INTRODUCTION
The craniovertebral junction is a critical anatomical region that contains the lower brainstem, upper cervical spinal cord, and lower cranial nerves, while enabling precisely coordinated multidirectional movements. Owing to its depth and close proximity to vital structures such as the nasopharynx, oropharynx, and cervical internal carotid arteries, surgical access to this area is particularly challenging. Lesions in the ventral craniovertebral junction—including neoplastic, congenital, or inflammatory pathologies—can cause severe neurological dysfunction and may require surgical intervention [1,2]. Traditionally, these lesions have been approached via microscopic transoral or transcervical techniques [3-5]. However, each of these approaches presents specific challenges. The transoral route provides direct midline access but requires incision of the oropharyngeal mucosa, increasing the risk of serious complications such as upper airway obstruction and mediastinitis [3,6]. The transcervical approach offers a sterile corridor but may inadequately expose the rostral craniovertebral junction and typically results in a narrow, deep operative field [7-9]. Furthermore, these traditional approaches may require additional invasive procedures—such as Le Fort osteotomy or glossotomy—to ensure sufficient exposure, thereby increasing surgical morbidity [10-12]. With recent advances in endoscopic technology and an improved understanding of microsurgical anatomy in this region, the endoscopic endonasal approach has emerged as a less invasive alternative [3,5,13,14]. This approach provides access to the anterior craniovertebral junction through the nasal cavity via an incision in the posterior end of the nasal septal mucosa and opening the nasopharyngeal wall. Because it proceeds caudally through the nasal cavity, invasive procedures such as mandibular splitting or glossotomy are avoided. Moreover, since the pharyngeal incision remains above the oropharynx, the risk of serious complications—such as airway obstruction and mediastinitis—is reduced. While the clinical application of this approach continues to expand because of its advantages, optimal strategies for perioperative airway management and occipitocervical stabilization remain to be established [14-16]. In this study, we aimed to describe the surgical technique of the endoscopic endonasal transnasopharyngeal approach and to present surgical outcomes, along with clinical strategies for approach selection, postoperative complications, and perioperative airway and spinal stabilization management.
MATERIALS AND METHODS
1. Patients and Clinical Data
This retrospective observational study involved 7 consecutive patients with ventral craniovertebral junction lesions who underwent surgery using the endoscopic endonasal transnasopharyngeal approach at Dokkyo Medical University Hospital between December 2021 and January 2025. Clinical data were retrospectively reviewed, including physical findings, preoperative evaluations, surgical records, perioperative complications, radiological imaging, the need for spinal stabilization, airway management strategies, and postoperative follow-up outcomes. The study was approved by the Institutional Review Board (IRB) of Dokkyo Medical University Hospital (R-52-8J) and conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all patients.
2. Preoperative Planning
1) Approach selection
The indication for the endoscopic endonasal transnasopharyngeal approach was primarily based on the vertical and lateral extent of the lesions (Fig. 1A and B). The inferior limit of this approach was estimated using bony landmarks such as the rhinion, anterior, and posterior nasal spine [17-19]. When the lesion extended beyond these caudal limits, a transoral route or a combined endonasal-transoral approach was considered, as previously reported [20,21]. If the lesion extended laterally beyond the cervical internal carotid artery, a transcervical approach was selected because the transnasopharyngeal approach was considered insufficient. Pharyngeal mucosal incision types included inverted U-shaped or linear incisions, selected based on the lateral extent of the lesion. A linear incision was used for lesions confined to the midline, while an inverted U-shaped incision was selected when broader lateral exposure was required.
2) Airway management
All patients underwent preoperative evaluation for potential airway difficulty. Anesthesiologists conducted the assessments using standardized scoring systems [22]. Particular attention was given to patients with upper airway obstruction caused by a tumor, prior occipitocervical fixation, pediatric status, or symptoms indicating lower cranial nerve dysfunction, such as recurrent laryngeal nerve palsy—all of which may increase the risk of difficult intubation or ventilation. Although tracheostomy is generally not required for the endoscopic endonasal transnasopharyngeal approach, it was performed preoperatively in selected high-risk patients to secure the airway. In other patients assessed as having high airway risk, awake or fiberoptic-assisted intubation was planned.
