gms | German Medical Science

Objective: Malignant intramedullary tumors usually require radiation therapy which might lead to cord edema with severe neurological decline. Spinal expansile duraplasty enlarges the intradural space and might thus inhibit radiation-induced neurological deficits. The aim of this study is to evaluate the indications and the clinical course of patients undergoing expansile duraplasty.

Methods: In this retrospective observational analysis (2007 – 2016), we included all patients with primary and secondary intramedullary tumors who underwent spinal expansile duraplasty with either autologous or allogenic material. To analyze the degree of preoperative edema and the relative space occupying effect we formed the ratio of the diameter of the spinal cord adjacent to the tumor (median of a.p. diameter one level apical and a.p. diameter one level caudal to tumor) divided by the diameter measured at the level of the tumor (ratio > 0.8 was regarded as a significant swelling). The presence of CSF fringe around the medulla was analyzed. Postoperative course was analyzed with special emphasis on neurological deficits, wound breakdown and CSF fistula.

Results: We screened records from 985 patients, identified by the surgical procedural codes of 5-036.8 and 5-021.3. Thirty-one patients suffered from primary or secondary intramedullary tumors. Fifteen patients were excluded since a tumor debulking was carried out simultaneously, thus, 16 patients met the inclusion criteria. Twelve patients suffered from an intramedullary metastasis, four from an intramedullary glioma. Main localization was thoracic spine (6 patients; 5 cervical, 4 lumbar, 1 craniocervical junction). Mean craniocaudal tumor extension measured on the saggital T1 contrast enhanced image was 24.6 mm (SD 8.2 mm), the mean extension of intramedullary edema measured 116.2 mm (SD 48.7 mm). A diameter ratio > 0.8 representing a significant preoperative intramedullary swelling was seen in 87.5%. Allogenic material for expansile duraplasty (Goretex^®) was used in 81.3%. A CSF fringe surrounding the medulla at the level of the tumor was detected on the postoperative images of all patients, demonstrating successful decompression. Median length of follow-up was 3 months after the surgery. Postoperative course was uneventful in 87.5%, 1 patient developed wound infection, one patient died in the course because of ongoing primary bronchial carcinoma. All but one patient (93.8%) remained clinically stable during the course of radiotherapy.

Conclusion: Spinal expansile duraplasty for unresectable intramedullary tumor prior to radiation therapy is a previously undescribed method and a safe and feasible procedure. Wound breakdown or CSF fistula due to radiation associated wound healing disturbance was not observed. In selected cases the procedure seems to effectively prevent neurological decline due to radiation induce cord swelling.