Multi-subregional susceptibility patterns in gliomas using sub-voxel quantitative susceptibility mapping

Objective To evaluate glioma susceptibility patterns and their clinical/pathological correlations using a sub-voxel quantitative susceptibility mapping (QSM) technique that quantifies the paramagnetic (χ para ) and diamagnetic (χ dia ) components (APART-QSM). Materials and methods We enrolled 120 pathologically confirmed glioma patients ( n = 120; aged 47 ± 13 years, mean ± standard deviation; 51 women) retrospectively and 6 for histologic validation prospectively. χ para /χ dia maps were generated. Subregions included peritumoral edema (PE), subdivided into four equidistant subregions, tumor parenchyma (TP), and necrosis, cystic degeneration (CD). Associations between multi-subregional susceptibility and clinicopathological parameters and a conditional inference tree (CIT) model for genotyping were analyzed. Results Compared to contralateral normal-appearing tissue (CNAT), TP in high-grade and isocitrate dehydrogenase (IDH)-wild-type gliomas exhibited increased χ para ( p ≤ 0.044), low-grade and IDH-mutant gliomas lower χ bulk , χ para and χ dia ( p ≤ 0.015). PE demonstrated lower susceptibility than CNAT ( p < 0.001), with χ para /χ dia initially decreasing then increasing with peritumoral distance ( p ≤ 0.008). Necrosis showed elevated χ para /χ dia ( p ≤ 0.025), and CD in oligodendrogliomas displayed higher χ dia than CNAT ( p = 0.019). The CIT model (TP χ dia and PE χ para ) achieved 0.702 accuracy. Univariable analysis showed that higher TP χ para /χ dia correlated with shorter survival ( p < 0.001). In the prospective subgroup, TP χ para correlated with iron content ( r = 0.901, p = 0.014), TP χ dia /χ para correlated with Ki-67 ( r = 0.841, p = 0.036). Conclusion APART-QSM revealed distinct, clinically relevant susceptibility profiles across glioma subregions with potential associations with genotyping and prognosis, providing biomarkers for iron content and tumor proliferation. Distance-associated susceptibility pattern in PE may reflect peritumoral microenvironmental properties. Key Points Question Gliomas exhibit complex histopathology with heterogeneous magnetic susceptibility. Effective preoperative evaluation of multi-subregional susceptibility patterns in gliomas facilitates refined tissue characterization and guides clinical treatment. Findings Glioma subregions showed clinically relevant susceptibility profiles, correlating with iron content, tumor proliferation, and revealing a distance-associated peritumoral edema pattern. Relevance statement Our study used the APART-QSM technique to characterize multi-subregional susceptibility patterns and pathological changes in different glioma types, revealing its potential association with glioma genotyping and prognosis.