3) Occipitocervical instability
Occipitocervical instability was evaluated using clinical findings and imaging studies. Clinical indicators, including neck pain, a sensation of head instability, and signs of myelopathy were assessed. Dynamic flexion-extension cervical radiographs were obtained to detect significant changes in the atlantodental interval. Computed tomography (CT) was used to assess bony involvement or destruction at the atlanto-occipital and atlantoaxial joints [23]. Magnetic resonance imaging (MRI) was used to evaluate potential ligamentous injuries, particularly of the transverse ligament and tectorial membrane. In patients with prior occipitocervical fixation, CT was used to determine whether bony fusion had been completed. When instability was already present or anticipated owing to lesion resection, occipitocervical fixation was performed before the surgery. Ideally, the lesion was removed after confirming complete bony fusion; however, in cases involving rapidly growing tumors or severe myelopathy, surgery was undertaken before fusion was complete, with continued use of a halo vest for external stabilization during the perioperative period.
3. Surgical Technique
1) Operative setup
The surgeon and the endoscope-holding assistant were positioned on the right side of the patient, while the scrub nurse stood on the left. The assistant was seated in front of the surgeon and maintained control of the endoscope throughout the procedure. The operating table was elevated to place the patient in a vertexup position, facilitating caudal exposure. Neuronavigation was routinely prepared, and 4-mm endoscopes with both straight and angled lenses were used. Spinal drainage was selectively performed after inducing general anesthesia in cases at high risk of postoperative cerebrospinal fluid leakage, following previously described strategies [24]. The surgical field was prepared to allow harvesting of autologous fascia lata or fat grafts for reconstruction. All procedures were jointly performed by neurosurgeons and otolaryngologists using a 3- or 4-hand technique. Otolaryngologists performed the nasal procedures, while neurosurgeons conducted the subsequent surgical steps. For the endonasal approach, gauze soaked in 1:5,000 diluted epinephrine was inserted into the nasal cavity to induce mucosal decongestion before dissection.
2) Transnasopharyngeal procedure and lesion exposure
The posterior pharyngeal wall was exposed through the interseptal corridor between the nasal septal mucosae [25]. A Killian incision was made on one side of the nasal septum, and a horizontal posterior-superior mucosal incision was made on the opposite side [26]. The interseptal space was dissected, and the anterior wall of the sphenoid sinus was widely opened [27,28]. The pharyngeal mucosa over the floor of the sphenoid sinus and upper clivus was inferiorly dissected using monopolar cautery, followed by an incision of the posterior edge of the nasal septal mucosa to access the posterior pharyngeal wall (Fig. 2A). To enhance caudal visualization, the soft palate was retracted anteriorly using a Nelaton catheter, as it might otherwise obstruct the surgical view.
The posterior pharyngeal wall was incised using monopolar cautery. Incision types included inverted U-shaped and midline linear designs (Fig. 2B). When an inverted U-shaped incision was used, the lateral edges were placed just medial to the torus tubarius to avoid injury to the eustachian tubes [29]. The incised posterior pharyngeal wall, along with the underlying longus capitis muscle, rectus capitis anterior muscle, longus colli muscle, anterior atlanto-occipital membrane, and anterior longitudinal ligament, was dissected off the lower clivus and upper cervical spine in a caudolateral direction, creating sufficient working space. When the lesion involved the parapharyngeal space, it was exposed at this stage. Depending on the lesion location, instrument maneuverability was sometimes restricted. For laterally extending lesions, a medial maxillectomy was performed to overcome this limitation, and in some cases a transmaxillarypterygoid approach was also used [30,31]. For lesions extending toward the jugular tubercle or petrous apex, a far medial or translacerum approach was used for resection [32,33].
Intraoperative venous bleeding, particularly from the interdural space, was controlled either by applying flowable sealants or fibrin glue to fill the bleeding site, or by constructing a barrier with oxidized cellulose sheets. These methods allowed effective hemostasis while preserving visualization of the surgical field.
3) Reconstruction and closure
The focus of the reconstruction of the ventral craniovertebral junction was on preventing cerebrospinal fluid leakage and postoperative infection. When dural incisions or resections were made intraoperatively, primary dural closure was performed by suturing autologous fascia lata or abdominal fat grafts to the dural edge using 7-0 Prolene [24]. Dead space was minimized by packing the defect with fibrin glue and/or fat grafts. The mucosal flap was repositioned to its original location. In several cases, the pharyngeal mucosa was additionally sutured in place using absorbable sutures. In cases with a high risk of cerebrospinal fluid leakage, anticipated postoperative radiation therapy, or an exposed internal carotid artery, a pedicled nasoseptal flap was used to reinforce the repair [34]. When the reconstructed tissue tended to float or detach, a pharyngeal balloon was temporarily placed to maintain compression and support graft adherence.
