D.§ *Department of Neurosurgery, Hopital Sainte-Anne, Paris, France; †Department of ENT and Cervicofacial Surgery, Hopital Laennec, Paris, France; ‡Department of ENT and Cervicofacial Surgery, Institut Gustave Roussy, Villejuif, France; §Department of ENT and Cervicofacial Surgery, Hopital Tenon, Paris, France
Roux FX, Moussa R, Devaux B, Nataf F, Page P, Laccourreye O, Schwaab G, Brasnu D, Lacau Saint-Guily J. Subcranial frontoorbito-nasal approach for ethmoidal cancers: surgical techniques and results. Surg Neurol 1999;52:501–10. BACKGROUND
The authors report their experience with the subfrontoorbito-nasal approach (SFON) for the treatment of 30 patients suffering from ethmoidal cancers over the past 3 years. The advantages and pitfalls of this technique are described and compared with other classic approaches. METHODS
Among 156 patients suffering from ethmoidal cancers and treated between January 1984 and January 1998, 30 patients were operated on using the SFON approach during the past 3 years. There were 27 males and 3 females, ranging in age from 15 to 77 years. Histologic composition of the lesions was as follows: 15 adenocarcinomas, 6 esthesioneuroblastomas, 3 melanomas, 2 epidermoid carcinomas, 1 nondifferentiated carcinoma, 1 neuroendocrine carcinoma, 1 villous carcinoma, and 1 cystic adenoid carcinoma (cylindroma). According to the authors’ classification, 7% were T1, 6% T2, 22% T3, 38.5% T4a, and 26.5% T4b. All patients were operated on through a SFON approach, followed by removal of the tumor and reconstruction of the skull base with a pericranial flap. RESULTS
Since the mean follow-up was of short duration (12 months, ranging from 3 to 29 months), significant carcinologic results could not be obtained. However, a detailed analysis of the surgical procedure was performed. No patient died or had major complications related to the SFON approach. One cerebrospinal fluid (CSF) fistula and four oculomotricity dysfunctions were observed. Definitive anosmia was reported in all cases.
Address reprint requests to: F. X. Roux, Service de Neurochirurgie, Hopital Sainte-Anne, 1 Rue Cabanis 75014 Paris. Received July 16, 1998; accepted December 7, 1998. © 1999 by Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010
CONCLUSION
The advantages of the procedure include a wide exposure of the anterior skull base through a limited approach, the possibility of modifying the approach according to the size and location of the lesion, total resection of tumors, simplified skull base reconstruction technique, and reduction of postoperative confusion and hospital stay. © 1999 by Elsevier Science Inc. KEY WORDS
Subfronto-orbito-nasal approach, skull base surgery, ethmoidal cancers.
rom 1984 to 1995, we have been using a combined transcranial transfacial approach to remove ethmoidal or ethmoidosphenoidal cancers [21–24]. We believed that it would be reasonable to modify our technique and choose a more limited surgical route to minimize postoperative morbidity as much as possible. Our aim was to avoid, or at least decrease, the frequency of incidents or complications such as: 1) Dural tears, which can be significant when the dura is adherent to the inner table of the frontal skull, particularly in patients over 60 years old. Such dural tears can be responsible for postoperative cerebrospinal fluid leakage, and therefore require watertight reconstruction or reinforcement, which is more difficult to perform bedcause they are larger; 2) Immediate postoperative confusion, mainly in elderly patients, attributable to frontal lobe retraction during a subfrontal transcranial approach; 3) Retractile transfacial scar such as can be seen after radiotherapy when the patient has undergone a combined surgical approach. That is why we began using a subcranial
F
0090-3019/99/$–see front matter PII S0090-3019(99)00127-5
502 Surg Neurol 1999;52:501–10
1
Roux et al
Histological Data of Ethmoidal Cancers Operated Via the SFON Approach
NUMBER OF PERCENTAGE CASES (%) Adenocarcinoma Esthesioneuroblastoma Melanoma Epidermoid carcinoma Cylindroma (adenoid cystic) Neuroendocrine tumor Villous tumor Nondifferentiated tumor Total
15 6 3 2 1 1 1 1 30
50 20 10 8 3 3 3 3 100
approach inspired by the procedure described by Raveh in 1978 [17–18]. In this report, we describe our experience with the subcranial-fronto-orbito-nasal approach (SFON) as used for the treatment of 30 ethmoidal cancers over the past 3 years. We shall focus on the surgical data and postoperative results and compare them with those obtained with combined transcranial transfacial procedures for equivalent indications. We exclude carcinologic results that have been presented recently in other issues.
Materials and Methods PATIENT POPULATION From January 1984 to January 1998, 221 patients were referred to our institution for anterior cranial base procedures (excluding pituitary surgery). Among them, 156 patients suffered from ethmoidal cancer: 106 were operated through a combined transcranial and transfacial approach, 30 through an extended SFON route, 20 were not operated. There were 27 males and 3 females who underwent a SFON procedure. Patients’ ages ranged from 15 to 77 years, with a mean age of 57.5; 13 patients were over 65, and seven were over 70. Table 1 shows histologic composition and location of the lesions treated by the SFON approach. PREOPERATIVE GRADING Patients were evaluated by a team including a head and neck surgeon, anesthesiologist, and neurosurgeon. Computed tomography (CT) scans with axial and coronal views and magnetic resonance imaging (MRI) were performed in 29 and 28 patients, respectively. Radiological evaluation revealed a tumoral extension. Such grading is necessary to confirm surgical indication and determine the type of approach. Because of the absence of any international clas-
sification for ethmoidal cancers, we have been using a personal grading scale for the last 12 years [22]: T1: the tumor is developed in only one ethmoidal region, T2: the tumor is developed in at least two regions of the ethmoid sinuses with respect to the walls, T3: the tumor destroys any ethmoid cell wall, except the cribriform plates, T4a: the tumor destroys the cribriform plate(s) without intracranial extension or dural invasion, T4b: tumor with intracranial extension whatever its volume. In the present series, 7% of tumors were T1, 6% T2, 22% T3, 38.5% T4a, and 26.5% T4b. PREOPERATIVE MANAGEMENT Five of these 30 patients were referred to our team with a tumor recurrence: three had undergone a previous transfacial operation (1 epidermoid, 1 esthesioneuroma, 1 villous tumor); one had been treated by chemotherapy and radiotherapy with no surgery for an undifferentiated carcinoma; and one had transfacial surgery followed by radiotherapy for an adenocarcinoma. In all circumstances, histologic diagnosis was confirmed or reconfirmed by endonasal biopsy performed in the ENT department before combined surgery. An inductive chemotherapy was then started, consisting of 4-day continuous infusion of cisplatin (25 mg/m2/day) and 5-fluorouracyl (1 g/m2/day). An average of three cycles were administered at 3-week intervals. Eighteen patients went through chemotherapy: 11 adenocarcinomas, 4 esthesioneuromas, 1 epidermoid carcinoma, 1 melanoma, and 1 neuroendocrine carcinoma. Twelve patients had no chemotherapy either because they were harboring a tumor nonresponsive to chemotherapy, or because chemotherapy could not be used for some reason. Preoperative MRI evaluation was conducted 2 to 3 weeks after the last chemotherapy course, a few days before surgery. It was then possible to assess the quality of the response to chemotherapy [21, 23]. The operation was then performed, 3 to 4 weeks after the last chemotherapy course, by the neurosurgical team in collaboration with the head and neck surgeons. OPERATIVE TECHNIQUE Via a bicoronal skin incision, the scalp flap separated from the pericranium is elevated. Care is taken to preserve the pericranium for use during reconstruction of the skull base. Dissection of a pedicled pericranial flap measuring 10 to 12 cm in length and 8 to 10 cm in width is performed (Figure 1). Once it has been elevated over the calvarium
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
Intraoperative photograph showing the pedicled pericranial flap to be used for skull base reconstruction.
