- Email: [email protected]
Surgical Management of “Kissing” Spinal Plexiform Neurofibromas in Neurofibromatosis Type 1 Patients
Journal Pre-proof Surgical management of “kissing” spinal plexiform neurofibromas in NF1 patients Ben Shofty, Michal Mauda-Havakuk, Liat Ben-Sira, Felix Bokstein, Zvi Lidar, Khalil Salame, Akiva Korn, Shlomi Constantini PII:
S1878-8750(19)32964-X
DOI:
https://doi.org/10.1016/j.wneu.2019.11.124
Reference:
WNEU 13793
To appear in:
World Neurosurgery
Received Date: 6 August 2019 Revised Date:
20 November 2019
Accepted Date: 21 November 2019
Please cite this article as: Shofty B, Mauda-Havakuk M, Ben-Sira L, Bokstein F, Lidar Z, Salame K, Korn A, Constantini S, Surgical management of “kissing” spinal plexiform neurofibromas in NF1 patients, World Neurosurgery (2019), doi: https://doi.org/10.1016/j.wneu.2019.11.124. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Surgical management of “kissing” spinal plexiform neurofibromas in NF1 patients Ben Shoftya,b, Michal Mauda-Havakukb, Liat Ben-Sirab, Felix Boksteinb, Zvi Lidara, Khalil Salamea, Akiva Korna, and Shlomi Constantinib,c 1. Department of Neurosurgery 2. The Gilbert Israeli Neurofibromatosis Center 3. Department of Pediatric Neurosurgery All at the Tel-Aviv Medical Center, and Tel-Aviv University, Tel-Aviv, Israel Correspondence to: Prof. Shlomi Constantini, MD, MSc, Department of Pediatric Neurosurgery, Tel-Aviv Medical Center, 6 Weizmann St. Tel-Aviv, Israel. [email protected] Highest Academic Degree: Ben Shofty MD, PhD, Michal Mauda-Havakuk MD, PhD, Liat Ben-Sira MD, Felix Bokstein MD, Zvi Lidar MD, Khalil Salame MD, Akiva Korn MSc, and Shlomi Constantini MD, MSc Keywords: Neurofibromatosis, intradural, neurofibroma, NF1, kissing Short title: “kissing” spinal plexiform neurofibromas in NF1
Abstract Introduction “Kissing” neurofibromas (KN) are a unique group of spinal tumors found in neurofibromatosis type 1 (NF1) patients. These are bilateral neurofibromas that approximate each other at the same level, with significant impingement compression of the cord or thecal sac. The best management options and surgical strategies for NF1 patients with KN have not been standardized.
Methods We conducted a retrospective study evaluating adult NF1 patients with KN. All patients are followed routinely at the Gilbert Israeli NF Center. Patients files were reviewed for natural history, imaging features, surgical technique, and surgical outcome.
Results Twelve patients with at least 1 pair of KN were identified (6 females). Median age at spinal presentation was 24 (range 17-48). KN were located at the cervical (n=8) and lumbar (n=8) region, with no thoracic involvement. Seven of the 12 patients were operated; all underwent surgery due to cervical compression with progressive myelopathy. Four patients remained asymptomatic during the follow-up period. Three patients underwent multiple operations. Operative outcome was favorable in 71% of patients, with marked overall motor improvement or stabilization of neurological deterioration. Two patients that entered surgery with a low functional reserve deteriorated following surgery.
Conclusions In our series, KN caused progressive cord compression in 7 of the 8 patients with cervical tumors. No intervention was needed for lumbar tumors. Cervical tumors should be followed closely, with a low-threshold for intervention. NF1 patients harboring KN should be followed both clinically and radiologically for life.
1
Background Neurofibromatosis type 1 (NF1) is a common neurocutaneous syndrome that affects ~1:2000 individuals globally1,2. This syndrome causes tumors along the nervous system, including peripheralnerve-sheath tumors, neurofibromas, that often cause pain and neurological deficits, and may undergo malignant transformation3. Spinal involvement varies widely. Included in the spinal phenotype are structural abnormalities such as vertebral anomalies, dural ectasia, and various spinal tumors4. Interestingly, intra-dural tumors are relatively rare in NF1, known to occur in only ~2-6% of patients, as most neoplastic involvements are para-spinal5, extending from the neuroforamina laterally. Anatomically, most spinal tumors in NF1 are foraminal (~57% of patients with spinal involvement), followed by intradural-extramedullary (33%). Intramedullary tumors occur rarely and are usually of glial origin. In our center, out of ~300 adult NF1 patients, only 34 required focused imaging for work-up of their spinal disease. NF1 is a disease with unpredictable progression patterns, and highly variable clinical course and tumor load. Two major patterns of spinal involvement have been defined in NF1: “spinal NF1” in which the disease is limited to the spine, with little, if any, systemic or cutaneous manifestations6, and NF1 in which there is a significant involvement of the spine, and the rest of the body is also affected7. Therefore, treatment should be individualized, requiring an experienced multidisciplinary team and a dedicated disease-focused center. Surgery is reserved for progressive tumors or those causing neurological deficit. Often, in patients with significant spinal tumor burden, it is difficult to locate the culprit tumor, and choosing the best surgical strategy may be challenging. Evidence in the literature regarding surgical management of spinal NF1 is scarce, mainly focusing on deformity correction 7,8. Currently, surgical treatment decisions for these complex patients are typically based on personal experience and multi-disciplinary team discussions. Multiple, bilateral neurofibromas in NF1 patients pose a special challenge for the treating neurosurgeon. In this paper, we focus on a unique group of spinal tumors we termed “kissing” neurofibromas – bilateral neurofibromas that cause significant pinching compression on the thecal sac and approximate each other. These tumors may often be responsible for clinical deterioration and necessitate surgical intervention. The best management options and surgical techniques for NF1 patients with kissing neurofibromas (KN) are still unknown, as the natural history and the surgical outcome of patients harboring these tumors has not been analyzed.
