gms | German Medical Science

Objective: Standard of care in newly diagnosed glioblastoma (GBM) is maximal safe tumor resection followed by radio (RT)- and chemotherapy (CHT). At tumor relapse, however, the standard of care, in particular the value of re-resection, is still under debate. We therefore addressed the significance of recurrent GBM surgery with regards to functional outcome and survival times and special attention was paid to the extent of resection (EOR).

Method: By means of our tumor database, we identified 30 patients diagnosed with primary GBM between 2005 and 2010, who had been re-resected at least once for tumor relapse and were deceased at the time of analysis. Patients with biopsy at 1^st or 2^nd surgery were excluded. Patient demographics, functional status, adjuvant therapies and survival times were extracted from medical charts. EOR was determined on postoperative MRI scans where available. Survival after re-resection (SRR) was calculated by log-rank test.

Results: At initial diagnosis, our patient sample was treated by CR or partial resection (PR) followed by RT and/or temozolomide. Tumor relapse was mainly localized adjacent to the resection cavity (93%) and in 47% of patients in eloquent areas. 93% of patients received one, 7% two re-resections. Only 10% were treated with second-line RT or CHT prior to re-resection. At 1^st re-resection, CR was achieved in 50%, PR in 30%, EOR could not be determined in 20%. Postoperatively acquired neurologic deficits (NDs) were more frequent after re-resection than after 1^st surgery, but the rate of permanent deficits (10% vs. 13,3%) and KPS were comparable. After re-resection, patients received an average of 2 additional therapies with CHT most frequently applied. Analyzing only patients with CR (n=15) and PR (n=9) at 1^st re-resection, both groups were comparable with respect to demographic data, tumor eloquence, KPS, number of re-resections and adjuvant therapies for tumor relapse. It is important to note that SRR was significantly prolonged in completely resected patients (11 vs. 5 months; p=0,0497) while the rate of permanent NDs was comparable in both groups.

Conclusions: In this monocenter analysis, re-resection was the 1st treatment of choice in 90% of recurrent GBMs. Compared to PR, complete resection of recurrent GBM was associated with a significantly prolonged SRR without neurological impairment. The results of this study stress the significance of maximal tumor resection even at the time of relapse. Certainly, larger studies are warranted to confirm these findings.