1
and protected with wet gauze, subperiostal dissection is continued down to the periorbita, which is separated from the superior, medial and occasionally lateral walls of the orbit. The supraorbital nerves are preserved by freeing them from the supraorbital canals, maintaining continuity of the pericranium and periorbita. Care is taken to preserve the supraorbital arteries supplying the pericranial flap. Dissection is extended to expose the entire nasal pyramid down to the cartilage in the midline and laterally as far as the lacrymal crests. One or both lacrymal ducts may be cut obliquely to avoid or limit retraction and stenosis at their extremities. The osteotomy line is usually extended 3 to 4 cm cranially and 3 to 4 cm laterally, limited by the supraorbital canals, but may be extended deliberately according to tumor extension. One burr hole is then made at the midline and osteotomies are made across the frontal bone down to and along orbital roofs, down the medial orbital wall and along the nasomaxillary grooves just anterior to the lacrymal crest. The nasal pyramid is separated 1 to 5 mm above the insertion of the septal cartilage. A vertical osteotomy across the skull base just anterior to the crista galli, allows detachment of the frontonasal segment and exposure of the anterior skull base (Figure 2). The dura is carefully protected during the procedure. It is divided around the olfactory groove, severing the olfactory filaments. Then it is separated subcranially from the floor of the anterior fossa without frontal lobe retraction. If dura is resected because it was invaded, the defect is repaired either with pericranium or with synthetic dura. Broad access is provided to the nasal, maxillary, and sphenoethmoidal structures (Figure 3). With a high-speed drill, an electric
Surg Neurol 503 1999;52:501–10
saw, and/or chisels, cuts are made through the cribriform plates. The posterior limit of the resection is at the planum sphenoidale. The specimen is removed with the entire ethmoid block, including one or both lamina papyracea, the superior and middle turbinates on both sides, and the nasal septum. It must be stressed that removal of middle turbinates has been in a few patients a little more difficult to achieve than through a classic transfacial approach. The resection of the ipsilateral orbital floor was limited to a fourth or a third of its width. The integrity and vacuity of both maxillary sinuses were checked, as well as of the sphenoid sinus, which was systematically entered. Tumor removal was finally achieved as far as possible “en bloc”; in fact, it was more often performed in a piecemeal fashion because the lesion usually split into pieces (Figure 4). RECONSTRUCTIVE PROCEDURE The dura was closed meticulously in a watertight manner with stitches, reinforced with collagen glue. All dural lacerations were closed by primary suture; but, as noted previously, if necessary any dural defect was reconstructed or reinforced with a graft. We did not reconstruct the cranial base defect itself. We only used the pedicled pericranial flap, which was inserted deeply to the jugum sphenoidalis in two layers, fixed to the dura with sutures and reinforced with surgical glue. Such a pericranial flap is well vascularized by both supraorbital vessels and should be sufficient to ensure an efficient closure of the intracranial cavity. No canthopexy was performed. The bone flap was repositioned and secured either with microplates and screws (26 cases) or with nonabsorbable sutures (4 cases) (Figure 5). The burr hole(s) were filled in with bone dust. Four to six Merocel® tubes were left in the nasal cavities and removed on the third postoperative day. The bicoronal incision was closed in layers and drained with a suction drain. Perioperative lumbar drainage was not instituted. Antibiotic drugs (amoxicillin-clavulanic acid, Augmentin®) were given for 3 to 5 days postoperatively. POSTOPERATIVE COURSE The patient was allowed to walk on the first or second postoperative day. Subcutaneous (s.c.) drainage was removed on the second postoperative day. Endonasal drains were removed and antibiotic drugs withdrawn on the third postoperative day. Postoperative stay in the intensive care unit was limited to an average of 2 days and the patient was discharged from the hospital after a stay limited to 8 to 10 days. All patients had an MRI before leaving
504 Surg Neurol 1999;52:501–10
Roux et al
Intraoperative photograph (left) showing removal of the fronto nasal unit after the dura has been dissected and the osteotomy performed. The illustration (bottom right) shows the angle of view in the SFON approach (black arrow) compared with the transcranial approach (white arrow). The angle is almost doubled, reducing the need to retract frontal lobes.
2
the hospital confirming the extent of the tumor resection, which was total in all cases.
Results PATIENT FOLLOW-UP Patients were followed for a mean period of 12 months, ranging from 3 to 29 months. This follow-up is too short for significant carcinologic results, which is not the aim of this work, but does allow a detailed analysis and assessment of the surgical technique. MORTALITY One patient died postoperatively from acute brain swelling after removal of a metastatic lesion located in the left frontal lobe, which was performed during the same procedure done for his ethmoidal carcinoma. During the follow-up, one patient with an ethmoidal adenocarcinoma type T4b died 3 months after operation from multiple liver metastases.
POSTOPERATIVE COMPLICATIONS AND MORBIDITY No patient suffered from confusion in the immediate postoperative period, even elderly patients. One patient developed a CSF rhinorrhea that was noted after the removal of the endonasal packing. A ventriculoperitoneal shunt was placed 6 days later, and led to a quick and definitive stop of the leakage. In two other patients, a liquid discharge was observed after the removal of the nasal Merocel® meshes on the third postoperative day, but it was not as clear as CSF, and most probably was a mixture of water used to rinse the surgical cavity at the end of the procedure and exudation mucosities tinged with blood; nevertheless it was considered to be a possible CSF leak and was therefore managed with fluid restriction, acetazolamide, and daily lumbar punctures for 3 days. No cases of meningitis or deep suppuration were noted. Two s.c. infections were reported; they were cured easily with 5 days of antibiotic therapy.
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
3
Intraoperative view showing the wide exposure of nasal, maxillary, and sphenoidal structures.
Oculomotricity dysfunction was reported in four cases; it was caused by a slight desaxation of the eyeball, and associated with a moderate enophthalmia in three cases and with a watering eye in one case. All these patients displayed almost complete
4
Surg Neurol 505 1999;52:501–10
5
Intraoperative view showing the flap replaced and secured with miniplates.
resolution with time. Two other patients had watering eyes. Anosmia was, of course, observed in all patients. Cosmetic results were satisfactory in all cases except one patient who received preoperative radi-
CT scan, coronal views. Left: preoperative view showing the location and extension of an adenocarcinoma. Right: postoperative view with bone window showing total removal of the lesion.
506 Surg Neurol 1999;52:501–10
ation therapy, in whom miniplates were visible beneath the skin.
Discussion The anterior base of the skull can be approached by different routes, according to location, size, and histologic type of the lesions. For benign tumors and traumatic lesions, endocranial or subcranial approaches offer a broad operative visualization of the region. For malignant tumors, a combined transcranial transbasal approach may be needed for resection of paranasal tumors invading the skull base [4,8,13–15,20]. Dandy [2], and Ray and McLean [19] first described, in 1941 and 1943 respectively, a combined approach for resection of orbital tumors. Smith et al [28], Ketcham et al [14], and Van Buren et al [30] were among the first authors to report on a combined transcranial and transfacial resection of a paranasal sinus tumor, in 1953. In 1972, P. Derome [5] described the transbasal removal of sphenoethmoidal tumors. Since that time, modifications of these approaches have been undertaken by many head and neck surgeons and neurosurgeons around the world [8 –10,24]. The subcranial approach was first described by Raveh in 1978, primarily for the treatment of skull base trauma and craniofacial anomalies. Since 1980, he has extended the indications to tumor resection [17,18]. The advantages of this approach have been confirmed by Sekhar et al [24,25], Spetzler et al [29], Delfini et al [4], and Darrouzet et al [3]. These advantages are based on the surgical exposure and direction of operative approach, avoidance of frontal lobe retraction, avoidance of facial incision, en bloc resection of intra- and extradural tumors, and watertight skull base reconstruction. In our department, and in accordance with the ENT teams we work with, we consider that small and medium size tumors, i.e., T1 and T2 tumors, as long as they are at a minimal distance of 10 mm from the cribriform plates, should be operated through an only transfacial and/or endonasal route. All other tumors (T3, T4a, and T4b) justify a combined approach. Because we decided to operate ethmoidal cancers through a SFON approach, we reserve combined craniofacial approaches for paranasal lesions involving the infrastructure, the palate and/or the maxillary bone and sinus since in such situations the SFON exposure is too limited. Surgery is not recommended for patients harboring large carcinomas presenting with a voluminous intracerebral extension, bilateral orbital development, and/or posterior invasion toward the sphenoid and
Roux et al
