Clinical Investigation
Multifocal Glioblastoma Multiforme: Prognostic Factors and Patterns of Progression

https://doi.org/10.1016/j.ijrobp.2007.03.045Get rights and content

Purpose

To assess the progression patterns in patients with multifocal glioblastoma multiforme who had undergone whole brain radiotherapy (WBRT), the historical standard, versus three-dimensional conformal radiotherapy, and to identify predictive treatment and pretreatment factors.

Methods and Materials

The records of 50 patients with multifocal glioblastoma multiforme treated with RT were reviewed. Univariate analyses were performed using survival methods and the Cox proportional hazards regression method. Multivariate analyses were performed using the Cox proportional hazards regression method.

Results

The mean age was 61 years, and 71% had a Karnofsky performance status (KPS) score of ≥70. Of the 50 patients, 32% underwent WBRT and 68%, three-dimensional conformal RT. Progression was local in all evaluable patients, as determined by imaging in 38 patients and early neurologic progression in 12. The median time to progression (TTP) was 3.1 months, and the median survival time (MST) was 8.1 months. The significant independent predictors of TTP on multivariate analysis were a KPS score <70 (p = 0.001), the extent of surgery (p = 0.040), a radiation dose <60 Gy (p = 0.027), and the lack of chemotherapy (p = 0.001). The significant independent predictors of a reduced MST were a KPS score <70 (p = 0.022) and the absence of salvage surgery (p = 0.011) and salvage chemotherapy (p = 0.003).

Conclusion

Local progression was observed in all patients. On multivariate analysis, no significant difference was found in the TTP or MST between three-dimensional conformal radiotherapy and WBRT. The KPS was a consistent independent predictor of both TTP and MST. On the basis of the progression pattern, we do not recommend WBRT as a mandatory component of the treatment of multifocal glioblastoma multiforme.

Introduction

Glioblastoma multiforme (GBM) is the most common primary brain tumor encountered in adults. Median survival with optimal treatment is <15 months 1, 2. Although solitary lesions are typical for GBM, multiple synchronous gliomatous foci are found at diagnosis with a reported incidence of 0.5–20% 3, 4, 5, 6, 7. Multiple synchronous gliomas are categorized either as multifocal if a microscopic connection is presumed to exist or as multicentric if a route of dissemination is inapparent 5, 8, 9. Given the diffusely infiltrative nature of GBM, the clinical utility of this classic subcategorization of multiple lesions may be scant (10). Patients with disseminated GBM have inferior survival and a shortened time to recurrence (11).

The standard treatment for GBM includes conformal radiotherapy (RT) encompassing the tumor volume plus margin and concurrent and adjuvant temozolomide 1, 2. Tumor progression or recurrence typically occurs locally within the irradiated volume, despite dose escalation with conformal RT techniques 12, 13. Although the entity of multiple synchronous gliomas has been recognized for decades, no clear consensus has been reached regarding the RT technique for those tumors. In the past, standard therapy for multifocal GBM consisted of whole brain RT (WBRT) with or without a smaller cone-down volume to boost the tumor region (14). The inclusion of whole brain fields in the RT plan for multifocal GBM is controversial because of the consistent pattern of failure within the original tumor volume, as well as the limited doses that can be delivered safely to whole brain fields 3, 5. At our institution, the standard approach changed from the inclusion of WBRT to the use of three-dimensional conformal RT (3D-CRT), with involved-field techniques, allowing for the study of both approaches in this report. This retrospective, single-institution analysis was performed to describe the influence of prognostic and treatment factors on local tumor control and patterns of recurrence.

Section snippets

Methods and Materials

A total of 497 patients who were diagnosed with GBM between 1995 and 2005 were identified from the Thomas Jefferson University tumor registry database. Of the 497 patients, 50 had had multifocal GBM and had undergone RT in the Jefferson Health System and were included as the study sample for this report. Multifocal disease was defined as multiple tumor sites with clear separation between foci; lesions with >2 cm of separation or in contralateral lobes were also labeled as multicentric. The

Results

A total of 50 patients with multifocal GBM were included in this analysis, representing 10% of all 497 patients with GBM treated at our institution during the study period. Patient age ranged from 19 to 86 years (mean, 61). A KPS score of <70 was seen in 29% of the subjects. Most patients (88%) had two lesions.

Subtotal resection was performed more commonly (66%) than were other operations. RT included WBRT in 32% of patients and was initiated urgently for 2 patients. The mean cumulative

Discussion

In our sample of 50 patients with multiple synchronous foci of disease at presentation, no recurrences were seen distant from the original foci in the absence of significant local progression. On multivariate analysis, no significant TTP or survival benefit was found in this sample for 3D-CRT compared with WBRT. This could have been related to small sample size or to the relatively more direct influence of performance status on the selection of the treatment approach and survival. In GBM,

Conclusion

In our series of 50 patients undergoing RT for multifocal GBM, no instance of intracranial dissemination in the absence of local progression was found. The radiation technique, 3D-CRT or WBRT, did not produce a statistically significant difference in the TTP or MST on multivariate analysis. KPS was the only predictor of both TTP and survival. Given the dominant pattern of local progression and the radiotherapeutic constraints imposed by the inclusion of WBRT, we recommend the use of standard

References (24)

  • H.B. Mishra et al.

    Multicentric gliomas: Two case reports and a review of the literature

    Br J Neurosurg

    (1990)
  • M.F. Pell et al.

    Multicentric malignant glioma

    Br J Neurosurg

    (1991)
  • Cited by (0)

    Conflict of interest: none.

    View full text