A meta-analysis of tumor control rates and treatment-related morbidity for patients with glomus jugulare tumors.

OBJECT Because of the rarity of glomus jugulare tumors, a variety of treatment paradigms are currently used. There is no consensus regarding the optimal management to control tumor burden while minimizing treatment-related morbidity. In this study, the authors assessed data collected from 869 patients with glomus jugulare tumors from the published literature to identify treatment variables that impacted clinical outcomes and tumor control rates. METHODS A comprehensive search of the English-language literature identified 109 studies that collectively described outcomes for patients with glomus jugulare tumors. Univariate comparisons of demographic information between treatment cohorts were performed to detect differences in the sex distribution, age, and Fisch class of tumors among various treatment modalities. Meta-analyses were performed on calculated rates of recurrence and cranial neuropathy after subtotal resection (STR), gross-total resection (GTR), STR with adjuvant postoperative radiosurgery (STR+SRS), and stereotactic radiosurgery alone (SRS). RESULTS The authors identified 869 patients who met their inclusion criteria. In these studies, the length of follow-up ranged from 6 to 256 months. Patients treated with STR were observed for 72 ± 7.9 months and had a tumor control rate of 69% (95% CI 57%-82%). Those who underwent GTR had a follow-up of 88 ± 5.0 months and a tumor control rate of 86% (95% CI 81%-91%). Those treated with STR+SRS were observed for 96 ± 4.4 months and had a tumor control rate of 71% (95% CI 53%-83%). Patients undergoing SRS alone had a follow-up of 71 ± 4.9 months and a tumor control rate of 95% (95% CI 92%-99%). The authors' analysis found that patients undergoing SRS had the lowest rates of recurrence of these 4 cohorts, and therefore, these patients experienced the most favorable rates of tumor control (p < 0.01). Patients who underwent GTR sustained worse rates of cranial nerve (CN) deficits with regard to CNs IX-XI than those who underwent SRS alone; however, the rates of CN XII deficits were comparable. CONCLUSIONS The authors' analysis is limited by the quality and accuracy of these studies and may reflect source study biases, as it is impossible to control for the quality of the data reported in the literature. Finally, due to the diverse range of data presentation, the authors found that they were limited in their ability to study and control for certain variables. Some of these limitations should be minimized with their use of meta-analysis methods, which statistically evaluate and adjust for between-study heterogeneity. These results provide the impetus to initiate a prospective study, appropriately controlling for variables that can confound the retrospective analyses that largely comprise the existing literature.