clivus. These patients undergo a radiotherapy and/or chemotherapy regimen. Among different advantages of a SFON approach, performing a small subcranial osteotomy should reduce the risks of dural tears, and therefore should reduce the frequency of postoperative CSF leakage. Indeed, dural tears are less frequent and less extended with such a limited approach. This is particularly noted in elderly patients in whom the dura is often very adherent in the frontal area, particularly near the midline. Delfini [4] and Raveh et al [17,18] emphasized this fact, recommending the use of dural suspension and watertight closure. But it is always difficult to reach the goal of no leakage. Raveh recorded a 2.8% rate of CSF leakage [18]. In our series, we had one postoperative CSF rhinorrhea, corresponding to a 3.5% figure. It can be compared with the rate of rhinorrhea we recorded after combined procedures: 5 of 82 patients (6%). The difference between these figures is informative, but not significant. Concerning the treatment of small leaks, most authors recommend spinal fluid drainage or lumbar punctures. The SFON approach allows a direct visualization of the anterior skull base parallel to its plane [17]; consequently, because there is minimal or no retraction of the frontal lobes, it avoids or at least limits postoperative confusion and frontal syndrome as are sometimes encountered after a combined subcranial and transfacial approach, mainly in elderly patients. Indeed, none of our patients operated through this approach developed such symptoms, whereas 11 of 82 patients (13.5%) had a transient postoperative confusion after combined procedures. Therefore, this technique accelerates postoperative recovery, and reduces intensive care unit and hospital stay as well as time taken to return to work and independent life. The cost of treatment is consequently reduced, which nowadays is an important point to be considered [7]. The SFON approach does not seem to be responsible for more postoperative diplopia than combined surgery. Raveh recorded five oculomotor, trochlear or abducens nerve chronic dysfunctions [18], i.e,. 4.8%. We did not find any information concerning transient postoperative diplopia in his papers. Darrouzet reported two transient diplopias in four patients [3]. As for us, we noted four postoperative transient (13.3%) and one persistent (3.3%) diplopias in our 30 patients who had a SFON procedure. The figures were slightly different for the 82 patients we operated through a combined procedure: four transient (4.8%), and six persistent diplopias (7.3%). These functional results are comparable and almost identical, indeed the variations of
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
Surg Neurol 507 1999;52:501–10
figures are not significant. Raveh et al [18] and Darrouzet [3] performed canthal ligament fixation, we did not. Therefore, we think canthopexy is not really useful in most cases. The SFON approach to the anterior skull base allows en bloc removal of malignant tumors [17], because it offers a broad access to sphenoethmoidal structures as far as the jugum and clivus posteriorly. Nasal and paranasal cavities, including both maxillary sinuses, can be easily controlled, as well as orbital structures when necessary [1,22]. The ability to modulate the approach, depending on the tumor extension and surgical needs, must be emphasized. Osteotomy can be performed either through an enlarged route as described by Cophignon et al [1], or through a limited glabellar approach as described by Jho and Ko [11]. For instance, extensive intradural or brain involvement may require an enlargement of midfrontal osteotomy. An extended glabellar approach along the eyebrows, as reported by Osguthorpe [16] may also offer excellent exposure of the anterior skull base; it provides sufficient pericranial tissue for skull base reconstruction and subfrontal dura lining, and good cosmetic results. Furthermore, ethmoidosphenoidal tumors extending laterally may require an associated transfacial approach, which can be easily undertaken with a SFON; this is particularly the case when the tumor extends toward the orbit and/or the maxillary sinus. In two of our patients (Cases 4 and 15), we performed such a combined procedure to obtain complete resection of the tumor. Despite allowing a rather large access to the cranial base, the SFON approach of course cannot solve all problems. Its usual anatomic limits are: laterally, the optic nerves, cavernous sinuses, and carotid arteries; posteriorly, the clivus, basilar artery, and brain stem; caudally, the maxillary sinuses, inferior turbinates, and cavum. Therefore, tumors extending laterally require a combined subtemporal and/or infratemporal approach [23,24]. This was used in two cases of chordoma we operated on and that were not included in the present series. Nowadays, there is a consensus among the majority of authors to simplify the reconstruction of skull base defects after resection of tumors [4,12, 15,17,21,26,27]. The main point on which everyone agrees is that the dura should be always closed meticulously in a watertight manner. Synthetic dura (Neuropatch®) can be used to repair dural defects when reapproximation of dural edges is not feasible. Concerning the bone defect corresponding to the removed ethmoid block, we consider, as many do [4,9,13,15,26,28], that there is no need to insert bone graft or any surgical material to isolate the
cranial cavity from the nasal fossa. Indeed, the lack of reconstruction of the bone has no particular consequence in terms of rhinorrhea risks, superficial or deep suppuration, even in patients who received preoperative and/or postoperative radiotherapy [21]. From 1984 to 1992, we used autografts (i.e., bone sheets taken from the inner table of the frontal flap), allografts of irradiated bone or coral blocks [21]. Starting in January 1993, we stopped using any of these materials and tried, in 14 patients, to reinforce the defect with a simple vicryl plate (VTC®). We now use only the pedicled pericranium flap, which is brought into the craniotomy site, folded in two layers, fixed with fibrin glue, and sutured to peripheral dura. We believe there is no need to add musculocutaneous flap as it was advocated by some authors [12,25]. We do not perform canthopexy, as already stated above. Great care is taken to remove frontal sinus mucosa because it reduces the risk of mucocele and infections.
Conclusion The subcranial fronto-orbito-nasal approach is the most direct procedure for management of skull base lesions extending from the frontal sinus to the clivus. The advantages of wide exposure, en bloc resection of malignant tumors, reduced complications and hospital stay, as well as the possibility of combination with other craniofacial approaches make the SFON approach an excellent surgical procedure. We recommend this approach for the treatment of ethmoido-sphenoidal tumors, providing the lesion has been well defined on neuroimaging studies. For management of chordomas and other uncommon tumors of the clivus, the SFON approach can be combined with transcranial procedures. The development of interventional interactive surgical navigation will most probably facilitate such surgical procedures. Planning of skin incision, skull bone flap size, and facial osteotomies can nowadays be facilitated by computerized three-dimensional (3D) pre- and intraoperative assistance provided by a surgical navigation unit [6]. The last three operations we performed were planned and controlled using such 3D interactive procedures. REFERENCES 1. Cophignon J, George B, Marchac D, Roux FX. Voie transbasale e´largie par mobilisation du bandeau fronto orbitaire me´dian. Neurochirurgie 1989;29:407– 10. 2. Dandy WE. Orbital tumors. Results following the transcranial operative attack. New York: Oskar Piest Publications, 1941.
508 Surg Neurol 1999;52:501–10
3. Darrouzet V, San Galli F, Liguoro D, Pinsolle J, Stolll D, Bebear JP. La voie sous-fronto-nasale. Evolution de la chirurgie par voie mixte des tumeurs ethmoı¨dales. Expe´rience de 12 cas. Rev Laryngol Otol Rhinol 1993; 114:217–20. 4. Delfini R, Iannetti G, Belli E, Santoro A, Ciappetta P, Cantore G. Cranio-facial approaches for tumors involving the anterior half of the skull base. Acta Neurochir (Wien) 1993;124:53– 60. 5. Derome P. Les tumeurs sphe´no-orbitaires. Possibilite´s d’exe´re`se et de re´parations chirurgicales. Neurochirurgie 1972;18(suppl. 1):164. 6. Ghande AJ, Hill DLG, Studholme C, Hawkes DJ, Ruff CF, Cox TCS, Gleeson MJ, Strong AJ. Combined three dimensional rendered multimodal data for planning Cranial base surgery: a prospective evaluation. Neurosurgery 1994;35:463–71. 7. Holmes B, Sekhar L, Sofaer S, Holmes KL, Wright DC. Outcomes and analysis in cranial base surgery. Preliminary results. Acta Neurochir (Wien) 1995;134: 136 – 8. 8. Jackson RT, Fitz-Hugh GS, Constable WC. Malignant neoplasms of the nasal cavities and paranasal sinuses. A retrospective study. Laryngoscope 1977;87: 726 –36. 9. Jackson RT, Adham MN, Marsh WR. Use of the galeal frontalis myofascial flap in craniofacial surgery. Plast Reconstr Surg 1986;77:905–20. 10. Jane JA, Park TS, Peberskin LH. The supraorbital approach: technical note. Neurosurgery 1982;11:537– 42. 11. Jho HD, Ko Y. Glabellar approach: simplified midline anterior skull base approach. Minim Invas Neurosurg 1997;40:62–7. 12. Johns ME, Winn HR, McLean WC, Cantrell R. Pericranial flap for the closure of defects of craniofacial resections. Laryngoscope 1981;6:952– 8. 13. Johns ME, Kaplan MJ. Advances in the management of paranasal sinus tumors. In: Gregory T, ed. Head and neck oncology. Boston: Martinus Nijhoff Publishers, 1984:25–52. 14. Ketcham AS, Wilkins RH, Van Buren JM. A combined intra-cranial facial approach to the paranasal sinuses. Am J Surg 1963;103:698 –703. 15. Ohata K, Hakuba A, Nagai K, Morino M, Iwa Y. A biorbitofrontobasal interhemispheric approach for suprasellar lesions. Mount Sinai J Med. 1997;64:217– 21. 16. Osguthorpe JD. Sinus neoplasia. Arch Otolaryngol Head Neck Surg 1994;120:19 –25. 17. Raveh J, Laedrach K, Speiser M, Chen J, Vuillemin T, Seiler R, Ebeling U, Leibinger K. The subcranial approach for fronto-orbital and anterioposterior skullbase tumors. Arch Otolaryngol Head Neck Surg 1993; 119:385–93. 18. Raveh J, Turk J, Ladrach K, Seiler R, Godoy N, Chen J, Paladino J, Virag M, Leibinger K. Extended anterior subcranial approach for skull base tumors: long-term results. J Neurosurg 1995;82:1002–10. 19. Ray BS, McLean JM. Combined intracranial and orbital operation for retinoblastoma. Arch Ophtalmol 1943;30:437– 45. 20. Rinehart GC, Jackson IT, Potp AZ. Management of locally aggressive sinus disease using craniofacial ex-
Roux et al
21. 22.