4. Postoperative Management
1) Airway and nutritional support
Because of the potential risk of postoperative upper airway obstruction caused by prevertebral hematoma, the feasibility of extubation was carefully evaluated. Extubation was performed promptly once the airway was deemed stable. In patients with tracheostomy, the tracheostomy tube was removed once airway patency was confirmed. Oral intake was gradually initiated following fiberoptic evaluation of velopharyngeal insufficiency risk and swallowing function. Until oral feeding could be resumed, enteral nutrition was provided via an oral or nasal feeding tube inserted under fiberoscopic guidance to prevent inadvertent entry into the cranial cavity. Postoperatively, patients performed daily nasal irrigation, and regular otolaryngology follow-up included thorough cleaning of the nasal and pharyngeal cavities. Prophylactic intravenous antibiotics were administered perioperatively. In cases without intraoperative cerebrospinal fluid leakage, cefazolin was administered until postoperative day 1. When cerebrospinal fluid leakage occurred intraoperatively, ceftriaxone was administered instead and continued for several days.
2) Occipitocervical stabilization
In patients who underwent occipitocervical fixation preoperatively, stability was assessed to ensure the maintenance of fusion and alignment. In patients without prior fixation, careful monitoring was conducted to detect signs of new-onset neck pain suggestive of occipitocervical instability, and cervical motion was minimized during the early postoperative period. Postoperative CT was used to evaluate the atlanto-occipital and atlantoaxial joints for signs of articular damage. When joint destruction was suspected, cervical immobilization with a neck collar was applied.
RESULTS
1. Summary of Clinical Cases
Seven consecutive patients with ventral craniovertebral junction lesions underwent surgery using the endoscopic endonasal transnasopharyngeal approach, as summarized in Table 1. Patient age ranged from 4–73 years. Six patients had neoplastic lesions, including chordoma (n=3), Langerhans cell histiocytosis (n=1), adenoid cystic carcinoma (n=1), and recurrent foramen magnum meningioma (n=1). One patient had a nonneoplastic lesion—basilar invagination—which was treated with odontoidectomy. Four patients had a history of prior surgical interventions, such as craniotomy or endoscopic endonasal surgery. Among the neoplastic lesions, the mean maximum diameter was 39.4 mm, and the mean lateral extension from the midline was 26.1 mm. Intracranial extension was observed in 4 cases, intraspinal extension in 2 cases, and parapharyngeal space extension in 3 cases. The caudal margin of the lesion reached the lowest margin of the clivus in 2 cases, the inferior margin of the anterior arch of the atlas in 2 cases, the base of the odontoid process of the axis in 1 case, and the inferior margin of the axis body in 1 case. Occipitocervical fixation was performed preoperatively in 2 patients, and tracheostomy was performed in 2 patients. One patient underwent prophylactic tracheostomy due to tumor-induced upper airway obstruction, while the other had preexisting severe lower cranial nerve dysfunction, warranting airway protection. Pharyngeal mucosal incisions included inverted U-shaped (n=5) and linear (n=2). Among the tumor cases, gross total resection was achieved in 1 case, subtotal (90%–99%) in 4 cases, and partial (<90%) in 1 case. In the 4 cases of subtotal resection, the residual tumor was located adjacent to the cervical internal carotid artery (case 2), adherent to the brainstem (cases 3 and 4), or infiltrating the petrous bone (case 5), and in the partial resection case (case 6), it remained adherent to both the brainstem and the vertebral arteries. Reconstruction with a nasoseptal flap was performed or reused in 3 cases. Severe intraoperative cerebrospinal fluid leakage occurred in 3 cases; however, no postoperative cerebrospinal fluid leakage was observed. Among the 5 orally intubated patients, 3 were immediately extubated after surgery, and 2 on postoperative day (POD) 2. Oral feeding was initiated by POD 10 in all but 1 patient with preexisting severe lower cranial nerve dysfunction. One patient who underwent odontoidectomy experienced transient worsening of preexisting dysarthria, which resolved within 1 month. Postoperative radiotherapy was performed in all tumor cases in which gross total resection was not achieved, including proton beam therapy, stereotactic radiosurgery or radiotherapy, and intensity-modulated radiation therapy. The follow-up period ranged from 4 to 42 months. The Karnofsky Performance Status (KPS) recovered to 90–100 in all cases except for 1 patient with recurrent chordoma (case 3). In this patient, the KPS improved from 50 to 70 immediately postoperatively but tumor regrowth occurred at 6 months postoperatively, and the patient died of disease at 10 months.
