Impact of adjuvant chemotherapy for gliomatosis cerebri

The chemotreatment group received adjuvant TMZ or nitrosourea-based chemotherapy following radiotherapy (n = 19), and the control group underwent radiotherapy alone (n = 18). There were no statistically significant differences in baseline characteristics between the 2 groups (Table 1). For this cohort group, twenty-two patients were men (59.5%) and 15 patients were women (40.5%). The mean age at presentation was 41.2 years (range: 11-67 years). The median KPS at the time of initial diagnosis was 80 (range: 70-100). The MR images demonstrated a variety of findings including presence of: a dominant mass formation (18 cases) or a diffuse infiltration pattern (19 cases), initial dimly enhanced lesions or not (6 vs. 31 cases), brainstem involvement or not (11 vs. 26 cases), and tumor extent, i.e., involving two lobes or ≥ 3 lobes (13 vs. 24 cases). Based on the histopathological data, 23 patients were diagnosed as having a tumor with low-grade features (grade II) and 14 patients were diagnosed as having tumor with high-grade features (grade III).

19 of 37 patients (51.4%) received adjuvant chemotherapy after radiotherapy, while 18 other patients underwent local radiotherapy alone. Chemotherapeutic regimens included temozolomide (TMZ) (200 mg/m² × consecutive 5 days per month, 3-9 cycles) and nitrosourea-based chemotherapy such as BCNU (carmustine-[1,3-bis (2-chloroethyl)-1-nitrosourea]), PCV (procarbazine/lomustine/vincristine). The choice of chemotherapeutic regimens was dependent upon how much tumors had oligodendroglial components. At an earlier period, PCV chemotherapy was given for 5 patients with tumor which dominantly contained oligodendroglial components, while BCNU treatment (4 patients) was performed for tumors with dominant astrocytic components. However, since TMZ treatment had been widely introduced into our institution in 2002, 10 patients received TMZ treatment regardless of the cellular components.

All patients were assessed by serial MR images with contrast enhancement. The follow-up images were evaluated for progression versus no progression by a radiologist (Kim ST) blinded to the pathological and clinical findings. During the follow-up period, we sometimes observed new contrast-enhancing lesions, which had been absent in the previous studies. These contrast-enhancing lesions were patched or punctuated lesions[14, 15] and located in the mid-portion of widely infiltrated lesions on FLAIR images. They could be differentiated from radiation necrosis by several MR sequences (diffusion/perfusion weighted images, MR spectroscopy, or FDG-PET) [16‐20]. It was difficult to apply the tumor response criteria commonly used for high-grade gliomas because of the lack of discrete and measurable tumor margins. Tumor progression was assessed by semi-quantitative analysis of the tumor extent in T2-weighted or FLAIR-weighted MR images using commercially available image analysis software (SCION Corp.).

In cases of tumor recurrence or progression, we performed surgical resection of the enhancing lesions in 10 patients or gamma knife radiosurgery in 6 patients. All patients with tumor recurrence or progression received 2^nd-line chemotreatment.

Overall survival and progression-free survival were calculated using the Kaplan-Meier method and groups were compared with the log-rank test. For statistical analysis, age, gender, KPS, histological grade, degree of resection, and involvement of brainstem were compared using Fisher exact test or Mann-Whitney test between the chemotreatment group and the control group. The Cox proportional hazards model with a stepwise procedure was used in the multi-variate survival analysis to assess the prognostic factors for survival. P values of less than 0.05 were considered statistically significant for all tests, using the SPSS software (SPSS version 10.0, Chicago, IL).

Twenty-five patients were still alive at the time of this analysis and the median follow-up period was 12.6 months (range: 7.1-49.4 months). Median overall survival in the chemotreatment group was 24.2 months (95% CI: 23.4-24.9 months), compared with 13.1 months (95% CI: 10.4-15.9 months) in the control group. Kaplan-Meier analysis showed that median overall survival in the chemotreatment group was longer than that in the control group (p = 0.045 by log-rank test, Fig. 2). Age, gender, KPS, cellular components, degree of resection and brainstem involvement were not identified as independent prognostic variables. For Cox proportional hazard model, no chemotherapy (HR 6.385, 95% CI: 1.8-22.6, p = 0.004) was strongly associated with poor survival even after adjustment for age, sex, KPS, cellular components, and brainstem involvement. Higher histological grade (HR 4.434, 95% CI: 1.312-14.982, p = 0.017) was also associated with dismal prognosis (table 2). Between patients treated with TMZ and those with nitrosourea-based chemotreatment, there was no significant difference of survival (p > 0.05 by log-rank test).