2
The aim of this report is to describe this unique group of patients, and our personal clinical and surgical experience.
Methods All NF1 patients in this study were followed and treated at the Gilbert Israeli International NF Center (GIINFC). We conducted a retrospective study that was approved by the institutional review board (IRB) of Tel-Aviv Sourasky Medical Center. Imaging of 34 patients with NF1 and spinal involvement were reviewed and analyzed for the presence of KN. KN were defined as bilateral foraminal-extramedullary tumors approximating each other at the same level, with significant compression of the thecal sac. Included in this study were all patients with KN and sufficient followup and imaging data. Clinical, radiological, and surgical data was extracted from the patients’ electronic medical records. Neurological course was determined according to the neurological examination as documented during subsequent clinic follow-up. Information about structural changes and tumor progression were based on sequential MR imaging and the report of the neuroradiologist at the NF1 center. All imaging of the included patients was reviewed by a radiology resident (Dr. Mauda-Havakuk), under the supervision of a senior neuro-radiologist (Dr. Ben-Sira). Clinical outcome such as neurological and functional status were based on serial reports of the treating neurologist.
Results Following retrospective review of adult NF1 patients with spinal disease (n=34), twelve patients were found to have at least one pair of KN and were included in this study. Six were females. Median age at diagnosis with KN was 24 years (range 17-41 years, mean 28±11 years). Five patients had positive familial history of NF1. Out of 8 patients who were symptomatic at presentation, 6 presented with motor symptoms that correlated with their spinal disease, and 2 patients presented with pain that led to diagnosis. Four patients were asymptomatic and were either diagnosed via routine screening or identified incidentally. Mean follow-up time was 48 months (±32 months, range 14-104 months). Anatomically, all patients in this series suffered from extensive spinal involvement, with a minimum of 3 spinal levels involved, and a maximum of 20. Of the 12 patients, 10 had cervical tumors, 4 of whom also had lumbar KN. Two patients had lumbar tumors only. Nine patients had anteriorly-kissing NF, and 3 had posterior tumors.
3
During the follow-up period, 7 patients experienced neurological deterioration associated with their KN, and were operated. All pathological reports were consistent with a WHO I neurofibroma. The other 5 patients were clinically and radiologically stable. Four of the neurologically stable patients were asymptomatic at presentation, 1 presented with tumor-associated pain. All of the 7 deteriorating patients underwent surgical resection of their culprit lesions. All of the surgeries were for progressive, compressing cervical tumors. No intervention was needed for lumbar KN. Six patients underwent more than one operation (a total of 16 surgeries in 6 patients), for progression of tumor at a different spinal level. Major surgical complications occurred in 1 patient, however, this patient entered the surgery with significant cord compression and minimal motor reserve and required prolonged postoperative mechanical ventilation. Two other patients suffered from wound related issues that resolved with conservative management. Postoperative hydrocephalus necessitating shunt placement occurred in 3/7 operated patients, without any evidence for postoperative CNS infection, bleed or other obvious mechanism. During the post-operative follow-up period, of the operated patients, 2 experienced worsening of their neurological condition, 3 patients stabilized, and two patients improved. Ultimately, in our series, the natural history of this group is unfavorable, as 7/12 patients experienced progressive neurological deterioration due to cord compression prior to neurosurgical intervention. This led to severe motor disability and even death from related complications in one patient. Of note, none of the patients in this series suffered from clinically significant postoperative deformity that required intervention.
Example Case A 36-year-old female, diagnosed with familial NF1 since birth. Known history of multiple cutaneous and sub-cutaneous neurofibromas, some of which underwent resections, with a growing overall disease burden, and a significant spinal disease. Presented with 4 months of cervical and thoracic mechanical back pain, with right C5-7 polyradiculopathy, and signs and symptoms of cervical myelopathy. Follow-up cervical MRIs demonstrated progressive bilateral “kissing” neurofibromas (KN) at C6-7 with significant cord compression (Figure 1). She then underwent C6-7 laminectomy and resection of the intradural component emerging from the sensory rootlet of C7 (see surgical technique in the following section). Postoperative course was uneventful and complete recovery was documented. Follow-up imaging demonstrated complete decompression of the cord at C6-7 with no residual 4
tumor (Figure 2 A+B), and small neurofibromas at C1-2 with no cord compression (Figure 2C, blue arrows). Three years later, the patient presented with recurrence of the cervical myelopathic symptoms, with progressive cord compression at the level of C1-2 due to bilateral KN. Rapid radiological and clinical progression over 4 months was noted. The patient underwent a second surgery with C1-2 laminectomy and intradural resection of the KN. Postoperative recovery was slow, with spastic weakness of the right hand, recovering fully following rehabilitation.
