Early Temporal Cortex Changes During Radiotherapy in Cortical Mean Diffusivity and Thickness Derived From MRI

Temporal lobe (TL) damage is a common complication of radiotherapy (RT) for nasopharyngeal carcinoma (NPC), yet early morphological changes during RT remain poorly characterized. Identifying these early changes could facilitate timely interventions to mitigate long-term cognitive impairments.

To characterize early TL structural changes during RT using cortical thickness (CTh) and cortical mean diffusivity (cMD), and to explore potential dose-dependent relationships.

Prospective, longitudinal.

40 patients with NPC (5 females, 12.5%, mean age 52.7 years).

3 T, diffusion tensor imaging (DTI) (spin echo echo-planar imaging sequence); T1-weighted images (magnetization-prepared rapid gradient echo (MPRAGE) sequence).

MRI was acquired at three phases: pre-treatment (tp₁), mid-treatment (tp₂), and post-treatment (tp₃). At each time point, CTh and cMD were quantified from 18 bilateral TL regions defined by the Desikan-Killiany atlas using surface-based analysis. Radiation dose was mapped onto these regions using CT-MRI co-registration.

Linear mixed-effects models assessed cortical changes across treatment phases (tp₁-tp₂ [P1], tp₂-tp₃ [P2], tp₁-tp₃ [P3]). Dose-response relationships were evaluated using Pearson correlations and permutation testing. Family-wise error (FWE) correction (Bonferroni method) was applied, with significance set at corrected p < 0.05.

Across treatment, cMD significantly increased while CTh significantly decreased, particularly evident at the final post-treatment phase (P3: CTh in 15/18 regions; cMD in 8/18 regions). Early significant changes (P1) appeared in fewer regions (CTh: 3/18; cMD: 5/18). Dose-dependent analysis revealed small-to-moderate correlations; after strict FWE correction, significant associations remained only in the right banks of the superior temporal sulcus (cMD, r = 0.57) and left inferior temporal gyrus (CTh, r = -0.44) at P1.

RT-induced TL structural changes emerge early and intensify progressively. Despite limited statistical power, the early dose associations highlight the need for larger studies to further elucidate early dose-response dynamics. EVIDENCE LEVEL: 2.