23.
24.
25.
26.
27. 28. 29. 30.
31.
posure and the galeal frontalis fascia muscla flap. Plast Reconstr Surg 1993;92:1219 –25. Roux FX, Brasnu D, Loty B, George B, Guillemin G. Madreporic coral: a new bone graft substitute for cranial surgery. J Neurosurg 1988;69:510 –13. Roux FX, Brasnu D, Devaux B, Chabardes E, Schwaab G, Laccourreye O, Menard M, Janot F, Nguyen S, Bertrand J, Meder JF. Ethmoid sinus carcinomas: results and prognosis after neoadjuvant chemotherapy and combined surgery. Surg Neurol 1994;42:98 –104. Roux FX, Laccourreye O, Devaux B, Cioloca C, Brasnu D. Evolution des techniques de reconstruction de la base ante´rieure du cra ˆne apre`s exe´re`se tumorale. Ann Otolaryngol Chir Cervicofac 1996;113:29 –33. Roux FX, Devaux B, Nataf F, Pages JC, Laccourreye O, Menard M, Brasnu D. Tumeurs malignes de la re´gion ethmoidale: techniques neurochirurgicale. Neurochirurgie 1997;43:92–9. Sekhar LN, Janecka IP, Jones NF. Subtemporal infratemporal and basal subfrontal approach to extensive cranial base tumours. Acta Neurochirurgica 1988;92: 83–92. Sekhar LN, Sen CN. An extended frontal approach to tumors involving the skull base, In: Wilkins RH, Rengachary SS, eds. Neurosurgery update. I Diagnosis, operative technique and neurooncolgy. New York: McGraw-Hill, 1990:292–301. Shah JT. Surgical treatment of tumors involving the anterior skull base. In: Ward, PH, ed. Head & neck cancer, Vol.2. Toronto: B. C. Decker, 1990:393–9. Smith KR, Klopp CT, Williams JM. Surgical treatment of cancer of the frontal sinus and adjacent areas. Cancer 1954;7:991– 4. Spetzler RF, Herman JM, Beals S, Joganic E, Miligan J. Preservation of olfaction in anterior craniofacial approaches. J Neurosurg 1993;79:48 –52. Van Buren JM, Ommaya AK, Ketcham AS. Ten years’ experience with radical combined cranio-facial resection of malignant tumors of the paranasal sinuses. J Neurosurg 1968;28:341–50. Van Dijk JMC, Thomeer RTWM. Control of complications in the midfrontobasal approach. Acta Neurochirurgica 1997;139:355– 8.
COMMENTARY
In this paper, the authors describe their positive experience with the subcranial fronto-orbito-nasal approach for treatment of ethmoidal cancers. Although this approach has been widely used for several years, the series described represents an important contribution because the authors confirm its advantages, point out some important technical details, and describe an excellent outcome, rightly emphasizing the lower number of complications. Their classification of ethmoidal cancers is also praiseworthy since no classification permitting a rational preoperative grading scale for these lesions—particularly for the purposes of surgery and comparison of case material— has yet appeared in the relevant literature. In this context, I
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
Surg Neurol 509 1999;52:501–10
agree completely with their statement that patients harboring large carcinomas with a voluminous intracerebral extension should not be operated. Unfortunately, as the authors themselves point out, the follow-up period is not long enough to permit long-term assessment of the surgical results. This is extremely important, not only to verify the usefulness of a relatively complex operation, but also to determine survival times for the various oncotypes.
trauma, and congenital midface advancement patients, it is usually best to reconstruct the anterior skull base with cranial bone grafts at the time of the procedure. The donor site is already exposed and harvesting of the grafts can be performed simply and quickly. The reconstruction does not add significant time to the procedure. Bony continuity of the anterior skull base is definitely beneficial in terms of creating a seal between the nasal pharynx and the anterior cranial fossa. This will help reduce the incidence of postoperative CSF leaks as well. Once the bone graft reconstruction has been undertaken, it can then be lined with the soft tissue pericranial flap, adding a second protective seal layer. I think the bony reconstruction is particularly important in patients in whom the lateral nasal walls are removed as well. Skeletal reconstruction of the lateral nasal walls will reestablish preoperative nasal form and also allow a skeletal anchorage point for the required medial cathopexys. For patients in whom the facial soft tissues are not involved in the disease, restoration of the normal morphology, and architecture of the soft tissues should always be provided.
Roberto Delfini, M.D. Neurosurgery Clinic Universita ` degli Studi di Roma “La Sapienza” Rome, Italy The authors have a nice series, with a relatively short follow-up, of patients with mid-facial carcinomas that they treated surgically through a frontal orbital approach. This is a well-written article; however, I think the following comments should be made. 1. The authors state that transection of one or both of the lacrimal ducts is performed during surgery. They describe an oblique cut used to attempt to avoid stenosis at the cut end, with potential epiphora for patients postoperatively. When the lateral nasal walls are removed en bloc with their specimen as they describe, this can certainly be a problem. This can be nicely managed, however, through either pre- or postoperative stenting of the cannicular and duct system, leaving the stents in place for several weeks postoperatively. The stents can be brought out through the nasal vestibules and removed easily in the clinic several months after the surgery, after the period of wound contracture. This will ensure that the cut ends will not stenose and patients can avoid the bothersome epiphora that can occur with this type of surgery. 2. The authors state that they never perform a medial canthopexy for their patients. Although the authors did not mention this, postoperative medial canthal drift in this type of resection can certainly be not only a cosmetic problem, but potentially a functional problem as well. Bilateral medial canthopexys can certainly be performed and are enhanced by tagging the medial canthal ligaments prior to their takedown. This routine procedure should be performed any time the medial canthal structures are released from their periosteal attachments. 3. The authors state that the skeletal cranial base defect is not reconstructed in their patients. They use a pericranial flap for lining the anterior fossa in these patients. I believe that in tumor,
John W. Polley, M.D. The Craniofacial Center University of Illinois at Chicago Chicago, Illinois In this article, Roux et al present the results of treatment of ethmoidal cancers using the wellknown subcranial approach developed by Joram Raveh. Their surgical results in regard to resection and complications were good, but their long-term results of tumor management were not assessed because of stent follow-up. It is important to note that a number of their patients received preoperative chemotherapy, and some received postoperative radiotherapy, although the selection of patients for these treatments is not specified. During the last 5 years, I have treated extradural anterior cranial base malignancies and benign tumors using only the prevously described “extended subfrontal approach” [1]. This approach differs somewhat from the subcranial approach of Raveh in that a small bifrontal craniotomy is performed, followed by an orbito-frontal osteotomy that stops inferiorly just below the nasion. The nasal bone is not removed. The entire tumor removal takes place through this expanded exposure without the need for a facial incision. An endoscope is used if necessary to complete the resection in the inferior aspect. The inferior limit of this approach is the max-
510 Surg Neurol 1999;52:501–10
illary sinus, and if tumor extends there, a sublabial degloving approach can be used. It is important to note that with our technique, the lacrimal ducts are not transected. If extensive bone loss occurs around the orbit, it must be repaired with split calvarial bone grafts, otherwise significant ocular displacement and dysfunction may result. In most patients, however, this is not needed. Using this method, we did not experience any brain CSF leak, infections, or other complications in patients operated on during the last 5 years. Some potential disadvantages of the subcranial approach are the possible loss of the nasal bone caused by radiation or infection, and excessive tear-
Roux et al
ing attributable to interruption of the lacrimal duct. Otherwise, the reported technique is more than adequate to deal with ethmoidal malignancies. Laligam N. Sekhar, M.D., FACS Department of Neurological Surgery George Washington University Washington, DC
REFERENCE 1. Sekhar LN, Nanda A, Sen CN, Snyderman CH, Janecka IP. The extended frontal approach to tumors of the anterior, middle, and posterior skull base. J Neurosurg 1992;76:198 –206.
study of Bosnia refugees in a refugee camp shows those demonstrating psychiatric symptoms had increased risk for disability. Richard F. Mollica, M.D., M.A.R., of the Harvard Program in Refugee Trauma in Cambridge, MA, and colleagues surveyed 534 Bosnian refugee adults randomly sampled from families living in a camp established by the Croatian government near the city of Varazdin. The mean age of refugees in this study was 50 years, and 59 percent were female. While 55 percent of refugees surveyed reported no psychiatric symptoms, 39 percent reported symptoms of depression, and 26 percent reported symptoms of PTSD. Twentysix percent reported having a disability. Those refugees who reported symptoms for both depression and PTSD were two times more likely to report disability than those who reported no psychiatric symptoms.
