Structural network analysis in patients with IDH-mutant gliomas grade 2 and 3
Jan Kückelhaus (Erlangen), Alfred Gramelt (Erlangen), Endrit Vinca (Erlangen), Ludwig Singer (Erlangen), Nico Teske (Erlangen), Arnd Dörfler (Erlangen), Daniel Delev (Erlangen), Oliver Schnell (Erlangen), Dieter-Henrik Heiland (Erlangen), Philipp Karschnia (Erlangen)
Following resection, patients with IDH-mutant gliomas of grade 2 and 3 are managed with watch-and-scan strategies, adjuvant radiotherapy or chemotherapy, or combined radiochemotherapy. While aggressive treatment may improve survival, there is emerging concern about delayed toxicity following CNS-directed radiotherapy which may eventually translate into substantial cognitive deficits. This study aims to assess whether structural networks may serve as objective imaging-based biomarkers for treatment-induced neurodegeneration.
We retrospectively identified 126 patients with newly diagnosed IDH-mutant gliomas of grade 2 and 3, and cohorts were stratified according to the adjuvant first-line treatment (N chemo = 6; radio = 7; radiochemo = 76; watch-and-scan = 47). Anatomic areas were assigned according to the Aparc parcellation atlas and subsequently characterized by quantifying cortical thickness, surface area, volume, gyral curvature and sulcal depth using FreeSurfer (preliminary: n = 32). Structural similarity networks of the unaffected hemisphere were estimated by Morphometric Inverse Divergence (MIND).
Our proposed workflow focused exclusively on the non-lesioned hemisphere, eliminating confounding effects from the tumor lesion or the morphology of the resection cavity. This enabled reliable quantification of the cortical architecture from standard MR sequences and precise estimation of the cortical connectome. Analysis of post-treatment MRIs revealed increased cortical surface areas in memory and cognition-critical regions, including the parahippocampal gyrus, posterior cingulate cortex, and orbitofrontal cortex (all p < 0.005), in patients undergoing radiotherapy compared to those who did not. These increases may reflect radiation-induced processes such as myelination or gliogenesis, which could contribute to the observed cortical remodeling.
Our findings support the use of structural similarity networks as biomarkers for treatment-induced neurodegeneration in glioma patients. If validated in larger datasets, our preliminary findings suggest that surgery alone better maintains the physiological architecture of the healthy hemisphere compared to radiotherapy. It may therefore appear relevant to identify individuals who can be safely managed with resection alone without compromising survival.
Auf unserem Internetauftritt verwenden wir Cookies. Bei Cookies handelt es sich um kleine (Text-)Dateien, die auf Ihrem Endgerät (z.B. Smartphone, Notebook, Tablet, PC) angelegt und gespeichert werden. Einige dieser Cookies sind technisch notwendig um die Webseite zu betreiben, andere Cookies dienen dazu die Funktionalität der Webseite zu erweitern oder zu Marketingzwecken. Abgesehen von den technisch notwendigen Cookies, steht es Ihnen frei Cookies beim Besuch unserer Webseite zuzulassen oder nicht.