Surgical technique All patients were operated under general anesthesia in prone position, under full intraoperative electrophysiological monitoring. Following localization of the correct level, skin incision, fascial and muscle dissection, and laminectomy or laminotomy using the bone scalpel was done. Ultrasound is routinely used to identify those tumors that are not obvious on the surface prior to dural opening. The rest of the operation was done under the operative microscope. When an enlarged extradural component exists, it is usually debulked first from the two sides, followed by dural opening with identification of the dural orifice and careful microscopic dissection bilaterally, away from the cord. Spinal neurofibromas commonly demonstrate a dumbbell-shaped anatomy. This dumbbell appearance if often seen with both the extra- and intra-canalicular component, as well as the intraand extra-dural component. For those tumors kissing dorsally, the origin sensory rootlet is usually cut, and the tumor removed either with microdissection or with the ultrasonic aspirator. For those tumors kissing anteriorly, direct stimulation of the involved rootlet is performed, providing additional information used when deciding how radical a removal can be completed. Note that an anterior KN, in many cases, may create a continuity of tumor anteriorly to the cord, and therefore pose a significant technical challenge. In addition, radical removal of anteriorly-originating tumors may cause postoperative motor deficit. Thus, the surgeon should consider a more conservative intra-capsular removal, leaving the radical surgical approaches for dorsally-located tumors.
5
Discussion Progressive KN causing spinal cord compression in NF1 patients is an under-reported pathology that may have significant impact on the neurological status and quality of life of the patient. Here, we presented the first case-series focused on the surgical management of these unique tumors. Our study of 12 patients with KN raises several key points in the follow-up, surgical management, and postoperative care of these patients. In our center, more than 1400 NF1 patients are followed and treated by a dedicated, neurosurgicallyoriented multidisciplinary team. As the majority of the followed patients are pediatric, only 34/300 adult patients underwent clinically-indicated spinal imaging. Of this group, only 12 patients were diagnosed with at least one pair of KN. The relative rarity of this finding, occurring in 4% of adult NF1 patients, may explain the lack of publications dealing with this pathology. In our cohort, all of the patients were young adults at presentation, and progression was noted in 7/12 patients throughout the follow-up period of 4 years. It is important to notice that none of the incidentally diagnosed patients required any intervention or treatment during the follow-up period. In their report from 2007, Leonard et al. described cervical cord compression in 13 NF1 patients9. In their study, a discordance was noted between radiological cord compression and clinical cervical myelopathy, and they recommended against routine spinal imaging of NF1 patients. This recommendation correlates with the fact that in our series, only symptomatic patients required intervention. In addition, much like in our series, there was no spinal instrumentation following laminectomy for tumor resection. This finding differs from a recent publication from Taleb et al., describing 21 patients with NF1 and cervical spine manifestation. In their series, 85% of the patients eventually required spinal instrumentation and fusion8. Another interesting finding from our series is that none of the patients with KN of the lumber region required surgical intervention for tumors of this region. Recently, Dombi et al reported that MEK inhibitors show good response in inoperable plexiform neurofibromas (PN)10. This emerging medical treatment, and others, may in the future alter our strategy in multi-level diseases. Yet, as of now, its role in progressive spinal mechanical compression is unclear. A few sporadic case reports were published describing patients with KN and spinal cord compression11–13. In all these case reports, a positive outcome was noted following surgical decompression. Interestingly, in one report published by Montemurro et al. the patient suffered from 6
postoperative hydrocephalus one month following resection of cervical KN, necessitating permanent CSF diversion11. In our series, 3 out of 7 operated patients required shunting. Given the possibility of a link between KN, surgery and hydrocephalus, this generally unusual sequela should raise a high level of suspicion in any ICP-related complaints of these patients, despite the unknown mechanism. Our present study, emphasizes for the first time, that the surgical outcome of the KN patients was reasonably good despite their complicated disease and heavy tumor load. This is especially true when considering that the progressive neurological deficit associated with the culprit lesion was stabilized or improved in 71% of patients in our series. Intradural spinal cord tumors in NF1 rarely transform into MPNST. Hence the decision making of timing and the technical ramifications directed towards treating a slowly progressive pinching effect on the spinal cord. For example, unilateral neurofibromas will mostly displace the cord from side to side. When the pressure is bilateral, clinical deterioration may occur without any such displacement. Despite that, the long-term outcome of these patients, given their heavy
spinal tumor burden, and other comorbidities is usually not favorable. Both patients that did not improve following surgery had a significant neurological impairment preoperatively, and following surgery experienced further motor and sensory deterioration. For KN patients, limiting the surgery for the culprit lesion causing the compression may aid in maximizing outcome while minimizing operative risks. As such, focused aggressive, early intervention should be considered as it may optimize the surgical outcome, in this unique group of patients.
