Room: Exhibit Hall
Purpose: The aim of this study was to determine whether radiomic features measured at baseline in Magnetic Resonance images (MRI) of acoustic neuromas (AN) can be used to predict treatment failure following Gamma Knife (GK) treatment. Validation based on audiology hearing examinations was a secondary goal.
Methods: The study was conducted on pre-GK and post-GK MRI-T2 scans obtained from thirty two patients who underwent stereotactic radiosurgery (SRS) for AN between 2010 and 2015. The prescription dose was 12 Gy to 50 % isodose volume. Radiomic features extracted using an in-house software include features based on Intensity, Fractals, Laplacian of Gaussian and textural Co-Occurrence, Run-length (RL), Size Zone, and Neighborhood Gray-Tone Difference matrices (NGTDM). Study subjects were classified as treatment failures (TF) if their tumor volume increased >10% following GK. Audiology reports acquired before and after SRS were used to validate against treatment outcome independently. Audiology report findings were graded on a Gardner-Robertson (GR) hearing scale based on pure-tone average and word recognition score.
Results: The mean Â± standard deviation (SD) time interval between MRI examinations before and after SRS was 207Â±72 days (range: 88â€“368 days). Fifteen subjects (47%) qualified as TFs. In univariate receiver operating characteristic (ROC) analysis, two radiomic features with highest areas under curves (AUC) were complexity in NGTDM (AUC>0.7) and run percentage in RL (AUC>0.65). Based on 14 pairs of audiology reports, univariate correlation analysis was insufficiently powered to establish a relationship between GR hearing score and treatment outcome.
Conclusion: This initial radiomic study establishes two features that illustrate the prognostic ability of the SRS treatment in AN. Independent method of validation of treatment technique based on the audiological findings were insufficiently powered to establish a relationship with SRS-based treatment outcome of AN.
IM/TH- Image Analysis (Single modality or Multi-modality): Computer-aided decision support systems (detection, diagnosis, risk prediction, staging, treatment response assessment/monitoring, prognosis prediction)