Journal of Neurosurgery

Although secondary gliosarcoma after treatment of primary glioblastoma multiforme has been described, little is known of these rare tumors. In this article the authors review the literature on secondary gliosarcoma, with attention to clinical course and pathological features.

A PubMed search of the key word intracranial “gliosarcoma” with and without “radiation” or “radiotherapy” in humans was performed. The 204 citations yielded were screened for relevancy to gliosarcomas that occur after treatment of previous intracranial neoplasms.

A search of the literature yielded 24 relevant articles, combined for a total of only 12 cases of secondary gliosarcoma and 12 cases of radiation-induced gliosarcoma. Of the 12 cases of secondary gliosarcoma, all were previously treated with surgery and radiotherapy (mean dose 50.7 Gy), with a mean survival of 13 months since time of gliosarcoma diagnosis (range 6.9–19.4 months). In the cases of radiation-induced gliosarcoma, the mean dose of previous radiotherapy was 51.3 Gy (median 54 Gy, range 24–60 Gy), and the mean survival since gliosarcoma diagnosis was 6.7 months (median 6 months, range 2–10 months).

Secondary gliosarcoma and radiation-induced gliosarcoma are exceedingly rare. The literature on secondary gliosarcoma illustrates a more favorable survival than for primary gliosarcoma but remains limited regarding clinical and radiographic presentation, response to treatment, and pathogenesis. The results of the present review also support the notion that secondary gliosarcomas and radiation-induced gliosarcomas are distinct entities, with longer survival and shorter latency of gliosarcoma induction seen in the former. Efforts to elucidate the role of radiotherapy in the induction of gliosarcomas may yield new insights into therapeutic risks of cranial radiation and CNS tumor pathogenesis.

Gliosarcoma can arise secondarily, after conventional adjuvant treatment of high-grade glioma. The current literature on the occurrence of secondary gliosarcoma (SGS) after glioblastoma multiforme (GBM) is limited, with only 12 reported cases. The authors present a large series of histologically confirmed SGSs, with follow-up to describe the clinical and radiological presentation, pathological diagnosis, and treatment outcomes.

Gliosarcoma cases were identified using the University of California, San Francisco's Departments of Neurological Surgery and Neuropathology databases. Through a retrospective chart review, cases of gliosarcoma were considered SGS if the following inclusion criteria were met: 1) the patient had a previously diagnosed intracranial malignant glioma that did not have gliosarcoma components; and 2) the histopathological tissue diagnosis of the recurrence confirmed gliosarcoma according to the most current WHO criteria. Extensive review of clinical, surgical, and pathology notes was performed to gather clinical and pathological data on these cases.

Thirty consecutive patients in whom SGS had been diagnosed between 1996 and 2008 were included in the analysis. All patients had previously received a diagnosis of malignant glioma. For the initial malignant glioma, all patients underwent resection, and 25 patients received both external-beam radiation and chemotherapy. Three patients received radiotherapy alone, 1 patient was treated with chemotherapy alone, and 1 patient's tumor rapidly recurred as gliosarcoma, requiring surgical intervention prior to initiation of adjuvant therapy. The median time from diagnosis of the initial tumor to diagnosis of gliosarcoma was 8.5 months (range 0.5–25 months). All but 1 patient (who only had a biopsy) underwent a second operation for gliosarcoma; 8 patients went on to receive radiotherapy (4 had brachytherapy, 3 had external-beam radiation, and 1 had Gamma Knife surgery); and 14 patients received additional chemotherapy. The median length of survival from the time of gliosarcoma diagnosis was 4.4 months (range 0.7–46 months). The median survival from the time of the original GBM diagnosis was 12.6 months (range 5.7–47.4 months). Patients who had received concurrent and adjuvant temozolomide for GBM had worse outcomes than those who had not (4.3 and 10.5 months, respectively; p = 0.045). There was no difference in time to diagnosis of gliosarcoma in these 2 groups (8 and 8.5 months; p = 0.387). Two patients who had not received radiation therapy for GBM had an anecdotally very prolonged survival (20.9 and 46.4 months).

The data underscore the difficulty associated with management of this disease. The strikingly poor survival of patients with SGS who had previously received combined radiation and temozolomide chemotherapy for GBM may reflect a unique molecular profile of GBM that eventually recurs as SGS. Further work will be required, controlling for multiple prognostic factors with larger numbers of patients.