Conclusion NF1 patients with KN and associated neurological deficit may benefit from surgical decompression. As these patients usually harbor significant spinal disease, surgery should be limited to the culprit lesion, and kept as minimally invasive as possible. Timed decompression may lead to good surgical outcome. Hydrocephalus may be an unexpected post-surgical sequela in these patients.
7
References 1. Gutmann DH, Ferner RE, Listernick RH, Korf BR, Wolters PL, Johnson KJ. Neurofibromatosis Type 1. Vol 3.; 2017. doi:10.1038/nrdp.2017.4 2. Shofty B, Constantini S, Ben-Shachar S. Advances in Molecular Diagnosis of Neurofibromatosis Type 1. Seminars in Pediatric Neurology. 2015;22(4). doi:10.1016/j.spen.2015.10.007 3. Murovic JA, Kim DH, Kline DG. Neurofibromatosis-associated nerve sheath tumors. Neurosurg Focus. 2006;20(6). 4. Restrepo CS, Riascos RF, Hatta AA, Rojas R. Neurofibromatosis type 1: spinal manifestations of a systemic disease. J Comput Assist Tomogr. 2005;29(4):532-539. 5. Thakkar SD, Feigen U, Mautner VF. Spinal tumours in neurofibromatosis type 1: an MRI study of frequency, multiplicity and variety. Neuroradiology. 1999;41(9):625-629. 6. Ruggieri M, Polizzi A, Spalice A, et al. The natural history of spinal neurofibromatosis: a critical review of clinical and genetic features. Clin Genet. 2015;87(5):401-410. doi:10.1111/cge.12498 7. Mauda-Havakuk M, Shofty B, Ben-Shachar S, Ben-Sira L, Constantini S, Bokstein F. Spinal and paraspinal plexiform neurofibromas in patients with neurofibromatosis type 1: A novel scoring system for radiological-clinical correlation. American Journal of Neuroradiology. 2017;38(10). doi:10.3174/ajnr.A5338 8. Taleb FS, Guha A, Arnold PM, Fehlings MG, Massicotte EM. Surgical management of cervical spine manifestations of neurofibromatosis Type 1: long-term clinical and radiological follow-up in 22 cases. J Neurosurg Spine. 2011;14(3):356-366. doi:10.3171/2010.9.SPINE09242 9. Leonard JR, Ferner RE, Thomas N, Gutmann DH. Cervical cord compression from plexiform neurofibromas in neurofibromatosis 1. J Neurol Neurosurg Psychiatry. 2007;78(12):1404-1406. doi:10.1136/jnnp.2007.121509 10. Dombi E, Baldwin A, Marcus LJ, et al. Activity of Selumetinib in Neurofibromatosis Type 1Related Plexiform Neurofibromas. N Engl J Med. 2016;375(26):2550-2560. doi:10.1056/NEJMoa1605943 11. Montemurro N, Cocciaro A, Meola A, Lutzemberger L, Vannozzi R. Hydrocephalus following bilateral dumbbell-shaped c2 spinal neurofibromas resection and postoperative cervical pseudomeningocele in a patient with neurofibromatosis type 1: a case report. Evid Based Spine Care J. 2014;5(2):136-138. doi:10.1055/s-0034-1387805 12. Bigder M, Szelemej P, Berrington N. Extensive Mirror-Image Neurofibromas of Entire Spine Resulting in Spastic Tetraplegia. World Neurosurg. 2017;108:985.e7-985.e8. doi:10.1016/j.wneu.2017.08.139 13. Pandey S, Singh K, Sharma V, Khan MT, Ghosh A, Santhosh D. Bilateral Mirror Image Cervical Neurofibroma in an Adult with Neurofibromatosis Type 1. The Malaysian journal of medical sciences: MJMS. 2017;24(1):117.
8
Figure 1. Diagnostic MRI done at presentation, as well as intraoperative image. (A) Sagittal T2 demonstrating an intraduralextramedullary laterally located mass lesion (blue arrow) at the level of C6 vertebral body. (B) Axial T2 MRI at the level of C6 demonstrates bilateral, kissing neurofibromas (blue arrows) with significant cord compression. (C) Intraoperative image of the left tumor resection (white arrow) following resection of the right tumor (black arrow). Figure 2. (A) Axial T2 MR images at a 3-year interval. Bilateral, small tumors were noted at C1 level on axial T2 (blue arrows) at the time of the first surgery. (B) Follow up imaging 3 years later demonstrate significant progression of the C1 lesions, now kissing (blue arrows).
