Medical treatment of recurrent MG is challenging. The poor availability of active chemotherapeutic drugs represents a major limitation in the decision-making process guiding treatment choice after failure of first-line chemotherapy. In addition to this, rapid tumor progression and low KPS of the patient are other important factors hampering re-treatment with chemotherapy. Against this background, recommendations for medical treatment of recurrent pretreated MGs are based almost exclusively on uncontrolled phase II studies.
In the present analysis, we found that fotemustine monotherapy is able to provide a response of 20% for an overall DCR of 47.5%, which is remarkable since these findings were observed in a heavily pretreated population (all patients had received at least one line of chemotherapy and 25% of patients had received two prior lines). These activity results are in line with those reported in the literature with single-agent fotemustine at the conventional schedule of 100 mg/m2, where a response rate of 15.5%–26% and a DCR of 50%–76% have been reported [7–9]. However, it should be noted that a less extensively pretreated population was included in the oldest studies (0–26.5% of patients had received previous chemotherapy) and, more importantly, CT scan instead of MRI was uniformly used to assess tumor response [7, 8]. On the other hand, our PFS at 6 months of 27% compares favorably with a PFS-6 months of 15% observed in a very recent report by Trevisan et al. . To this regard, it is reasonable to argue whether or not this difference in PFS should be attributed to the lower dose of fotemustine adopted in the majority of our patients which might have resulted into better tolerability of treatment. In fact, the population of the two studies is similar in that both reports include patients with recurrent MGs pretreated with ≥ 1 line of chemotherapy; yet, a much higher incidence of severe myelosuppression was recorded by Trevisan et al. with the use of a conventional schedule of fotemustine at 100 mg/m2 .
Interestingly, in our study fotemustine monotherapy produced more enthusiastic results than those reported with "older" nitrosureas. In a retrospective study, Kappelle et al. found a response of only 3% with the classical triple combination PCV in recurrent GBM . More recently, Rosenthal et al. reported 4% of responses with the use of BCNU for recurrent MG patients pretreated with temozolomide . Better activity results were obtained in GBM patients with the combination of BCNU and irinotecan , although polychemotherapy for recurrent MGs is usually associated with higher toxicity and is bound to a strong bias of patients selection.
Notably, our analysis also showed the absence of cross-resistance between fotemustine and temozolomide, since all responses were observed in temozolomide-pretreated patients. This observation is worthy of being pointed out in view of the recent incorporation of temozolomide in the standard treatment of newly diagnosed glioblastoma multiforme . Similarly, other authors have reported an activity of 30% with the use of fotemustine in glioblastoma patients pretreated with temozolomide . On the other hand, the lack of activity observed for fotemustine in PCV-pretreated patients, suggests the presence of cross-resistance between fotemustine and other nitrosureas as hypothesized preclinically .
Furthermore, our results showed that both the activity of first-line therapy and treatment with fotemustine were positive prognostic factors for OS; these findings confirm that the use of chemotherapy in recurrent MGs has a positive impact on patients outcome .
The better activity recorded in groups A and B where fotemustine was given at doses ranging from 65 mg/m2 to 85 mg/m2 is supposedly to ascribe to the absence of severe myelotoxicity that has allowed the administration of a higher dose intensity. In fact, at these doses we found no case of severe thrombocytopenia and/or neutropenia, whereas the same adverse events were 20% and 15% respectively in group C, where fotemustine was given at doses ranging from 75–85 mg/m2 to 100 mg/m2. Importantly, our data appear to rule out the hypothesis that the better activity reported for low-dose fotemustine could be attributed to an imbalanced distribution toward group A and B of tumors with a worse prognosis such as GBM (table 3).
Interestingly, fotemustine-induced myelosuppression can also be lowered by delaying the intervals between each cycle, as shown recently in two studies exploring fotemustine in combination with either dacarbazine or procarbazine for the treatment of recurrent GBM [25, 26]. However, the modest response (3%–11%) observed in these studies showed also that the use of a fotemustine schedule not including the weekly induction phase might compromise the activity of fotemustine itself through a higher rate of early progressions, thus invalidating the benefits potentially obtainable by the addition of a second cytotoxic [25, 26]. To this regard, in our study no cases of early progression were recorded after the induction phase and all patients received at least six cycles of fotemustine.
In the 19 patients with tissue available for assessment of the MGMT promoter methylation status, we found a considerable rate of disease control in patients with methylated MGMT (8 out of 12, 66.5%) (table 6). More interestingly, all of the 7 unmethylated patients were found progressive to fotemustine monotherapy. Despite the low number of patients analyzed and the heterogeneity of patients histotypes, these data suggest that the presence of MGMT promoter methylation is a crucial prerequisite for response to fotemustine although it does not guarantee sensitivity to treatment. On the other hand, patients with unmethylated MGMT do not appear to benefit at all from fotemustine chemotherapy. However, our number of patients was too low to observe a significant difference in terms of efficacy according to MGMT promoter methylation status. For this reason, the role of MGMT should be further assessed prospectively in a larger cohort of patients undergoing fotemustine chemotherapy, possibly re-assessing MGMT status at the time of tumor recurrence. In fact, a recent study suggested that changes in the status of MGMT promoter methylation may occur after primary treatment for newly diagnosed GBM .