Effectiveness of preoperative facial nerve diffusion tensor imaging tractography for preservation of facial nerve function in surgery for large vestibular schwannomas: Results of a prospective randomized study
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.253631
Source of Support: None, Conflict of Interest: None
Keywords: Diffusion tensor imaging, facial nerve preservation, vestibular schwannoma
Large vestibular schwannomas (VS) pose a unique problem to the neurosurgeons., As these tumors grow, they cause distortion and compression of other nearby structures including the facial nerve and brainstem, resulting in cranial nerve dysfunction. Postoperative facial nerve injury is one of the most important causes of morbidity following resection of large vestibular schwannomas. Facial nerve injury can lead to a decreased quality of life postoperatively and ultimately requires other surgical interventions to alleviate the consequences of facial paralysis. Knowing the location and the course of the facial nerve preoperatively can lead to safer resections by reducing the incidence of inadvertent injury to the facial nerve intraoperatively.,
Standard preoperative imaging typically comprises T1-weighted images obtained before and after contrast administration, T2-weighted images, and a cisternography image. However, these are not useful in determining the location of the facial nerve in large VS. This is due to the fact that the signal characteristics of the nerve and tumor capsule are similar on T1- and T2-weighted images, as also the intimate association of the nerve fibers to the tumor capsule and compression of the facial nerve due to the large tumor. Various studies have attempted to address this issue and possibilities of integration of various modalities of imaging such as diffusion tensor imaging.,,
Diffusion tensor imaging is a newer imaging modality which measures the direction of diffusion of water molecules by combining multiple diffusion weighted image scans constructed in multiple directions. The diffusion of water molecules is anisotropic inside white matter fiber tracts and hence maximal along the direction of nerve fiber. By this method, we can reconstruct a nerve fiber tract between two regions of interest. Three dimensional (3D) diffusion tensor imaging has been widely used to construct white fiber tracts such as pyramidal tract, optic radiation, arcuate fasciculus, etc.,, Diffusion tensor imaging can be used to the determine the position of facial nerve accurately in relation to the tumor. By knowing the position of facial nerve preoperatively, the surgeon can modify his/her intraoperative strategies to better preserve facial nerve function.
[Table 1] shows the results of various studies regarding prediction of facial nerve position with DTI tractography.,,,,,,,,,,,, These studies have reported a concordance of 80% to 100% between preoperative diffusion tensor imaging and intraoperative findings.,, The senior investigator (SAB) had previously carried out a study to establish the validity of preoperative DTI tractography for prediction of facial nerve position in large vestibular schwannomas (>3 cm). The study concluded that there is 89% concordance between the DTI tractography-predicted facial nerve position and the intra-operative position of facial nerve.
The authors now aim to translate this knowledge to improve the clinical outcome. In this prospective randomized study, the investigators have recruited 100 consecutive patients with a large vestibular schwannoma (>3 cm) and undergoing surgery, into two groups using a computer generated randomization chart – group I (DTI tractography was done) and group II (DTI tractography was not done). The operating surgeon was informed of the facial nerve position as predicted by the DTI prior to surgery. The facial nerve preservation rates were compared between both the groups.
In this prospective randomized study, we attempted to assess if knowing facial nerve position preoperatively using high definition DTI tractography could translate into better facial nerve preservation rates during surgery for large vestibular schwannomas. The study was registered under Clinical Trial Registry of India (CTRI/2017/04/008345).
After obtaining the Institute Ethics Committee (IEC) approval (IECPG-57/2016) and after obtaining a written informed consent from the patient and/or his relative, 100 consecutive patients of either gender with a large vestibular schwannoma (>3 cm) undergoing surgery, were recruited in this study. We excluded patients in whom there was computerized tomographic (CT) evidence of calcification or haemorrhage as these are known to interfere with DTI tractography. After exclusion, 94 patients were randomized into two groups using a computer generated randomization chart – group I (DTI tractography done) and group II (DTI tractography not done). The intent of surgery was complete excision of the tumor and preservation of facial nerve function. The operating surgeon was informed about the DTI tractography predicted facial nerve position prior to surgery. The facial nerve preservation rates between the two groups were compared.