9
B
A
C c6
A
B
c6
Abbreviation list Kissing neurofibromas – KN Neurofibromatosis type 1 – NF1 Institutional review board – IRB
Disclosure statement All authors have nothing to disclose, and have no conflict of interest in regard to this work Ben Shofty (on behalf of all authors)
Ben Shofty: conceptualization, writing Michal Mauda-Havakuk: investigation, writing-review and editing, analysis Liat Ben-Sira: writing-review and editing, analysis, Felix Bokstein writing-review and editing, analysis Zvi Lidar: writing-review and editing, analysis, supervision Khalil Salame: writing-review and editing, analysis Akiva Korn: writing-review and editing, analysis Shlomi Constantini: conceptualization , writing, supervision
S1878-8750(19)32964-X
DOI:
https://doi.org/10.1016/j.wneu.2019.11.124
Reference:
WNEU 13793
To appear in:
World Neurosurgery
Received Date: 6 August 2019 Revised Date:
20 November 2019
Accepted Date: 21 November 2019
Please cite this article as: Shofty B, Mauda-Havakuk M, Ben-Sira L, Bokstein F, Lidar Z, Salame K, Korn A, Constantini S, Surgical management of “kissing” spinal plexiform neurofibromas in NF1 patients, World Neurosurgery (2019), doi: https://doi.org/10.1016/j.wneu.2019.11.124. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Surgical management of “kissing” spinal plexiform neurofibromas in NF1 patients Ben Shoftya,b, Michal Mauda-Havakukb, Liat Ben-Sirab, Felix Boksteinb, Zvi Lidara, Khalil Salamea, Akiva Korna, and Shlomi Constantinib,c 1. Department of Neurosurgery 2. The Gilbert Israeli Neurofibromatosis Center 3. Department of Pediatric Neurosurgery All at the Tel-Aviv Medical Center, and Tel-Aviv University, Tel-Aviv, Israel Correspondence to: Prof. Shlomi Constantini, MD, MSc, Department of Pediatric Neurosurgery, Tel-Aviv Medical Center, 6 Weizmann St. Tel-Aviv, Israel. [email protected] Highest Academic Degree: Ben Shofty MD, PhD, Michal Mauda-Havakuk MD, PhD, Liat Ben-Sira MD, Felix Bokstein MD, Zvi Lidar MD, Khalil Salame MD, Akiva Korn MSc, and Shlomi Constantini MD, MSc Keywords: Neurofibromatosis, intradural, neurofibroma, NF1, kissing Short title: “kissing” spinal plexiform neurofibromas in NF1
Abstract Introduction “Kissing” neurofibromas (KN) are a unique group of spinal tumors found in neurofibromatosis type 1 (NF1) patients. These are bilateral neurofibromas that approximate each other at the same level, with significant impingement compression of the cord or thecal sac. The best management options and surgical strategies for NF1 patients with KN have not been standardized.
Methods We conducted a retrospective study evaluating adult NF1 patients with KN. All patients are followed routinely at the Gilbert Israeli NF Center. Patients files were reviewed for natural history, imaging features, surgical technique, and surgical outcome.
Results Twelve patients with at least 1 pair of KN were identified (6 females). Median age at spinal presentation was 24 (range 17-48). KN were located at the cervical (n=8) and lumbar (n=8) region, with no thoracic involvement. Seven of the 12 patients were operated; all underwent surgery due to cervical compression with progressive myelopathy. Four patients remained asymptomatic during the follow-up period. Three patients underwent multiple operations. Operative outcome was favorable in 71% of patients, with marked overall motor improvement or stabilization of neurological deterioration. Two patients that entered surgery with a low functional reserve deteriorated following surgery.
Conclusions In our series, KN caused progressive cord compression in 7 of the 8 patients with cervical tumors. No intervention was needed for lumbar tumors. Cervical tumors should be followed closely, with a low-threshold for intervention. NF1 patients harboring KN should be followed both clinically and radiologically for life.
1
Background Neurofibromatosis type 1 (NF1) is a common neurocutaneous syndrome that affects ~1:2000 individuals globally1,2. This syndrome causes tumors along the nervous system, including peripheralnerve-sheath tumors, neurofibromas, that often cause pain and neurological deficits, and may undergo malignant transformation3. Spinal involvement varies widely. Included in the spinal phenotype are structural abnormalities such as vertebral anomalies, dural ectasia, and various spinal tumors4. Interestingly, intra-dural tumors are relatively rare in NF1, known to occur in only ~2-6% of patients, as most neoplastic involvements are para-spinal5, extending from the neuroforamina laterally. Anatomically, most spinal tumors in NF1 are foraminal (~57% of patients with spinal involvement), followed by intradural-extramedullary (33%). Intramedullary tumors occur rarely and are usually of glial origin. In our center, out of ~300 adult NF1 patients, only 34 required focused imaging for work-up of their spinal disease. NF1 is a disease with unpredictable progression patterns, and highly variable clinical course and tumor load. Two major patterns of spinal involvement have been defined in NF1: “spinal NF1” in which the disease is limited to the spine, with little, if any, systemic or cutaneous manifestations6, and NF1 in which there is a significant involvement of the spine, and the rest of the body is also affected7. Therefore, treatment should be individualized, requiring an experienced multidisciplinary team and a dedicated disease-focused center. Surgery is reserved for progressive tumors or those causing neurological deficit. Often, in patients with significant spinal tumor burden, it is difficult to locate the culprit tumor, and choosing the best surgical strategy may be challenging. Evidence in the literature regarding surgical management of spinal NF1 is scarce, mainly focusing on deformity correction 7,8. Currently, surgical treatment decisions for these complex patients are typically based on personal experience and multi-disciplinary team discussions. Multiple, bilateral neurofibromas in NF1 patients pose a special challenge for the treating neurosurgeon. In this paper, we focus on a unique group of spinal tumors we termed “kissing” neurofibromas – bilateral neurofibromas that cause significant pinching compression on the thecal sac and approximate each other. These tumors may often be responsible for clinical deterioration and necessitate surgical intervention. The best management options and surgical techniques for NF1 patients with kissing neurofibromas (KN) are still unknown, as the natural history and the surgical outcome of patients harboring these tumors has not been analyzed.
