INTRAMEDULLARY SPINAL-CORD TUMORS

Since Guidetti and Slooff's masterbooks (1964), numerous papers have b een devoted to intramedullary spinal cord tumors (IMT) and their treat ment, most of them were focused on ependymomas and astrocytomas. Infor med opinion was that these tumors were difficult to cure and that biop sy plus radiotherapy was the treatment of choice. Thanks to microsurge ry and bipolar coagulation surgeons grew bolder and more efficient as illustrated by the contributions from Hurth or Resche on hemangioblast omas, Fischer on ependymomas, Epstein on childhood astrocytomas and Gu idetti, Malis, Stein on IMT in general. Meanwhile, in the eighties, Ma gnetic Resonance Imaging (MRI) and Cavitron Ultrasonic Surgical Aspira tor (CUSA) drastically modified diagnostic and therapeutic strategies of IMTs, as may be judged by the significant increase in the number of publications on this topic in current medical literature. However, no updated work of synthesis is currently available. This prompted us to publish our common experience based on 171 patients and 200 surgical procedures. Thanks to the collaboration of many colleagues of the ''So ciete de Neurochirurgie de Langue Francaise'' (SNCLF) who completed a questionnaire sent to them, we are now in the position to give an epid emiological estimation based on 1117 cases. This work is divided in th ree main parts. The first part deals with considerations common to all IMTs, the second with particular aspects specific of each tumoral typ e, the third with the results and concluding recommendations. CHAPTER I. GENERAL STATEMENTS TO ALL IMTs. A thorough knowledge of spinal cord anatomy is crucial to the understanding of the basic surgical princip les involved, in particular the anatomy as seen through the eyes of th e Neurosurgeon with his special interest in the surgical approach to I MTs. Such knowledge may help us to understand why it is better to open the spinal cord on the midline rather than laterally, why special car e must be devoted to arterial blood supply, particularly at thoracic a nd thoraco-lumbar levels, and how IMTs can be removed without neurolog ical deficit. Electrophysiology helps the surgeon to carry out complet e removal of IMTs without sequellae and guides him in preventing neuro logical deficits during surgery. Somatosensory Evoked Potentials (SEP) are useful for the diagnosis of IMTs and for the post-operative follo w-up. SEP were monitored during surgery in 33 patients of this series in 70% of them they showed changes in dorsal column conduction. When S EPs changes were not reversible, a permanent sensitive deficit was cli nically observed after surgery. On the other hand, transient modificat ions or unchanged SEPs during surgery were predictive of a good clinic al outcome. However, SEP monitoring, while obviously useful, is diffic ult to implement under good conditions in the operating theater. Moreo ver, SEPs give no information on the function of the pyramidal tracts which can only be assessed by monitoring Motor Evoked Potentials, a te chnique which is currently still at the experimental stage. Neuroradio logy has undergone revolutionary changes since the introduction of MRI . The surgeon is now able to see inside the spinal cord to distinguish between the solid and the cystic parts of the tumor, to assess the ex tent of associated syrinx, to detect a small tumor responsible for hyd romyelia, to discover whether the lesion is lateral or medial, deep or superficial, exophytic, extra or intramedullary. All such knowledge i s essential for the surgeon; to obtain it, MRI should be performed usi ng sagittal, coronal and axial slices with T1, T2 weighted images, and also post-contrast T1-weighted images. Unfortunately, MRI does not pr ovide accurate histological diagnosis in all cases : hemangioblastomas , lipomas, dermoid and epidermoid cysts have a typical MRI feature but low and high-grade gliomas (ependymomas, astrocytomas, ganglioglomas) , lymphomas, metastases and granulomas have no specific images. Theref ore MRI fails to predict the possibility of a complete tumor removal a nd the prognosis. Surgery and histology examination are mandatory in a ll cases. The histological findings in our series were reviewed by thr ee pathologists using the new WHO classification. In several cases, th is histological diagnosis was uneasy and the use of markers only could assess the tumor type. Surgery included laminotomy in children, and l aminectomy in adults, limited to the solid part of the tumor + one lev el above and one level below. The poles of the lesion have to be reach ed by opening the dura. Surgical removal of the tumor is performed thr ough the midline in all cases except when the tumor is superficial and lateralized. The term ''myelotomy'' should.be discarded as no cutting is done into the parenchyma. In order to minimize the risk of postope rative sensory deficit, the posterior columns must be carefully separa ted through the posterior sulcus and kept open by gentle pial sutures without any tension. Tumor debulking with CUSA must be performed prior to the search for any cleavage plane. The best landmark being the dif ference in color between the tumor and the spinal cord, use of the las er which chars the tissue is not recommendable. For the same reason, w e do not use laser to open the posterior sulcus nor to vaporize the tu mor. Since the introduction of MRI, ultrasonography has lost much of i ts justification. The goal of surgery should be a complete tumor remov al whenever possible. When no clear interface is seen between the tumo r and the spinal cord, we recommend halting the surgical procedure. Cl osing the pia and the arachnoid avoids adhesion between an open spinal cord and the dura. CHAPTER II. THE DIFFERENT SPINAL CORD TUMORS OF OU R SERIES Ependymomas of the filum terminale were eliminated from our s tudy since they are not IMTs, even though they enter the conus like th e finger of a glove. They are mostly of myxopapillary type and should be removed in the same way as extra-medullary tumors, i.e. without ope ning the conus. Epidemiologically, the estimated incidence of true int ramedullary ependymomas was 35% of all IMTs reported in the SNCLF inve stigation. We did not observe any typical clinical presentation in our series of 64 cases. MRI may show peripheral signal loss areas capping both poles of the tumor. This sign was observed in ependymomas, but n ot in all cases. Associated cysts were frequently present. All but one of our low-grade ependymomas were completely removed. No ad