Electrically evoked acoustic brainstem responses in vestibular schwannoma surgery with cochlear implant: inter and intra-rater agreement on waveform interpretation

Introduction There is still lack of consensus on whether the Auditory Nerve Test System (ANTS) via the Intra-cochlear Test Electrode (ITE) by MED-EL is the most appropriate one, given the recognized difficulty in correctly interpreting eABR traces, even by the most experienced raters. We aim to assess the intra-rater and inter-rater agreement score and by confirming the presence or absence of waves. This will also allow for a possible recalculation of the ANTS accuracy values. Materials and methods We included in our cohort the same patients from the previous study from Zanoletti-Concheri et al. plus one, for a total of 13 patients. All the patients had a small VS, treated with TLAB approach and simultaneous CI. All the patients were implanted regardless of their eABR rating during ANTS. Five raters were enrolled in our study, and asked to blindly retrospectively rate these traces. The traces provided were the most accurate ones after tumour removal, to eliminate all the biases caused by intraoperative disturbances. The raters had to score the traces according to Walton score. The following outcomes of agreement between raters were measured: concordance, number of assigned categories (NOAC), inter-rater agreement, intra-rater agreement. The results were then compared with the accuracy values found in the work by Zanoletti-Concheri et al. Results Inter-rater agreement of eABR traces assessment according to Kendall coefficient of concordance (W) was calculated both for the 2 and 3 category score. For the 3-category classification the score was W=0.71 (substantial agreement), while for the 2-category one the score was W=0.60 (moderate agreement). Intra-rater agreement, calculated based on the Cohen’s weighed kappa (κ), was, according to Landis and Koch 1977 interpretation, almost perfect for the 2-category classification, substantial for the 3-category classification. In this study the ability of predicting the CI performance based on the eABR interpretation showed: a sensibility of 0.571, a specificity of 0.667, a PPV of 0.667, a NPV of 0.571, and an accuracy of 0.615 Conclusion Our result show that eABR traces evaluation, even though the raters differ in experience and expertise, has an acceptable degree of inter rater and intra rater concordance. The values were generally better when considering a binary evaluation of the traces, so a simplified version of the Walton score might be used in the future, to further increase the agreement among raters. When compared to the results from the previous literature, this study demonstrates that there is indeed a bias in interpreting eABR traces, deriving from environmental factors. We redefined the values of sensitivity, specificity, PPV, NPV, accuracy compared to the previous work from Zanoletti-Concheri et al. The redefinition of the ANTS accuracy values shows that there is the tendency of a more optimistic interpretation of traces, when the biases are reduced, with possible economical negative repercussions on the Health System.