2
The aim of this report is to describe this unique group of patients, and our personal clinical and surgical experience.
Methods All NF1 patients in this study were followed and treated at the Gilbert Israeli International NF Center (GIINFC). We conducted a retrospective study that was approved by the institutional review board (IRB) of Tel-Aviv Sourasky Medical Center. Imaging of 34 patients with NF1 and spinal involvement were reviewed and analyzed for the presence of KN. KN were defined as bilateral foraminal-extramedullary tumors approximating each other at the same level, with significant compression of the thecal sac. Included in this study were all patients with KN and sufficient followup and imaging data. Clinical, radiological, and surgical data was extracted from the patients’ electronic medical records. Neurological course was determined according to the neurological examination as documented during subsequent clinic follow-up. Information about structural changes and tumor progression were based on sequential MR imaging and the report of the neuroradiologist at the NF1 center. All imaging of the included patients was reviewed by a radiology resident (Dr. Mauda-Havakuk), under the supervision of a senior neuro-radiologist (Dr. Ben-Sira). Clinical outcome such as neurological and functional status were based on serial reports of the treating neurologist.
Results Following retrospective review of adult NF1 patients with spinal disease (n=34), twelve patients were found to have at least one pair of KN and were included in this study. Six were females. Median age at diagnosis with KN was 24 years (range 17-41 years, mean 28±11 years). Five patients had positive familial history of NF1. Out of 8 patients who were symptomatic at presentation, 6 presented with motor symptoms that correlated with their spinal disease, and 2 patients presented with pain that led to diagnosis. Four patients were asymptomatic and were either diagnosed via routine screening or identified incidentally. Mean follow-up time was 48 months (±32 months, range 14-104 months). Anatomically, all patients in this series suffered from extensive spinal involvement, with a minimum of 3 spinal levels involved, and a maximum of 20. Of the 12 patients, 10 had cervical tumors, 4 of whom also had lumbar KN. Two patients had lumbar tumors only. Nine patients had anteriorly-kissing NF, and 3 had posterior tumors.
3
During the follow-up period, 7 patients experienced neurological deterioration associated with their KN, and were operated. All pathological reports were consistent with a WHO I neurofibroma. The other 5 patients were clinically and radiologically stable. Four of the neurologically stable patients were asymptomatic at presentation, 1 presented with tumor-associated pain. All of the 7 deteriorating patients underwent surgical resection of their culprit lesions. All of the surgeries were for progressive, compressing cervical tumors. No intervention was needed for lumbar KN. Six patients underwent more than one operation (a total of 16 surgeries in 6 patients), for progression of tumor at a different spinal level. Major surgical complications occurred in 1 patient, however, this patient entered the surgery with significant cord compression and minimal motor reserve and required prolonged postoperative mechanical ventilation. Two other patients suffered from wound related issues that resolved with conservative management. Postoperative hydrocephalus necessitating shunt placement occurred in 3/7 operated patients, without any evidence for postoperative CNS infection, bleed or other obvious mechanism. During the post-operative follow-up period, of the operated patients, 2 experienced worsening of their neurological condition, 3 patients stabilized, and two patients improved. Ultimately, in our series, the natural history of this group is unfavorable, as 7/12 patients experienced progressive neurological deterioration due to cord compression prior to neurosurgical intervention. This led to severe motor disability and even death from related complications in one patient. Of note, none of the patients in this series suffered from clinically significant postoperative deformity that required intervention.
Example Case A 36-year-old female, diagnosed with familial NF1 since birth. Known history of multiple cutaneous and sub-cutaneous neurofibromas, some of which underwent resections, with a growing overall disease burden, and a significant spinal disease. Presented with 4 months of cervical and thoracic mechanical back pain, with right C5-7 polyradiculopathy, and signs and symptoms of cervical myelopathy. Follow-up cervical MRIs demonstrated progressive bilateral “kissing” neurofibromas (KN) at C6-7 with significant cord compression (Figure 1). She then underwent C6-7 laminectomy and resection of the intradural component emerging from the sensory rootlet of C7 (see surgical technique in the following section). Postoperative course was uneventful and complete recovery was documented. Follow-up imaging demonstrated complete decompression of the cord at C6-7 with no residual 4
tumor (Figure 2 A+B), and small neurofibromas at C1-2 with no cord compression (Figure 2C, blue arrows). Three years later, the patient presented with recurrence of the cervical myelopathic symptoms, with progressive cord compression at the level of C1-2 due to bilateral KN. Rapid radiological and clinical progression over 4 months was noted. The patient underwent a second surgery with C1-2 laminectomy and intradural resection of the KN. Postoperative recovery was slow, with spastic weakness of the right hand, recovering fully following rehabilitation.