Preoperative T1W, T2W and contrast MRI were carried out to know the size of the tumor, associated brainstem compression and hydrocephalus, etc., in all patients. Diffusion tensor imaging sequences (using 32 channel head coil, no intersection gap) was acquired with a 3-T MR imaging scanner in group I. The MR imaging data was stored in a digital imaging and communication in medicine (DICOM) file and transferred via network for image processing using DTI software. The automatic fusion algorithm of the software was used to merge the DTI imaging and conventional MR images. Then the image fusion was verified for accuracy and the fiber tracking process was performed. Multiple regions of interest were used to depict the fibers of the facial nerve. The tracking algorithm was initiated using two regions: the internal auditory meatus and facial nerve entry area at the brainstem. Any of the tracts that were found to be in different locations than that predicted by previous anatomical data, such as those crossing the midline and those ascending or descending in the brainstem, were filtered (eliminated). The operating surgeon was informed about the DTI tractography-predicted position of the facial nerve. A comparative analysis was then made during operation. The location of the facial nerve in relation to the tumor was recorded during surgery using facial nerve stimulator (NIM Eclipse Medtronic, USA), which records electromyographic responses of facial nerve innervated muscles. Nerve location was categorized using the classification proposed by Sampath et al., (anterior, posterior, superior, or inferior to the tumor). Anterior and posterior locations were further subdivided into 3 subcategories: upper, middle, and lower thirds. The facial nerve preservation rates were compared between the two groups. The goal of surgery was always complete tumor removal with preservation of facial nerve. Total removal of tumor was attempted in all cases. A subgroup analysis was also done depending upon the level of experience of the surgeons. The clinical outcome of surgery in terms of facial nerve preservation was recorded at the time of discharge and follow up period of at least 3 months. The outcomes in the two groups were compared.
A total of 100 patients were recruited in the study. After exclusion, a total of 94 patients were randomized. In group I (DTI group), there were 47 patients, and in group II (DTI not done), there were 47 patients. The flow chart of this study protocol is shown in [Figure 1]. Out of the 47 patients in group I in which DTI tractography was done preoperatively, it was not possible to identify facial nerve in 5 patients (technical failure). The technical failure in five patients was due to software malfunction. We could not fuse the anatomical and diffusion tensor images. It was not related to tumor characteristics. It was possible to preoperatively identify facial nerve in rest of the 42 cases.
All tumors in both the groups were more than 3 cm in dimension. The mean tumor volume was 27.14 ± 14.3 ml. The tumor volume in Group I and Group II was 25.4 ± 13.0 ml and 28.6 ± 13.3 ml, respectively. There was no statistically significant difference between the two groups in terms of sex distribution, age of patient, presence of neurofibromatosis type II, hydrocephalus and tumor volume. The summary of the demographic profile of the study population has been shown in [Table 2].
There were 51 patients who had hydrocephalus, of whom 22 patients required a ventriculoperitoneal shunt surgery prior to definitive surgery. There were a total of 9 patients with neurofibromatosis type II who were almost equally distributed in both the groups. [Figure 2] shows the MRI images of patients with neurofibromatosis type II.
All patients presented with sensorineural hearing loss as the primary symptom. Preoperative facial nerve function was graded based on the House-Brackmann scale. Most (70.9%) patients had grade II facial nerve paralysis at presentation followed by grade I (19.7%) paresis. The difference in the distribution of the preoperative facial nerve paralysis grades in both the groups was not statistically significant.
In Group I, DTI tractography was carried out preoperatively to ascertain the position of facial nerve. The location of the nerve was categorized into anterior superior, anterior middle, anterior inferior and posterior, based upon the classification proposed by Sampath et al. The most common location was anterior superior in 24 (60%) followed by anterior middle in 10 (25%) patients [Table 3]. There were 36 patients in whom the facial nerve was compact, and 4 patients in whom the facial nerve was thinned out and splayed over the surface of the tumor. [Figure 3] and [Figure 4] show the DTI reconstructed images of the course of the facial nerve in relation to the tumor.
The DTI predicted facial nerve position was informed preoperatively to the operating surgeon on the day of surgery in all these patients. Out of the 47 patients in whom DTI was carried out, it was not possible to preoperatively identify facial nerve in 5 patients (technical failure). In 2 patients of Group I, surgery was not done. Out of these 2 patients, one had a large multinodular goitre and underwent total thyroidectomy. She had tracheomalacia and symptomatic hypocalcemia and hence surgery was deferred. The second patient did not consent for surgery. In group II patients, preoperative DTI tractography was not done. One patient was not operated as the patient did not give consent for surgery, leaving 46 patients for the final comparison.
Out of 40 patients available for analysis in the DTI group, the preoperative DTI predicted facial nerve position was concordant with the intraoperative position of facial nerve in 39 cases (97.5% concordance). In one patient, DTI predicted position of facial nerve was in anterior middle one-third of the tumor. However, the facial nerve was not identified intraoperatively. This was confirmed by talking to the operating surgeon at the end of the surgery. It was possible to preserve the facial nerve anatomically and functionally in 36 out of 40 cases in this group (90%). Facial nerve was anterosuperior to the tumor in 24 patients, anterior-middle in 10 patients, anteroinferior in 5 patients and posterior to tumor in one patient. Facial nerve could not be preserved in two patients in each anterosuperior and anterior-middle group (P = 0.6).
In group II, comprising of 46 patients in whom preoperative DTI tractography was not done, it was possible to preserve the facial nerve anatomically and functionally in 29 (63.0%)cases. The difference between the two groups was found to be significant (P value = 0.002). [Table 4] shows the details of facial nerve preservation rates in both the groups.