2. Representative Case
A 5-year-old girl, listed as case 2 in Table 1, presented with progressive snoring due to upper airway obstruction. Imaging revealed a tumor extensively involving the lower clivus, with intracranial extension causing the brainstem compression and lateral extension into the parapharyngeal space. The mass abutted the cervical internal carotid artery without encasing it. Preand postoperative MRI findings are shown in Fig. 3A–D. She had previously undergone posterior fossa decompression and biopsy at another institution. Occipitocervical instability was not observed preoperatively; however, it was anticipated following tumor resection. Therefore, occipitocervical fixation was performed before tumor resection. Owing to the high risk of airway compromise, tracheostomy was also concurrently performed with occipital to C2–3 posterior fixation. Given the rapid growth of the tumor, resection was conducted 10 days later without waiting for complete bony fusion. A halo vest was used perioperatively for external stabilization.
The surgical steps are illustrated in Fig. 4 and Supplementary Video Clip 1. As the tumor extended caudally into the parapharyngeal space, the procedure was performed via the endonasal route, with preparations in place for simultaneous transoral access if needed. Bilateral endoscopic medial maxillectomies were performed to enhance lateral and caudal maneuverability. The space between the bilateral nasal septal mucosa was dissected, and the anterior wall of the sphenoid sinus was opened (Fig. 4A). The posterior edge of the nasal septal mucosa was subsequently incised to expose the posterior pharyngeal wall (Fig. 4B). The vomer bone was removed, and the nasopharyngeal mucosa was dissected off the clivus. Lateral incisions were made just medial to the torus tubarius to create an inverted U-shape pharyngeal mucosal flap (Fig. 4C and D). Elevation of the flap, exposed the entire ventral surface of the clivus and the parapharyngeal component of the tumor (Fig. 4E). After resection of the parapharyngeal portion, the lesion extending to the lateral aspect of the dural sac was removed (Fig. 4F). The periosteal dura was disrupted; however, the meningeal dural layer remained intact, and no cerebrospinal fluid leakage occured intraoperatively. Because the tumor was invasive and poorly defined from the surrounding soft tissue, a small portion adjacent to the left cervical internal carotid artery was intentionally left in place. Subtotal resection of the tumor was achieved while preserving the atlantooccipital and atlantoaixial joints (Fig. 4G). The mucosal flap was subsequently repositioned and reinforced with a sealant (Fig. 4H).
No postoperative complications occurred. Oral intake was initiated on POD 10, and the tracheostomy tube was removed 3 months after surgery. The patient subsequently underwent adjuvant proton beam therapy and has remained recurrencefree for 3 years.
DISCUSSION
1. Surgical Indications and Technical Considerations
Traditionally, ventral craniovertebral junction lesions have been managed using microscopic transoral or transcervical approaches. The transoral route, while providing a direct midline access to the anterior craniovertebral junction, necessitates incision of the oropharyngeal mucosa, which is associated with prolonged intubation, delayed resumption of oral intake, velopharyngeal insufficiency, and an increased risk of mediastinitis [3,6]. In some cases, more invasive procedures—such as labiomandibular or palatal splitting, maxillary down-fracture, or glossotomy— are necessary to ensure adequate exposure [11,12]. An alternative approach is the transcervical method, which offers a sterile operative field but is limited in its angle of attack by anatomical constraints such as mandibular position and thoracic height [7-10]. With the advancement of neuroendoscopy and an improved understanding of microsurgical anatomy, the endoscopic endonasal approach has gained wider application for ventral craniovertebral junction lesions [3,5,13,14]. Although our present series is limited by the small number of patients (n=7), which inevitably restricts the generalizability of the findings, it demonstrates the feasibility and potential advantages of the endoscopic endonasal transnasopharyngeal approach. In our series, this approach involved an incision at the posterior end of the nasal septal mucosa to access the nasopharynx, followed by an incision of the nasopharyngeal mucosa to reach the ventral craniovertebral junction. This approach provides excellent visualization of the surgical field because of the close-up and panoramic views afforded by the endoscope. In addition, avoiding incisions through the oropharyngeal mucosa may reduce surgical morbidity compared with that of the transoral route [35]. In our series, the tumors were relatively large, with a mean maximum diameter of 39.4 mm. However, subtotal or greater resection (>90%) was achieved in 5 of the 6 tumor cases, excluding the meningioma case.