Surgical technique All patients were operated under general anesthesia in prone position, under full intraoperative electrophysiological monitoring. Following localization of the correct level, skin incision, fascial and muscle dissection, and laminectomy or laminotomy using the bone scalpel was done. Ultrasound is routinely used to identify those tumors that are not obvious on the surface prior to dural opening. The rest of the operation was done under the operative microscope. When an enlarged extradural component exists, it is usually debulked first from the two sides, followed by dural opening with identification of the dural orifice and careful microscopic dissection bilaterally, away from the cord. Spinal neurofibromas commonly demonstrate a dumbbell-shaped anatomy. This dumbbell appearance if often seen with both the extra- and intra-canalicular component, as well as the intraand extra-dural component. For those tumors kissing dorsally, the origin sensory rootlet is usually cut, and the tumor removed either with microdissection or with the ultrasonic aspirator. For those tumors kissing anteriorly, direct stimulation of the involved rootlet is performed, providing additional information used when deciding how radical a removal can be completed. Note that an anterior KN, in many cases, may create a continuity of tumor anteriorly to the cord, and therefore pose a significant technical challenge. In addition, radical removal of anteriorly-originating tumors may cause postoperative motor deficit. Thus, the surgeon should consider a more conservative intra-capsular removal, leaving the radical surgical approaches for dorsally-located tumors.
5
Discussion Progressive KN causing spinal cord compression in NF1 patients is an under-reported pathology that may have significant impact on the neurological status and quality of life of the patient. Here, we presented the first case-series focused on the surgical management of these unique tumors. Our study of 12 patients with KN raises several key points in the follow-up, surgical management, and postoperative care of these patients. In our center, more than 1400 NF1 patients are followed and treated by a dedicated, neurosurgicallyoriented multidisciplinary team. As the majority of the followed patients are pediatric, only 34/300 adult patients underwent clinically-indicated spinal imaging. Of this group, only 12 patients were diagnosed with at least one pair of KN. The relative rarity of this finding, occurring in 4% of adult NF1 patients, may explain the lack of publications dealing with this pathology. In our cohort, all of the patients were young adults at presentation, and progression was noted in 7/12 patients throughout the follow-up period of 4 years. It is important to notice that none of the incidentally diagnosed patients required any intervention or treatment during the follow-up period. In their report from 2007, Leonard et al. described cervical cord compression in 13 NF1 patients9. In their study, a discordance was noted between radiological cord compression and clinical cervical myelopathy, and they recommended against routine spinal imaging of NF1 patients. This recommendation correlates with the fact that in our series, only symptomatic patients required intervention. In addition, much like in our series, there was no spinal instrumentation following laminectomy for tumor resection. This finding differs from a recent publication from Taleb et al., describing 21 patients with NF1 and cervical spine manifestation. In their series, 85% of the patients eventually required spinal instrumentation and fusion8. Another interesting finding from our series is that none of the patients with KN of the lumber region required surgical intervention for tumors of this region. Recently, Dombi et al reported that MEK inhibitors show good response in inoperable plexiform neurofibromas (PN)10. This emerging medical treatment, and others, may in the future alter our strategy in multi-level diseases. Yet, as of now, its role in progressive spinal mechanical compression is unclear. A few sporadic case reports were published describing patients with KN and spinal cord compression11–13. In all these case reports, a positive outcome was noted following surgical decompression. Interestingly, in one report published by Montemurro et al. the patient suffered from 6
postoperative hydrocephalus one month following resection of cervical KN, necessitating permanent CSF diversion11. In our series, 3 out of 7 operated patients required shunting. Given the possibility of a link between KN, surgery and hydrocephalus, this generally unusual sequela should raise a high level of suspicion in any ICP-related complaints of these patients, despite the unknown mechanism. Our present study, emphasizes for the first time, that the surgical outcome of the KN patients was reasonably good despite their complicated disease and heavy tumor load. This is especially true when considering that the progressive neurological deficit associated with the culprit lesion was stabilized or improved in 71% of patients in our series. Intradural spinal cord tumors in NF1 rarely transform into MPNST. Hence the decision making of timing and the technical ramifications directed towards treating a slowly progressive pinching effect on the spinal cord. For example, unilateral neurofibromas will mostly displace the cord from side to side. When the pressure is bilateral, clinical deterioration may occur without any such displacement. Despite that, the long-term outcome of these patients, given their heavy
spinal tumor burden, and other comorbidities is usually not favorable. Both patients that did not improve following surgery had a significant neurological impairment preoperatively, and following surgery experienced further motor and sensory deterioration. For KN patients, limiting the surgery for the culprit lesion causing the compression may aid in maximizing outcome while minimizing operative risks. As such, focused aggressive, early intervention should be considered as it may optimize the surgical outcome, in this unique group of patients.