We did a subgroup analysis depending upon the level of experience of the surgeon (<2 years, 2-5 years, 5-10 years and more than >10 years) and the facial nerve preservation rates were not statistically different among the different subgroups. This was probably due to the small number of patients operated by surgeons with experience less than 5 years (P value- 0.98). [Table 5] shows the impact of the surgeon's experience in achieving preservation of the facial nerve.
The operating surgeons felt that the preoperative knowledge of the position of the facial nerve helped them in modifying their intraoperative strategies and dissection, leading to better facial nerve preservation rates in group I.
The patients were followed up on a regular basis. The preservation rates of facial nerve were maintained in the follow up period. The mean follow up period was 7.9 months with the minimum follow up period of 3.2 months and the maximum follow up period of 12.8 months. [Figure 5] shows the preoperative and postoperative contrast enhanced images in the DTI group, which shows total tumor removal with facial nerve preservation.
The goal of modern day surgery for vestibular schwannomas is complete tumor removal as well as preservation of neurological functions and quality of life., Preservation of the facial nerve during surgery is the key determinant of quality of life in the postoperative period. However, preservation of facial nerve becomes difficult as the size of the tumor increases. The nerve in its CPA segment is displaced by the expanding tumor in an unpredictable manner and is often changed morphologically, particularly in large tumors. Identification of the facial nerve is very important during surgery for its preservation. In large VSs, however, the preservation of the facial nerve may be challenging. Early identification of the facial nerve during surgery, based on its relationship to anatomical landmarks and on the application of electrical stimulation and electromyographic monitoring are among the main prerequisites for its preservation. In large tumors, facial nerve sometimes is splayed and thinned out over the surface of tumor, which makes it difficult to identify the nerve intraoperatively. Differentiation of wall of the cystic tumor and the thinned out facial nerve is also challenging during surgery. In a large surgical series, the most common position of facial nerve in relation to the tumor was anterosuperior followed by anterior middle one third of the tumor. Therefore, an imaging-based technique that demonstrates the location of the nerve in relation to the VS preoperatively could be useful to the surgeon.
As the VSs enlarge and compress the facial nerve, their preoperative identification using standard MR imaging becomes increasingly difficult. The facial nerve gets stretched thin and can be intimately associated with the tumor, and often the nerve is indistinguishable from the tumor. Standard preoperative MRI techniques (T1, T2, T1 contrast) have provided two-dimensional resolution of distinct structures but have lacked the ability to adequately demonstrate spatial relationships. The more recent MR imaging modalities, such as DT imaging, have allowed mapping of the cranial nerves in healthy individuals. The difference between the fractional anisotropy (FA) values of facial nerve and tumor permits us to distinguish between the nerve and the VS. Using FA values, facial nerve can be traced from the brainstem to the internal auditory canal. Taoka et al., applied these DT imaging techniques to patients with VSs, and the authors reported that they were able to identify the location of the facial nerve in 7 out of 8 patients. Gerganov et al., in a study of 22 patients with a large VS concluded that the position of the facial nerve in relation to the tumor can be predicted reliably (in 91% patients) using DT imaging–based fiber tracking. Roundy et al., in a prospective series of 5 patients with a large VS were able to identify the position of facial nerve in all patients using high definition DT imaging (HD-DT). Yoshino et al., (2015) and Borkar et al., (2016) also concluded that DT imaging had good concordance with intraoperative findings. These studies concluded that there was good concordance between DTI tractography predicted facial nerve position and the intra-operative position of the facial nerve. DT imaging is particularly useful in large cystic tumors and in cases where facial nerve is splayed and thinned out.
The facial nerve preservation is a very important aspect of VS surgery. Facial nerve paresis is a devastating complication and has an adverse psychosocial impact. In this prospective randomized study, we have attempted to assess whether or not knowing facial nerve position preoperatively using DTI tractography could translate into better facial nerve preservation rates and better functional outcome in surgery for large VSs. Our study showed that there was a good concordance of DTI tractography predicted facial nerve position and intraoperative position of the facial nerve (97.5%). The facial nerve preservation rates also improved in these patients when compared to patients in whom DT imaging was not done. The reliable preoperative visualization of the facial nerve location in relation to the VS has allowed surgeons to plan the tumor removal accordingly. This has increased the safety of surgery. This was also suggested by near-similar facial preservation rates across surgeons with varied levels of experience. The clinical outcomes were maintained even in the follow up period. We conclude that that DTI tractography for facial nerve is a powerful, accurate, and rapid method for preoperatively identifying the facial nerve in relation to large VSs and should be a routine investigative modality in all such cases.
The results of this prospective randomized study established the role of preoperative DTI tractography for better facial nerve preservation in surgery for large vestibular schwannomas (>3 cm). DTI tractography for facial nerve should be a routine investigative modality in all cases of large vestibular schwannomas for ensuring a better facial nerve preservation, thereby achieving a better surgical outcome.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]