Assessment of Volumetric Growth Rates of Spinal Intradural Extramedullary Schwannoma

Overview

Spinal intradural schwannoma detected incidentally increased recently. Because there is little knowledge about natural history of spinal schwannoma, there is no consensus of treatment. Our hypothesis is as follows; 1. Some schwannomas keep growing, the others do not. 2. Foraminal schwannomas do not usually grow. The investigators analyzed natural history and characteristics of 56 spinal schwannomas observed initially with accurate and reliable methods. Here the investigators displayed growing pattern and differential point of spinal schwannoma.

Study design and participants The investigators included all patients who were diagnosed of spinal IDEM schwannoma by MR imaging and checked follow-up MR imaging not undergoing surgical resection (including radiosurgery). The patients usually checked follow-up MR imaging annually or biannually. The patients who underwent surgical resection during the surveillance interval were also enrolled. Exclusion criteria of this study were as follows: 1) surgical resection at the diagnosis; 2) von Recklinghausen' disease (neurofibromatosis); and 3) recurred or residual tumor. For patients lost follow-up, the investigators checked whether the patients underwent a surgery or not by telephone and obtained recent radiologic studies as possible. Medical record and radiologic studies of enrolled patients were evaluated for patient demographics, tumor level, main location (intracanal or foraminal), shape (globus, rugby ball, or dumbbell), date at diagnosis, follow-up date at check follow-up MR imaging, and length of follow-up. These serial studies were compared with tumor growth to evaluate for any correlation. Tumor volume measurement and growth The investigators measured tumor volume by volume quantification method using the Leksell Gamma Plan® system (version, 10.1.1. Elekta Instrument AB, Stockholm, Sweden) in Figure 1. It calculates tumor volume by summation of cylindrical masses from each image and has an advantage to measure with the best accuracy especially non-globus tumor.6 The process of measuring tumor volume was rigorously designed to minimize observer subjectivity. Stack images without gap were loaded to the program. The investigators draw the tumor outlining by manually in select planes of each orthogonal view. The program measures tumor volume using an algorithm based on scattered data interpolation with radial basis functions. All tumor volumes were measured independently by two observers. Interobserver reliabilities were assessed by the Brand-Altman plot. The plot was applied to look for trends in the data that produce similar results and to evaluate the agreement between two observers.7 Under the confidence of excellent interobserver agreement marks, the investigators finally estimated the tumor volume as the mean value of two measured data. Volumetric growth rate (VGR) of the tumor was calculated volume change to the baseline volume and presented in percent. Overall VGR (%) was calculated using the following formula: (Final volume-Baseline volume)x100/Baseline volume. Annual VGR (%) was calculated as follows; Annual VGR (%)=(Final vol.-Baseline vol.)/(Baseline vol.)\*FU duration(days)\*12/365\*100 FU and vol. mean follow-up and volume, respectively. There is no consensus about definition and amount of tumor growth. A prior study reported that the cut-off value of growth rate of vestibular schwannoma extending into the cerebello-pontine angle was 10.4%. The investigators set 10% increase of VGR as a cut-off value. The investigators made two standards, 10% increase of overall and annual VGR, and evaluate validity by the receiver operating characteristic (ROC) curve. Enrolled patients were divided into two groups by the standard. The first group, growing tumor, was defined as mean ≥10% increase of tumor volume. The other group, stable tumor, was defined as mean \<10% increase, stationary, or decrease of tumor volume. Statistical analysis The investigators used the Brand-Altman plot for interobserver reliability. The investigators used the ROC analysis to compare predictive validity of standards by overall VGR and annual VGR, and to find out their optimal cut-off values of annual VGR in growing schwannomas. Because the area under the curve is a measure of the diagnostic power of a test, larger one is chosen as a standard dividing growing versus stable tumor. ROC curves were plotted using measures of sensitivity and specificity based on various cut-off values. Cut-off value was selected as the values closer to the upper left corner. For comparative analyses between the growing and stable tumor group, independent t tests for continuous variables, Fisher's exact test, and chi-square test were used. The significance level was set at p \< 0.05. The investigators used MedCalc Statistical Software version 12.7.2 (MedCalc Software bvba, Ostend, Belgium) for the Bland-Altman plot and the SPSS statistical package (Version 20.0; SPSS Inc, Chicago, IL, USA) for the other data analysis.