A
—AMA Press Release September 8, 1999
Roux FX, Moussa R, Devaux B, Nataf F, Page P, Laccourreye O, Schwaab G, Brasnu D, Lacau Saint-Guily J. Subcranial frontoorbito-nasal approach for ethmoidal cancers: surgical techniques and results. Surg Neurol 1999;52:501–10. BACKGROUND
The authors report their experience with the subfrontoorbito-nasal approach (SFON) for the treatment of 30 patients suffering from ethmoidal cancers over the past 3 years. The advantages and pitfalls of this technique are described and compared with other classic approaches. METHODS
Among 156 patients suffering from ethmoidal cancers and treated between January 1984 and January 1998, 30 patients were operated on using the SFON approach during the past 3 years. There were 27 males and 3 females, ranging in age from 15 to 77 years. Histologic composition of the lesions was as follows: 15 adenocarcinomas, 6 esthesioneuroblastomas, 3 melanomas, 2 epidermoid carcinomas, 1 nondifferentiated carcinoma, 1 neuroendocrine carcinoma, 1 villous carcinoma, and 1 cystic adenoid carcinoma (cylindroma). According to the authors’ classification, 7% were T1, 6% T2, 22% T3, 38.5% T4a, and 26.5% T4b. All patients were operated on through a SFON approach, followed by removal of the tumor and reconstruction of the skull base with a pericranial flap. RESULTS
Since the mean follow-up was of short duration (12 months, ranging from 3 to 29 months), significant carcinologic results could not be obtained. However, a detailed analysis of the surgical procedure was performed. No patient died or had major complications related to the SFON approach. One cerebrospinal fluid (CSF) fistula and four oculomotricity dysfunctions were observed. Definitive anosmia was reported in all cases.
Address reprint requests to: F. X. Roux, Service de Neurochirurgie, Hopital Sainte-Anne, 1 Rue Cabanis 75014 Paris. Received July 16, 1998; accepted December 7, 1998. © 1999 by Elsevier Science Inc. 655 Avenue of the Americas, New York, NY 10010
CONCLUSION
The advantages of the procedure include a wide exposure of the anterior skull base through a limited approach, the possibility of modifying the approach according to the size and location of the lesion, total resection of tumors, simplified skull base reconstruction technique, and reduction of postoperative confusion and hospital stay. © 1999 by Elsevier Science Inc. KEY WORDS
Subfronto-orbito-nasal approach, skull base surgery, ethmoidal cancers.
rom 1984 to 1995, we have been using a combined transcranial transfacial approach to remove ethmoidal or ethmoidosphenoidal cancers [21–24]. We believed that it would be reasonable to modify our technique and choose a more limited surgical route to minimize postoperative morbidity as much as possible. Our aim was to avoid, or at least decrease, the frequency of incidents or complications such as: 1) Dural tears, which can be significant when the dura is adherent to the inner table of the frontal skull, particularly in patients over 60 years old. Such dural tears can be responsible for postoperative cerebrospinal fluid leakage, and therefore require watertight reconstruction or reinforcement, which is more difficult to perform bedcause they are larger; 2) Immediate postoperative confusion, mainly in elderly patients, attributable to frontal lobe retraction during a subfrontal transcranial approach; 3) Retractile transfacial scar such as can be seen after radiotherapy when the patient has undergone a combined surgical approach. That is why we began using a subcranial
F
0090-3019/99/$–see front matter PII S0090-3019(99)00127-5
502 Surg Neurol 1999;52:501–10
1
Roux et al
Histological Data of Ethmoidal Cancers Operated Via the SFON Approach
NUMBER OF PERCENTAGE CASES (%) Adenocarcinoma Esthesioneuroblastoma Melanoma Epidermoid carcinoma Cylindroma (adenoid cystic) Neuroendocrine tumor Villous tumor Nondifferentiated tumor Total
15 6 3 2 1 1 1 1 30
50 20 10 8 3 3 3 3 100
approach inspired by the procedure described by Raveh in 1978 [17–18]. In this report, we describe our experience with the subcranial-fronto-orbito-nasal approach (SFON) as used for the treatment of 30 ethmoidal cancers over the past 3 years. We shall focus on the surgical data and postoperative results and compare them with those obtained with combined transcranial transfacial procedures for equivalent indications. We exclude carcinologic results that have been presented recently in other issues.
Materials and Methods PATIENT POPULATION From January 1984 to January 1998, 221 patients were referred to our institution for anterior cranial base procedures (excluding pituitary surgery). Among them, 156 patients suffered from ethmoidal cancer: 106 were operated through a combined transcranial and transfacial approach, 30 through an extended SFON route, 20 were not operated. There were 27 males and 3 females who underwent a SFON procedure. Patients’ ages ranged from 15 to 77 years, with a mean age of 57.5; 13 patients were over 65, and seven were over 70. Table 1 shows histologic composition and location of the lesions treated by the SFON approach. PREOPERATIVE GRADING Patients were evaluated by a team including a head and neck surgeon, anesthesiologist, and neurosurgeon. Computed tomography (CT) scans with axial and coronal views and magnetic resonance imaging (MRI) were performed in 29 and 28 patients, respectively. Radiological evaluation revealed a tumoral extension. Such grading is necessary to confirm surgical indication and determine the type of approach. Because of the absence of any international clas-
sification for ethmoidal cancers, we have been using a personal grading scale for the last 12 years [22]: T1: the tumor is developed in only one ethmoidal region, T2: the tumor is developed in at least two regions of the ethmoid sinuses with respect to the walls, T3: the tumor destroys any ethmoid cell wall, except the cribriform plates, T4a: the tumor destroys the cribriform plate(s) without intracranial extension or dural invasion, T4b: tumor with intracranial extension whatever its volume. In the present series, 7% of tumors were T1, 6% T2, 22% T3, 38.5% T4a, and 26.5% T4b. PREOPERATIVE MANAGEMENT Five of these 30 patients were referred to our team with a tumor recurrence: three had undergone a previous transfacial operation (1 epidermoid, 1 esthesioneuroma, 1 villous tumor); one had been treated by chemotherapy and radiotherapy with no surgery for an undifferentiated carcinoma; and one had transfacial surgery followed by radiotherapy for an adenocarcinoma. In all circumstances, histologic diagnosis was confirmed or reconfirmed by endonasal biopsy performed in the ENT department before combined surgery. An inductive chemotherapy was then started, consisting of 4-day continuous infusion of cisplatin (25 mg/m2/day) and 5-fluorouracyl (1 g/m2/day). An average of three cycles were administered at 3-week intervals. Eighteen patients went through chemotherapy: 11 adenocarcinomas, 4 esthesioneuromas, 1 epidermoid carcinoma, 1 melanoma, and 1 neuroendocrine carcinoma. Twelve patients had no chemotherapy either because they were harboring a tumor nonresponsive to chemotherapy, or because chemotherapy could not be used for some reason. Preoperative MRI evaluation was conducted 2 to 3 weeks after the last chemotherapy course, a few days before surgery. It was then possible to assess the quality of the response to chemotherapy [21, 23]. The operation was then performed, 3 to 4 weeks after the last chemotherapy course, by the neurosurgical team in collaboration with the head and neck surgeons. OPERATIVE TECHNIQUE Via a bicoronal skin incision, the scalp flap separated from the pericranium is elevated. Care is taken to preserve the pericranium for use during reconstruction of the skull base. Dissection of a pedicled pericranial flap measuring 10 to 12 cm in length and 8 to 10 cm in width is performed (Figure 1). Once it has been elevated over the calvarium
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
Intraoperative photograph showing the pedicled pericranial flap to be used for skull base reconstruction.
