Ten Years of VASARI Glioma Features: Systematic Review and Meta-Analysis of Their Impact and Performance
Aynur Azizova, Yeva Prysiazhniuk, Ivar J. H. G. Wamelink, Jan Petr, Frederik Barkhof, Vera C. Keil
Abstract
<h3>BACKGROUND:</h3> Visually Accessible Rembrandt (Repository for Molecular Brain Neoplasia Data) Images (VASARI) features, a vocabulary to establish reproducible terminology for glioma reporting, have been applied for a decade, but a systematic performance evaluation is lacking. <h3>PURPOSE:</h3> Our aim was to conduct a systematic review and meta-analysis of the performance of the VASARI features set for glioma assessment. <h3>DATA SOURCES:</h3> MEDLINE, Web of Science, EMBASE, and the Cochrane Library were systematically searched until September 26, 2023. <h3>STUDY SELECTION:</h3> Original articles predicting diagnosis, progression, and survival in patients with glioma were included. <h3>DATA ANALYSIS:</h3> The modified Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool was applied to evaluate the risk-of-bias. The meta-analysis used a random effects model and forest plot visualizations, if ≥5 comparable studies with a low or medium risk of bias were provided. <h3>DATA SYNTHESIS:</h3> Thirty-five studies (3304 patients) were included. Risk-of-bias scores were medium (<i>n</i> = 33) and low (<i>n</i> = 2). Recurring objectives were overall survival (<i>n</i> = 18) and isocitrate dehydrogenase mutation (<i>IDH</i>; <i>n</i> = 12) prediction. Progression-free survival was examined in 7 studies. In 4 studies (glioblastoma <i>n</i> = 2, grade 2/3 glioma <i>n</i> = 1, grade 3 glioma <i>n</i> = 1), a significant association was found between progression-free survival and single VASARI features. The single features predicting overall survival with the highest pooled hazard ratios were multifocality (hazard ratio = 1.80; 95%-CI, 1.21–2.67; I<sup>2</sup> = 53%), ependymal invasion (hazard ratio = 1.73; 95% CI, 1.45–2.05; I<sup>2</sup> = 0%), and enhancing tumor crossing the midline (hazard ratio = 2.08; 95% CI, 1.35–3.18; I<sup>2</sup> = 52%). <i>IDH</i> mutation-predicting models combining VASARI features rendered a pooled area under the receiver operating characteristic curve of 0.82 (95% CI, 0.76–0.88) at considerable heterogeneity (I<sup>2</sup> = 100%). Combined input models using VASARI plus clinical and/or radiomics features outperformed single data-type models in all relevant studies (<i>n</i> = 17). <h3>LIMITATIONS:</h3> Studies were heterogeneously designed and often with a small sample size. Several studies used The Cancer Imaging Archive database, with likely overlapping cohorts. The meta-analysis for <i>IDH</i> was limited due to a high study heterogeneity. <h3>CONCLUSIONS:</h3> Some VASARI features perform well in predicting overall survival and <i>IDH</i> mutation status, but combined models outperform single features. More studies with less heterogeneity are needed to increase the evidence level.