Conclusion NF1 patients with KN and associated neurological deficit may benefit from surgical decompression. As these patients usually harbor significant spinal disease, surgery should be limited to the culprit lesion, and kept as minimally invasive as possible. Timed decompression may lead to good surgical outcome. Hydrocephalus may be an unexpected post-surgical sequela in these patients.
7
References 1. Gutmann DH, Ferner RE, Listernick RH, Korf BR, Wolters PL, Johnson KJ. Neurofibromatosis Type 1. Vol 3.; 2017. doi:10.1038/nrdp.2017.4 2. Shofty B, Constantini S, Ben-Shachar S. Advances in Molecular Diagnosis of Neurofibromatosis Type 1. Seminars in Pediatric Neurology. 2015;22(4). doi:10.1016/j.spen.2015.10.007 3. Murovic JA, Kim DH, Kline DG. Neurofibromatosis-associated nerve sheath tumors. Neurosurg Focus. 2006;20(6). 4. Restrepo CS, Riascos RF, Hatta AA, Rojas R. Neurofibromatosis type 1: spinal manifestations of a systemic disease. J Comput Assist Tomogr. 2005;29(4):532-539. 5. Thakkar SD, Feigen U, Mautner VF. Spinal tumours in neurofibromatosis type 1: an MRI study of frequency, multiplicity and variety. Neuroradiology. 1999;41(9):625-629. 6. Ruggieri M, Polizzi A, Spalice A, et al. The natural history of spinal neurofibromatosis: a critical review of clinical and genetic features. Clin Genet. 2015;87(5):401-410. doi:10.1111/cge.12498 7. Mauda-Havakuk M, Shofty B, Ben-Shachar S, Ben-Sira L, Constantini S, Bokstein F. Spinal and paraspinal plexiform neurofibromas in patients with neurofibromatosis type 1: A novel scoring system for radiological-clinical correlation. American Journal of Neuroradiology. 2017;38(10). doi:10.3174/ajnr.A5338 8. Taleb FS, Guha A, Arnold PM, Fehlings MG, Massicotte EM. Surgical management of cervical spine manifestations of neurofibromatosis Type 1: long-term clinical and radiological follow-up in 22 cases. J Neurosurg Spine. 2011;14(3):356-366. doi:10.3171/2010.9.SPINE09242 9. Leonard JR, Ferner RE, Thomas N, Gutmann DH. Cervical cord compression from plexiform neurofibromas in neurofibromatosis 1. J Neurol Neurosurg Psychiatry. 2007;78(12):1404-1406. doi:10.1136/jnnp.2007.121509 10. Dombi E, Baldwin A, Marcus LJ, et al. Activity of Selumetinib in Neurofibromatosis Type 1Related Plexiform Neurofibromas. N Engl J Med. 2016;375(26):2550-2560. doi:10.1056/NEJMoa1605943 11. Montemurro N, Cocciaro A, Meola A, Lutzemberger L, Vannozzi R. Hydrocephalus following bilateral dumbbell-shaped c2 spinal neurofibromas resection and postoperative cervical pseudomeningocele in a patient with neurofibromatosis type 1: a case report. Evid Based Spine Care J. 2014;5(2):136-138. doi:10.1055/s-0034-1387805 12. Bigder M, Szelemej P, Berrington N. Extensive Mirror-Image Neurofibromas of Entire Spine Resulting in Spastic Tetraplegia. World Neurosurg. 2017;108:985.e7-985.e8. doi:10.1016/j.wneu.2017.08.139 13. Pandey S, Singh K, Sharma V, Khan MT, Ghosh A, Santhosh D. Bilateral Mirror Image Cervical Neurofibroma in an Adult with Neurofibromatosis Type 1. The Malaysian journal of medical sciences: MJMS. 2017;24(1):117.
8
Figure 1. Diagnostic MRI done at presentation, as well as intraoperative image. (A) Sagittal T2 demonstrating an intraduralextramedullary laterally located mass lesion (blue arrow) at the level of C6 vertebral body. (B) Axial T2 MRI at the level of C6 demonstrates bilateral, kissing neurofibromas (blue arrows) with significant cord compression. (C) Intraoperative image of the left tumor resection (white arrow) following resection of the right tumor (black arrow). Figure 2. (A) Axial T2 MR images at a 3-year interval. Bilateral, small tumors were noted at C1 level on axial T2 (blue arrows) at the time of the first surgery. (B) Follow up imaging 3 years later demonstrate significant progression of the C1 lesions, now kissing (blue arrows).
9
B
A
C c6
A
B
c6
Abbreviation list Kissing neurofibromas – KN Neurofibromatosis type 1 – NF1 Institutional review board – IRB
Disclosure statement All authors have nothing to disclose, and have no conflict of interest in regard to this work Ben Shofty (on behalf of all authors)
Ben Shofty: conceptualization, writing Michal Mauda-Havakuk: investigation, writing-review and editing, analysis Liat Ben-Sira: writing-review and editing, analysis, Felix Bokstein writing-review and editing, analysis Zvi Lidar: writing-review and editing, analysis, supervision Khalil Salame: writing-review and editing, analysis Akiva Korn: writing-review and editing, analysis Shlomi Constantini: conceptualization , writing, supervision