1
and protected with wet gauze, subperiostal dissection is continued down to the periorbita, which is separated from the superior, medial and occasionally lateral walls of the orbit. The supraorbital nerves are preserved by freeing them from the supraorbital canals, maintaining continuity of the pericranium and periorbita. Care is taken to preserve the supraorbital arteries supplying the pericranial flap. Dissection is extended to expose the entire nasal pyramid down to the cartilage in the midline and laterally as far as the lacrymal crests. One or both lacrymal ducts may be cut obliquely to avoid or limit retraction and stenosis at their extremities. The osteotomy line is usually extended 3 to 4 cm cranially and 3 to 4 cm laterally, limited by the supraorbital canals, but may be extended deliberately according to tumor extension. One burr hole is then made at the midline and osteotomies are made across the frontal bone down to and along orbital roofs, down the medial orbital wall and along the nasomaxillary grooves just anterior to the lacrymal crest. The nasal pyramid is separated 1 to 5 mm above the insertion of the septal cartilage. A vertical osteotomy across the skull base just anterior to the crista galli, allows detachment of the frontonasal segment and exposure of the anterior skull base (Figure 2). The dura is carefully protected during the procedure. It is divided around the olfactory groove, severing the olfactory filaments. Then it is separated subcranially from the floor of the anterior fossa without frontal lobe retraction. If dura is resected because it was invaded, the defect is repaired either with pericranium or with synthetic dura. Broad access is provided to the nasal, maxillary, and sphenoethmoidal structures (Figure 3). With a high-speed drill, an electric
Surg Neurol 503 1999;52:501–10
saw, and/or chisels, cuts are made through the cribriform plates. The posterior limit of the resection is at the planum sphenoidale. The specimen is removed with the entire ethmoid block, including one or both lamina papyracea, the superior and middle turbinates on both sides, and the nasal septum. It must be stressed that removal of middle turbinates has been in a few patients a little more difficult to achieve than through a classic transfacial approach. The resection of the ipsilateral orbital floor was limited to a fourth or a third of its width. The integrity and vacuity of both maxillary sinuses were checked, as well as of the sphenoid sinus, which was systematically entered. Tumor removal was finally achieved as far as possible “en bloc”; in fact, it was more often performed in a piecemeal fashion because the lesion usually split into pieces (Figure 4). RECONSTRUCTIVE PROCEDURE The dura was closed meticulously in a watertight manner with stitches, reinforced with collagen glue. All dural lacerations were closed by primary suture; but, as noted previously, if necessary any dural defect was reconstructed or reinforced with a graft. We did not reconstruct the cranial base defect itself. We only used the pedicled pericranial flap, which was inserted deeply to the jugum sphenoidalis in two layers, fixed to the dura with sutures and reinforced with surgical glue. Such a pericranial flap is well vascularized by both supraorbital vessels and should be sufficient to ensure an efficient closure of the intracranial cavity. No canthopexy was performed. The bone flap was repositioned and secured either with microplates and screws (26 cases) or with nonabsorbable sutures (4 cases) (Figure 5). The burr hole(s) were filled in with bone dust. Four to six Merocel® tubes were left in the nasal cavities and removed on the third postoperative day. The bicoronal incision was closed in layers and drained with a suction drain. Perioperative lumbar drainage was not instituted. Antibiotic drugs (amoxicillin-clavulanic acid, Augmentin®) were given for 3 to 5 days postoperatively. POSTOPERATIVE COURSE The patient was allowed to walk on the first or second postoperative day. Subcutaneous (s.c.) drainage was removed on the second postoperative day. Endonasal drains were removed and antibiotic drugs withdrawn on the third postoperative day. Postoperative stay in the intensive care unit was limited to an average of 2 days and the patient was discharged from the hospital after a stay limited to 8 to 10 days. All patients had an MRI before leaving
504 Surg Neurol 1999;52:501–10
Roux et al
Intraoperative photograph (left) showing removal of the fronto nasal unit after the dura has been dissected and the osteotomy performed. The illustration (bottom right) shows the angle of view in the SFON approach (black arrow) compared with the transcranial approach (white arrow). The angle is almost doubled, reducing the need to retract frontal lobes.
2
the hospital confirming the extent of the tumor resection, which was total in all cases.
Results PATIENT FOLLOW-UP Patients were followed for a mean period of 12 months, ranging from 3 to 29 months. This follow-up is too short for significant carcinologic results, which is not the aim of this work, but does allow a detailed analysis and assessment of the surgical technique. MORTALITY One patient died postoperatively from acute brain swelling after removal of a metastatic lesion located in the left frontal lobe, which was performed during the same procedure done for his ethmoidal carcinoma. During the follow-up, one patient with an ethmoidal adenocarcinoma type T4b died 3 months after operation from multiple liver metastases.
POSTOPERATIVE COMPLICATIONS AND MORBIDITY No patient suffered from confusion in the immediate postoperative period, even elderly patients. One patient developed a CSF rhinorrhea that was noted after the removal of the endonasal packing. A ventriculoperitoneal shunt was placed 6 days later, and led to a quick and definitive stop of the leakage. In two other patients, a liquid discharge was observed after the removal of the nasal Merocel® meshes on the third postoperative day, but it was not as clear as CSF, and most probably was a mixture of water used to rinse the surgical cavity at the end of the procedure and exudation mucosities tinged with blood; nevertheless it was considered to be a possible CSF leak and was therefore managed with fluid restriction, acetazolamide, and daily lumbar punctures for 3 days. No cases of meningitis or deep suppuration were noted. Two s.c. infections were reported; they were cured easily with 5 days of antibiotic therapy.
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
3
Intraoperative view showing the wide exposure of nasal, maxillary, and sphenoidal structures.
Oculomotricity dysfunction was reported in four cases; it was caused by a slight desaxation of the eyeball, and associated with a moderate enophthalmia in three cases and with a watering eye in one case. All these patients displayed almost complete
4
Surg Neurol 505 1999;52:501–10
5
Intraoperative view showing the flap replaced and secured with miniplates.
resolution with time. Two other patients had watering eyes. Anosmia was, of course, observed in all patients. Cosmetic results were satisfactory in all cases except one patient who received preoperative radi-
CT scan, coronal views. Left: preoperative view showing the location and extension of an adenocarcinoma. Right: postoperative view with bone window showing total removal of the lesion.
506 Surg Neurol 1999;52:501–10
ation therapy, in whom miniplates were visible beneath the skin.
Discussion The anterior base of the skull can be approached by different routes, according to location, size, and histologic type of the lesions. For benign tumors and traumatic lesions, endocranial or subcranial approaches offer a broad operative visualization of the region. For malignant tumors, a combined transcranial transbasal approach may be needed for resection of paranasal tumors invading the skull base [4,8,13–15,20]. Dandy [2], and Ray and McLean [19] first described, in 1941 and 1943 respectively, a combined approach for resection of orbital tumors. Smith et al [28], Ketcham et al [14], and Van Buren et al [30] were among the first authors to report on a combined transcranial and transfacial resection of a paranasal sinus tumor, in 1953. In 1972, P. Derome [5] described the transbasal removal of sphenoethmoidal tumors. Since that time, modifications of these approaches have been undertaken by many head and neck surgeons and neurosurgeons around the world [8 –10,24]. The subcranial approach was first described by Raveh in 1978, primarily for the treatment of skull base trauma and craniofacial anomalies. Since 1980, he has extended the indications to tumor resection [17,18]. The advantages of this approach have been confirmed by Sekhar et al [24,25], Spetzler et al [29], Delfini et al [4], and Darrouzet et al [3]. These advantages are based on the surgical exposure and direction of operative approach, avoidance of frontal lobe retraction, avoidance of facial incision, en bloc resection of intra- and extradural tumors, and watertight skull base reconstruction. In our department, and in accordance with the ENT teams we work with, we consider that small and medium size tumors, i.e., T1 and T2 tumors, as long as they are at a minimal distance of 10 mm from the cribriform plates, should be operated through an only transfacial and/or endonasal route. All other tumors (T3, T4a, and T4b) justify a combined approach. Because we decided to operate ethmoidal cancers through a SFON approach, we reserve combined craniofacial approaches for paranasal lesions involving the infrastructure, the palate and/or the maxillary bone and sinus since in such situations the SFON exposure is too limited. Surgery is not recommended for patients harboring large carcinomas presenting with a voluminous intracerebral extension, bilateral orbital development, and/or posterior invasion toward the sphenoid and
Roux et al
