Impact of tumor size on hearing outcome and facial function with the middle fossa approach for acoustic neuroma: a meta-analytic study.

OBJECTIVE

The purpose of this study was to review the English language literature concerning the effect of tumor size on hearing outcome and facial function after the middle fossa approach for acoustic neuroma in a large patient population.

MATERIAL AND

METHODS

The literature search identified a total of 11 studies reporting hearing outcome and facial function for a given tumor size. There were 1073 and 797 cases available for the analysis of hearing outcome and facial function, respectively. These cases were subdivided based on the way in which tumor size was measured: category 1 considered only the extracanalicular portion of the tumor; and category 2 considered the largest diameter of the tumor. In category 1, hearing and facial results were regrouped based on tumor size as follows: intracanalicular (IC) tumors; 1-9 mm tumors; 10-20 mm tumors; and a combined group of < 0.5 mm tumors, including IC tumors. In category 2, tumors were subdivided into 2 groups: those < 10 mm in diameter; and those 10-20 mm in diameter. In each category, tumor size groups were compared using the chi2 test in terms of the rate of functional hearing preservation and good facial function.

RESULTS

In category 1, analysis of the rate of functional hearing preservation showed that IC tumors compared favorably with the 1-9 mm and 10-19 mm tumors (56.9% vs 45.6%, p = 0.016; and 56.9% vs 32.3%, p < 0.001, respectively). The IC tumor group had the best rate of good facial function, followed by the 1-9 mm and 10-19 mm tumors (98.9% vs 93.9%, p = 0.007: and 98.9% vs 85.6%, p < 0.001, respectively). In category 2, rates of functional hearing preservation and good facial function were almost the same for tumors < 10 mm in diameter and those 10-20 mm in diameter (p > 0.05).

CONCLUSIONS

The meta-analysis revealed that tumor size is an important variable determining hearing outcome and facial function. Inclusion of the IC portion of a tumor in the tumor size measurement apparently hampered the statistical power of the study, leading to an overestimation of the size of IC tumors.