Fractal radiomics as a virtual biopsy in glioma: A systematic review and meta-analysis of molecular status, radiogenomic associations, and survival outcomes.

In 30 seconds

This systematic review and meta-analysis assessed whether fractal-based imaging features can predict molecular alterations and outcomes in glioma. For IDH mutation prediction, the pooled diagnostic effect size was 2.10 (95 % CI: 1.20-3.01); reported fractal-dimension AUC values reached 0.97. TERTp prediction had a pooled effect size of 1.29, with peritumoral edema the most predictive region. For overall survival, the pooled HR was 2.10 (95 % CI: 1.48-2.97), indicating an association between specified patterns of geometric complexity and higher mortality risk. The authors propose fractal analysis as a potential non-invasive complement to conventional imaging.

Key findings

• IDH mutation prediction had a pooled diagnostic effect size of 2.10 (95 % CI: 1.20-3.01).
• Fractal dimension of tumor contours and enhancing subcomponents had AUC values up to 0.97.
• TERTp prediction had a pooled effect size of 1.29; peritumoral edema was the most predictive region of interest.
• The pooled overall-survival HR was 2.10 (95 % CI: 1.48-2.97) for reported high-complexity imaging patterns.

Why it matters

Molecular classification and prognosis are central to glioma assessment, while tissue sampling can be invasive and may not capture intratumoral heterogeneity. Fractal imaging features may provide additional information about molecular status and survival from imaging data. The pooled findings support further evaluation of this approach as a complement to, rather than a replacement for, established diagnostic assessment.

What to keep in mind

The abstract does not report the number of included studies or patients, and it provides no pooled performance details beyond the reported effect sizes and AUC. The findings are based on meta-analysis of imaging associations and do not establish clinical utility.