clivus. These patients undergo a radiotherapy and/or chemotherapy regimen. Among different advantages of a SFON approach, performing a small subcranial osteotomy should reduce the risks of dural tears, and therefore should reduce the frequency of postoperative CSF leakage. Indeed, dural tears are less frequent and less extended with such a limited approach. This is particularly noted in elderly patients in whom the dura is often very adherent in the frontal area, particularly near the midline. Delfini [4] and Raveh et al [17,18] emphasized this fact, recommending the use of dural suspension and watertight closure. But it is always difficult to reach the goal of no leakage. Raveh recorded a 2.8% rate of CSF leakage [18]. In our series, we had one postoperative CSF rhinorrhea, corresponding to a 3.5% figure. It can be compared with the rate of rhinorrhea we recorded after combined procedures: 5 of 82 patients (6%). The difference between these figures is informative, but not significant. Concerning the treatment of small leaks, most authors recommend spinal fluid drainage or lumbar punctures. The SFON approach allows a direct visualization of the anterior skull base parallel to its plane [17]; consequently, because there is minimal or no retraction of the frontal lobes, it avoids or at least limits postoperative confusion and frontal syndrome as are sometimes encountered after a combined subcranial and transfacial approach, mainly in elderly patients. Indeed, none of our patients operated through this approach developed such symptoms, whereas 11 of 82 patients (13.5%) had a transient postoperative confusion after combined procedures. Therefore, this technique accelerates postoperative recovery, and reduces intensive care unit and hospital stay as well as time taken to return to work and independent life. The cost of treatment is consequently reduced, which nowadays is an important point to be considered [7]. The SFON approach does not seem to be responsible for more postoperative diplopia than combined surgery. Raveh recorded five oculomotor, trochlear or abducens nerve chronic dysfunctions [18], i.e,. 4.8%. We did not find any information concerning transient postoperative diplopia in his papers. Darrouzet reported two transient diplopias in four patients [3]. As for us, we noted four postoperative transient (13.3%) and one persistent (3.3%) diplopias in our 30 patients who had a SFON procedure. The figures were slightly different for the 82 patients we operated through a combined procedure: four transient (4.8%), and six persistent diplopias (7.3%). These functional results are comparable and almost identical, indeed the variations of
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
Surg Neurol 507 1999;52:501–10
figures are not significant. Raveh et al [18] and Darrouzet [3] performed canthal ligament fixation, we did not. Therefore, we think canthopexy is not really useful in most cases. The SFON approach to the anterior skull base allows en bloc removal of malignant tumors [17], because it offers a broad access to sphenoethmoidal structures as far as the jugum and clivus posteriorly. Nasal and paranasal cavities, including both maxillary sinuses, can be easily controlled, as well as orbital structures when necessary [1,22]. The ability to modulate the approach, depending on the tumor extension and surgical needs, must be emphasized. Osteotomy can be performed either through an enlarged route as described by Cophignon et al [1], or through a limited glabellar approach as described by Jho and Ko [11]. For instance, extensive intradural or brain involvement may require an enlargement of midfrontal osteotomy. An extended glabellar approach along the eyebrows, as reported by Osguthorpe [16] may also offer excellent exposure of the anterior skull base; it provides sufficient pericranial tissue for skull base reconstruction and subfrontal dura lining, and good cosmetic results. Furthermore, ethmoidosphenoidal tumors extending laterally may require an associated transfacial approach, which can be easily undertaken with a SFON; this is particularly the case when the tumor extends toward the orbit and/or the maxillary sinus. In two of our patients (Cases 4 and 15), we performed such a combined procedure to obtain complete resection of the tumor. Despite allowing a rather large access to the cranial base, the SFON approach of course cannot solve all problems. Its usual anatomic limits are: laterally, the optic nerves, cavernous sinuses, and carotid arteries; posteriorly, the clivus, basilar artery, and brain stem; caudally, the maxillary sinuses, inferior turbinates, and cavum. Therefore, tumors extending laterally require a combined subtemporal and/or infratemporal approach [23,24]. This was used in two cases of chordoma we operated on and that were not included in the present series. Nowadays, there is a consensus among the majority of authors to simplify the reconstruction of skull base defects after resection of tumors [4,12, 15,17,21,26,27]. The main point on which everyone agrees is that the dura should be always closed meticulously in a watertight manner. Synthetic dura (Neuropatch®) can be used to repair dural defects when reapproximation of dural edges is not feasible. Concerning the bone defect corresponding to the removed ethmoid block, we consider, as many do [4,9,13,15,26,28], that there is no need to insert bone graft or any surgical material to isolate the
cranial cavity from the nasal fossa. Indeed, the lack of reconstruction of the bone has no particular consequence in terms of rhinorrhea risks, superficial or deep suppuration, even in patients who received preoperative and/or postoperative radiotherapy [21]. From 1984 to 1992, we used autografts (i.e., bone sheets taken from the inner table of the frontal flap), allografts of irradiated bone or coral blocks [21]. Starting in January 1993, we stopped using any of these materials and tried, in 14 patients, to reinforce the defect with a simple vicryl plate (VTC®). We now use only the pedicled pericranium flap, which is brought into the craniotomy site, folded in two layers, fixed with fibrin glue, and sutured to peripheral dura. We believe there is no need to add musculocutaneous flap as it was advocated by some authors [12,25]. We do not perform canthopexy, as already stated above. Great care is taken to remove frontal sinus mucosa because it reduces the risk of mucocele and infections.
Conclusion The subcranial fronto-orbito-nasal approach is the most direct procedure for management of skull base lesions extending from the frontal sinus to the clivus. The advantages of wide exposure, en bloc resection of malignant tumors, reduced complications and hospital stay, as well as the possibility of combination with other craniofacial approaches make the SFON approach an excellent surgical procedure. We recommend this approach for the treatment of ethmoido-sphenoidal tumors, providing the lesion has been well defined on neuroimaging studies. For management of chordomas and other uncommon tumors of the clivus, the SFON approach can be combined with transcranial procedures. The development of interventional interactive surgical navigation will most probably facilitate such surgical procedures. Planning of skin incision, skull bone flap size, and facial osteotomies can nowadays be facilitated by computerized three-dimensional (3D) pre- and intraoperative assistance provided by a surgical navigation unit [6]. The last three operations we performed were planned and controlled using such 3D interactive procedures. REFERENCES 1. Cophignon J, George B, Marchac D, Roux FX. Voie transbasale e´largie par mobilisation du bandeau fronto orbitaire me´dian. Neurochirurgie 1989;29:407– 10. 2. Dandy WE. Orbital tumors. Results following the transcranial operative attack. New York: Oskar Piest Publications, 1941.
508 Surg Neurol 1999;52:501–10
3. Darrouzet V, San Galli F, Liguoro D, Pinsolle J, Stolll D, Bebear JP. La voie sous-fronto-nasale. Evolution de la chirurgie par voie mixte des tumeurs ethmoı¨dales. Expe´rience de 12 cas. Rev Laryngol Otol Rhinol 1993; 114:217–20. 4. Delfini R, Iannetti G, Belli E, Santoro A, Ciappetta P, Cantore G. Cranio-facial approaches for tumors involving the anterior half of the skull base. Acta Neurochir (Wien) 1993;124:53– 60. 5. Derome P. Les tumeurs sphe´no-orbitaires. Possibilite´s d’exe´re`se et de re´parations chirurgicales. Neurochirurgie 1972;18(suppl. 1):164. 6. Ghande AJ, Hill DLG, Studholme C, Hawkes DJ, Ruff CF, Cox TCS, Gleeson MJ, Strong AJ. Combined three dimensional rendered multimodal data for planning Cranial base surgery: a prospective evaluation. Neurosurgery 1994;35:463–71. 7. Holmes B, Sekhar L, Sofaer S, Holmes KL, Wright DC. Outcomes and analysis in cranial base surgery. Preliminary results. Acta Neurochir (Wien) 1995;134: 136 – 8. 8. Jackson RT, Fitz-Hugh GS, Constable WC. Malignant neoplasms of the nasal cavities and paranasal sinuses. A retrospective study. Laryngoscope 1977;87: 726 –36. 9. Jackson RT, Adham MN, Marsh WR. Use of the galeal frontalis myofascial flap in craniofacial surgery. Plast Reconstr Surg 1986;77:905–20. 10. Jane JA, Park TS, Peberskin LH. The supraorbital approach: technical note. Neurosurgery 1982;11:537– 42. 11. Jho HD, Ko Y. Glabellar approach: simplified midline anterior skull base approach. Minim Invas Neurosurg 1997;40:62–7. 12. Johns ME, Winn HR, McLean WC, Cantrell R. Pericranial flap for the closure of defects of craniofacial resections. Laryngoscope 1981;6:952– 8. 13. Johns ME, Kaplan MJ. Advances in the management of paranasal sinus tumors. In: Gregory T, ed. Head and neck oncology. Boston: Martinus Nijhoff Publishers, 1984:25–52. 14. Ketcham AS, Wilkins RH, Van Buren JM. A combined intra-cranial facial approach to the paranasal sinuses. Am J Surg 1963;103:698 –703. 15. Ohata K, Hakuba A, Nagai K, Morino M, Iwa Y. A biorbitofrontobasal interhemispheric approach for suprasellar lesions. Mount Sinai J Med. 1997;64:217– 21. 16. Osguthorpe JD. Sinus neoplasia. Arch Otolaryngol Head Neck Surg 1994;120:19 –25. 17. Raveh J, Laedrach K, Speiser M, Chen J, Vuillemin T, Seiler R, Ebeling U, Leibinger K. The subcranial approach for fronto-orbital and anterioposterior skullbase tumors. Arch Otolaryngol Head Neck Surg 1993; 119:385–93. 18. Raveh J, Turk J, Ladrach K, Seiler R, Godoy N, Chen J, Paladino J, Virag M, Leibinger K. Extended anterior subcranial approach for skull base tumors: long-term results. J Neurosurg 1995;82:1002–10. 19. Ray BS, McLean JM. Combined intracranial and orbital operation for retinoblastoma. Arch Ophtalmol 1943;30:437– 45. 20. Rinehart GC, Jackson IT, Potp AZ. Management of locally aggressive sinus disease using craniofacial ex-
Roux et al
21. 22.
