Apparent diffusion coefficient for genetic characterisation of untreated adult gliomas: a meta-analysis stratified by methods

Background: Isocitrate dehydrogenase (IDH) mutation and chromosome 1p19q genotyping have become fundamental to the prognostic grouping of adult diffuse gliomas. Apparent diffusion coefficient (ADC) values may enable non-invasive prediction of glioma molecular status. The purpose of this systematic review and meta-analysis was to investigate the diagnostic accuracy of ADC for IDH and 1p19q genotyping, considering measurement techniques and tumour grade. // Methods: A systematic search of PubMed and Cochrane Library databases was performed in December 2024. Studies were grouped according to the ADC parameter measured and the measurement techniques used. A meta-analysis was performed, supplemented by Egger’s regression testing. The quality of studies was assessed with the QUADAS-2 tool. // Results: Thirty-three studies including a total of 4297 patients fulfilled the inclusion criteria. IDH mutation and 1p19q deletion status were assessed by 30 and 14 studies respectively. Pooled area under the curve (AUC) values for the prediction of an IDH mutation and 1p19q codeletion ranged from 0.743 (0.680-0.805) to 0.804 (0.689-0.919), and 0.678 (0.614-0.741) to 0.692 (0.600-0.783). No significant differences were identified between regional and volumetric measurements, between ADCmean and ADCmin values, or comparing normalised and raw ADC data. // Conclusion: This meta-analysis supports ADC as an imaging biomarker in untreated gliomas, specifically to predict IDH status. ROI measurement, particularly by a single ADCmean, is rapid, reproducible and appears statistically equivalent to volumetric read outs. We found no evidence for superior diagnostic accuracy by ADC normalisation. Published ADC thresholds have been summarised for consideration of prospective testing across institutions.