NIMG-23A CONNECTOMIC PHENOTYPE OF 1p/19q LOSS OF HETEROZYGOSITY IN LOW-GRADE GLIOMA

The efficiency of a healthy brain network (the ‘connectome’) is positively related to physical and cognitive function. We sought to determine whether changes in network efficiency were related to loss of heterozygosity (LOH) at 1p/19q, a favorable prognostic marker in low-grade glioma (LGG). The presurgical diffusion tensor imaging (DTI) of LGG patients (WHO grade II) with known 1p/19q status treated at MGH were retrospectively analyzed. Processing included co-registration with anatomical 3D images, probabilistic tractography with FSL, and binary network construction based on the number of streamlines connecting each parcel from the Automated Anatomical Labeling atlas for which a benchmark healthy connectome (HC) is available. Path length (the average shortest number of steps between each network region) of the whole brain and the hemisphere contralateral to the tumor were compared to the HC by one-sample Wilcoxon signed rank tests. Path lengths of patients with and without LOH were compared using Mann-Whitney U tests. Thirteen patients with images acquired between 2008-2014 were included (8 male, mean age 42 ± 14 years, 8 left-sided tumors, median KPS 90). Four patients showed LOH. At the group level, patients had longer whole-brain path length (2.5 ± 0.3 steps, p = 0.023) and contralateral path length (2.0 ± 0.2 steps, p = 0.033) than the HC (2.3 and 1.9 steps, resp.), indicating a less efficient connectome. Patients with LOH had shorter contralateral path length than those without (2.5 ± 0.3 versus 2.3 ± 0.1 steps, p = 0.011). Age, gender, presurgical tumor volume, and performance status did not influence these results significantly. These results indicate that patients with LGG have decreased connectome efficiency even in the contralateral hemisphere, highlighting the global impact of local tumors. 1p/19q LOH is associated with a more efficient connectome, supporting the more favorable performance status and prognosis in these patients. Connectivity is therefore a promising noninvasive biomarker for prognosis and holds potential for treatment monitoring.