Bild vergrößern

Median progression-free survival in the chemotreatment group was 16.0 months (95% CI: 11.7-20.2 months), compared with 6.0 months (95% CI: 4.9-7.1 months) in the control group. Adjuvant chemotherapy (OR 8.250, 95% CI: 1.8-38.7, p = 0.007 by log-rank test, Fig. 3) was a significant independent factor for progression after adjustment of age, sex, KPS, cellular components, and histological grade.

Bild vergrößern

GC is an infrequent and controversial diagnosis, because the definition and pathophysiology have not been accurately defined yet. Recent studies using antemortem diagnostic criteria by MRI have revealed that the prognosis for GC is generally poor and the overall survival relatively short [1, 2, 6, 8, 9, 21‐27]. In the respective of radiological view, 19 of 38 patients in this study could be classified as de novo (primary) GC, whereas the others resulted from the spreading of a focal glioma (secondary GC) [10]. Vates et al. suggested that GC begins as either a low- or high-grade neoplasm without a focal mass, which progresses to form a focal mass similar to the natural history of low grade glioma [11]. However, GC should not be diagnosed by histopathological finding, but by radiographic findings. Furthermore, in most cases, difficulties in interpreting histological specimen are attributable to scattering of tumor cells and small amount of tissue specimens obtained by stereotactic biopsy. Therefore, it was too difficult to clarify the specific histological types (astrocyte-dominant or oligodendroglial-dominant) based on the tissue specimen. In this study, an interval between early stage and tumor progression in GC was extremely variable in each other.

We speculated that this change of enhancement pattern might imply a malignant progression of GC. In the current study, we found that early appearance of new patched or punctuated enhancing lesions was closely correlated with the poor survival and tumor progression. As a result, these findings supported that malignant transformation of GC is similar to that of low-grade gliomas. Tumor progression of GC used to be more rapid than that of low-grade gliomas, which can be explained by the following hypotheses. GC may contain highly invasive stem-like tumor cells and therefore shows a tendency to involve more lobes and progress rapidly [28]. Otherwise, it can be interpreted as a transformation from angiogenesis-independent growth to an angiogenesis-dependent phenotype driven by stem-like cells [29]. In the future, a detailed review of the clinical course in GC will contribute to our understanding of malignant progression of tumors [30].

To date, little is known about the therapeutic effect of adjuvant chemotherapy for GC. However, some researchers demonstrated that upfront TMZ treatment was effective for patients with GC [10, 12, 13]. By using a retrospective case-cohort study, we investigated the efficacy of adjuvant chemotherapy such as TMZ or nitrosourea-based chemotherapy for GC. In this study, we demonstrated that adjuvant chemotherapy following radiotherapy was effective for the prolongation of survival and delay of tumor progression. This suggests that adjuvant chemotherapy as well as first-line chemotherapy also provide improvement of survival in patients with GC. Furthermore, multivariate analysis revealed that GC patients with higher histological grade had shorter survival and earlier tumor progression than those with lower histological grade. Therefore, during the follow-up period, higher histological grade at the initial presentation or earlier appearance of enhancement should warrant dismal prognosis.

However, it is not certain that adjuvant chemotherapy should be required following radiotherapy for GC, because this retrospective study has some drawbacks. Even though these two groups were relatively balanced with respect to known prognostic variables, they were not randomly-allocated groups, but were groups determined by physician's preference. We should recognize that these selection biases could account for its effect on survival. Moreover, it was another important limitation in this study that chemotreatment group was composed of heterogeneous populations, for example, those treated with TMZ or nitrosourea-based chemotherapy.