While this approach offers favorable resectability, a known anatomical limitation is its restricted inferior reach. Several anatomical lines have been proposed to estimate the caudal limit of endonasal exposure, including the nasopalatine, nasoaxial, and rhinopalatine lines [17-19]. In our series, approach selection was tailored to the anatomical extent of each lesion. In the representative case (case 2 in Table 1), the parapharyngeal component appeared to extend beyond the theoretical inferior limit predicted by these lines; therefore, we prepared to add an endoscopic transoral route intraoperatively. Adequate exposure was ultimately obtained through the endonasal transnasopharyngeal corridor alone. In cases 6 and 7, the caudal extent reached the level of the base of the odontoid but remained rostral to the predicted limit; thus, an endonasal only approach was planned from the outset. These case based decisions underscore that these lines should be regarded as practical guides rather than absolute boundaries [36]. One method to overcome this caudal limitation is removal of the posterior nasal spine, which has been reported to facilitate more inferior access [37]. If further caudal exposure is required, an endoscopic transoral approach—or a combined endoscopic endonasal and transoral approach—should be considered [20,21]. Regarding lateral extension, lesions extending beyond the cervical internal carotid artery may lie outside the reach of the endonasal corridor. In such cases, alternative approaches including the transcervical route or transoral approach with lateral mucosal incision may be more appropriate [38].
2. Potential Complications and Management Strategies
Several potential complications are associated with the endoscopic endonasal transnasopharyngeal approach. Perioperative airway management is critical to prevent life-threatening complications such as upper airway obstruction. As this approach does not require mucosal incision of the oropharynx, the risk of postoperative airway obstruction due to pharyngeal mucosal swelling is considered lower than that associated with the transoral approach [14]. Nevertheless, particular caution is warranted in cases involving parapharyngeal tumor extension, as exemplified in our representative case. Pediatric patients and those who have undergone occipitocervical fixation should similarly be recognized as being at higher risk of airway compromise; in such cases, prophylactic tracheostomy should be considered. In our series, 2 patients underwent prophylactic tracheostomy: one was the representative case described above, and the other had severe lower cranial nerve dysfunction caused by tumor invasion and underwent tracheostomy immediately before surgery. Of the 5 patients who were orally intubated, 3 were extubated immediately after surgery, and 2 by POD 2. Regarding postoperative nutritional management, it has been suggested in previous studies that oral feeding can be resumed earlier following endonasal surgery compared with after transoral procedures [4,14]. In our cohort, oral feeding was resumed by POD 10 in 6 of the 7 cases, with a median initiation on POD 4. To prevent wound infection and mediastinitis, secure reconstruction of the surgical site is essential. One technical limitation of this approach is the difficulty in achieving tight mucosal suturing. However, as the oropharyngeal mucosa—part of the alimentary tract—is not incised, tight suturing may be less critical. No infectious complications were observed in our series. One patient who underwent odontoidectomy developed transient velopharyngeal insufficiency with mild dysarthria, which resolved spontaneously. This may have been caused by soft palate dysfunction resulting from balloon compression of the posterior pharyngeal wall. While minimizing dead space and ensuring mucosal apposition is important, limiting the duration of balloon compression may help preserve soft palate function.
Lesions involving the craniovertebral junction may be associated with preexisting occipitocervical instability, or such instability may develop following lesion resection. In these situations, occipitocervical fixation is required. However, there is currently no consensus regarding the optimal timing of fixation in relation to lesion resection; fixation may be performed either concurrently with or before resection [15,16]. When fixation is performed following lesion resection, potential risks include surgical site infection due to prolonged operative time, cerebrospinal fluid leakage in the prone position, and spinal cord injury during intraoperative repositioning in cases of instability [15]. To minimize these risks, some authors recommend performing occipitocervical fixation before lesion resection, proceeding with tumor removal only after confirming bony fusion [35,39,40]. However, as demonstrated in our representative case, rapid tumor progression may necessitate early resection before fusion has been achieved. Limitations of occipitocervical fixation include restricted neck extension, which may increase the risk of difficult airway management, and the presence of instrumentation-related imaging artifacts. Nonetheless, such artifacts are generally not problematic for visualizing ventrally located lesions.
Beyond these perioperative considerations, this approach also involves a demanding learning curve. Surgeons must acquire familiarity with the endoscopic anatomy of the craniovertebral junction and gain proficiency in working through the narrow nasopharyngeal corridor. Technical proficiency improves with accumulated experience, but careful case selection and institutional expertise are crucial during the early phase of adopting this technique.
This study was limited by its small sample size and retrospective design. Analyses involving larger cohorts will be necessary to more accurately evaluate the utility of this approach and to better characterize the potential complication profile.
CONCLUSION
The endoscopic endonasal transnasopharyngeal approach provides a safe and effective route to ventral craniovertebral junction lesions, offering enhanced visualization and reduced invasiveness. In our series, satisfactory lesion resection and early postoperative recovery were achieved through appropriate case selection and multidisciplinary planning. While limitations remain—particularly in caudal access and secure mucosal closure— these challenges can be mitigated with refined anatomical knowledge and technical modifications. Careful management of the airway and spinal stability is essential. This approach may broaden surgical options for selected cases, and further studies are warranted to clarify its indications and outcomes.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Supplement
Print