23.
24.
25.
26.
27. 28. 29. 30.
31.
posure and the galeal frontalis fascia muscla flap. Plast Reconstr Surg 1993;92:1219 –25. Roux FX, Brasnu D, Loty B, George B, Guillemin G. Madreporic coral: a new bone graft substitute for cranial surgery. J Neurosurg 1988;69:510 –13. Roux FX, Brasnu D, Devaux B, Chabardes E, Schwaab G, Laccourreye O, Menard M, Janot F, Nguyen S, Bertrand J, Meder JF. Ethmoid sinus carcinomas: results and prognosis after neoadjuvant chemotherapy and combined surgery. Surg Neurol 1994;42:98 –104. Roux FX, Laccourreye O, Devaux B, Cioloca C, Brasnu D. Evolution des techniques de reconstruction de la base ante´rieure du cra ˆne apre`s exe´re`se tumorale. Ann Otolaryngol Chir Cervicofac 1996;113:29 –33. Roux FX, Devaux B, Nataf F, Pages JC, Laccourreye O, Menard M, Brasnu D. Tumeurs malignes de la re´gion ethmoidale: techniques neurochirurgicale. Neurochirurgie 1997;43:92–9. Sekhar LN, Janecka IP, Jones NF. Subtemporal infratemporal and basal subfrontal approach to extensive cranial base tumours. Acta Neurochirurgica 1988;92: 83–92. Sekhar LN, Sen CN. An extended frontal approach to tumors involving the skull base, In: Wilkins RH, Rengachary SS, eds. Neurosurgery update. I Diagnosis, operative technique and neurooncolgy. New York: McGraw-Hill, 1990:292–301. Shah JT. Surgical treatment of tumors involving the anterior skull base. In: Ward, PH, ed. Head & neck cancer, Vol.2. Toronto: B. C. Decker, 1990:393–9. Smith KR, Klopp CT, Williams JM. Surgical treatment of cancer of the frontal sinus and adjacent areas. Cancer 1954;7:991– 4. Spetzler RF, Herman JM, Beals S, Joganic E, Miligan J. Preservation of olfaction in anterior craniofacial approaches. J Neurosurg 1993;79:48 –52. Van Buren JM, Ommaya AK, Ketcham AS. Ten years’ experience with radical combined cranio-facial resection of malignant tumors of the paranasal sinuses. J Neurosurg 1968;28:341–50. Van Dijk JMC, Thomeer RTWM. Control of complications in the midfrontobasal approach. Acta Neurochirurgica 1997;139:355– 8.
COMMENTARY
In this paper, the authors describe their positive experience with the subcranial fronto-orbito-nasal approach for treatment of ethmoidal cancers. Although this approach has been widely used for several years, the series described represents an important contribution because the authors confirm its advantages, point out some important technical details, and describe an excellent outcome, rightly emphasizing the lower number of complications. Their classification of ethmoidal cancers is also praiseworthy since no classification permitting a rational preoperative grading scale for these lesions—particularly for the purposes of surgery and comparison of case material— has yet appeared in the relevant literature. In this context, I
Subfronto Orbito Nasal Approach for Ethmoidal Cancers
Surg Neurol 509 1999;52:501–10
agree completely with their statement that patients harboring large carcinomas with a voluminous intracerebral extension should not be operated. Unfortunately, as the authors themselves point out, the follow-up period is not long enough to permit long-term assessment of the surgical results. This is extremely important, not only to verify the usefulness of a relatively complex operation, but also to determine survival times for the various oncotypes.
trauma, and congenital midface advancement patients, it is usually best to reconstruct the anterior skull base with cranial bone grafts at the time of the procedure. The donor site is already exposed and harvesting of the grafts can be performed simply and quickly. The reconstruction does not add significant time to the procedure. Bony continuity of the anterior skull base is definitely beneficial in terms of creating a seal between the nasal pharynx and the anterior cranial fossa. This will help reduce the incidence of postoperative CSF leaks as well. Once the bone graft reconstruction has been undertaken, it can then be lined with the soft tissue pericranial flap, adding a second protective seal layer. I think the bony reconstruction is particularly important in patients in whom the lateral nasal walls are removed as well. Skeletal reconstruction of the lateral nasal walls will reestablish preoperative nasal form and also allow a skeletal anchorage point for the required medial cathopexys. For patients in whom the facial soft tissues are not involved in the disease, restoration of the normal morphology, and architecture of the soft tissues should always be provided.
Roberto Delfini, M.D. Neurosurgery Clinic Universita ` degli Studi di Roma “La Sapienza” Rome, Italy The authors have a nice series, with a relatively short follow-up, of patients with mid-facial carcinomas that they treated surgically through a frontal orbital approach. This is a well-written article; however, I think the following comments should be made. 1. The authors state that transection of one or both of the lacrimal ducts is performed during surgery. They describe an oblique cut used to attempt to avoid stenosis at the cut end, with potential epiphora for patients postoperatively. When the lateral nasal walls are removed en bloc with their specimen as they describe, this can certainly be a problem. This can be nicely managed, however, through either pre- or postoperative stenting of the cannicular and duct system, leaving the stents in place for several weeks postoperatively. The stents can be brought out through the nasal vestibules and removed easily in the clinic several months after the surgery, after the period of wound contracture. This will ensure that the cut ends will not stenose and patients can avoid the bothersome epiphora that can occur with this type of surgery. 2. The authors state that they never perform a medial canthopexy for their patients. Although the authors did not mention this, postoperative medial canthal drift in this type of resection can certainly be not only a cosmetic problem, but potentially a functional problem as well. Bilateral medial canthopexys can certainly be performed and are enhanced by tagging the medial canthal ligaments prior to their takedown. This routine procedure should be performed any time the medial canthal structures are released from their periosteal attachments. 3. The authors state that the skeletal cranial base defect is not reconstructed in their patients. They use a pericranial flap for lining the anterior fossa in these patients. I believe that in tumor,
John W. Polley, M.D. The Craniofacial Center University of Illinois at Chicago Chicago, Illinois In this article, Roux et al present the results of treatment of ethmoidal cancers using the wellknown subcranial approach developed by Joram Raveh. Their surgical results in regard to resection and complications were good, but their long-term results of tumor management were not assessed because of stent follow-up. It is important to note that a number of their patients received preoperative chemotherapy, and some received postoperative radiotherapy, although the selection of patients for these treatments is not specified. During the last 5 years, I have treated extradural anterior cranial base malignancies and benign tumors using only the prevously described “extended subfrontal approach” [1]. This approach differs somewhat from the subcranial approach of Raveh in that a small bifrontal craniotomy is performed, followed by an orbito-frontal osteotomy that stops inferiorly just below the nasion. The nasal bone is not removed. The entire tumor removal takes place through this expanded exposure without the need for a facial incision. An endoscope is used if necessary to complete the resection in the inferior aspect. The inferior limit of this approach is the max-
510 Surg Neurol 1999;52:501–10
illary sinus, and if tumor extends there, a sublabial degloving approach can be used. It is important to note that with our technique, the lacrimal ducts are not transected. If extensive bone loss occurs around the orbit, it must be repaired with split calvarial bone grafts, otherwise significant ocular displacement and dysfunction may result. In most patients, however, this is not needed. Using this method, we did not experience any brain CSF leak, infections, or other complications in patients operated on during the last 5 years. Some potential disadvantages of the subcranial approach are the possible loss of the nasal bone caused by radiation or infection, and excessive tear-
Roux et al
ing attributable to interruption of the lacrimal duct. Otherwise, the reported technique is more than adequate to deal with ethmoidal malignancies. Laligam N. Sekhar, M.D., FACS Department of Neurological Surgery George Washington University Washington, DC
REFERENCE 1. Sekhar LN, Nanda A, Sen CN, Snyderman CH, Janecka IP. The extended frontal approach to tumors of the anterior, middle, and posterior skull base. J Neurosurg 1992;76:198 –206.
study of Bosnia refugees in a refugee camp shows those demonstrating psychiatric symptoms had increased risk for disability. Richard F. Mollica, M.D., M.A.R., of the Harvard Program in Refugee Trauma in Cambridge, MA, and colleagues surveyed 534 Bosnian refugee adults randomly sampled from families living in a camp established by the Croatian government near the city of Varazdin. The mean age of refugees in this study was 50 years, and 59 percent were female. While 55 percent of refugees surveyed reported no psychiatric symptoms, 39 percent reported symptoms of depression, and 26 percent reported symptoms of PTSD. Twentysix percent reported having a disability. Those refugees who reported symptoms for both depression and PTSD were two times more likely to report disability than those who reported no psychiatric symptoms.
A
—AMA Press Release September 8, 1999