Weekly hippocampal diffusion changes during partial-brain radiation therapy for high-grade glioma: associations with dose and early memory change.

PURPOSE: To characterize the weekly trajectory of hippocampal diffusion during fractionated partial-brain radiation therapy (RT) for high-grade glioma and determine whether the magnitude of change is associated with planned hippocampal dose.

METHODS

Forty-five adults enrolled in a prospective observational imaging study and treated with 60 Gy in 30 fractions with concurrent temozolomide underwent diffusion MRI before RT and weekly during treatment. Relative bilateral hippocampal apparent diffusion coefficient (ADC) change and fractional anisotropy (FA) were evaluated longitudinally. Linear mixed-effects models included treatment week and planned mean bilateral hippocampal dose as fixed main effects. A secondary laterally-specific analysis evaluated baseline-to-Week-6 ADC change separately in the hippocampi ipsilateral and contralateral to the tumor.

RESULTS

Hippocampal ADC increased progressively from a median relative change of 0.5% at Week 1 to 12% (IQR, 6%-18%) at Week 6. The ADC time coefficient was + 2.2% points per week (95% CI, 1.8-2.5; P < 0.001). After accounting for week, higher planned mean bilateral hippocampal dose was associated with greater ADC change (+ 0.81% points/Gy; 95% CI, 0.76-0.93; P < 0.001). In the secondary laterally-specific analysis, higher dose was associated with greater Week-6 ADC change in both the ipsilateral hippocampus (β = 0.83% points/Gy; 95% CI, 0.52-1.26; P = 0.013) and the contralateral hippocampus (β = 0.36% points/Gy; 95% CI, 0.05-0.712; P = 0.046). FA remained stable through Week 3 and declined during Weeks 4-6.

CONCLUSION

Weekly diffusion MRI demonstrates a progressive, dose-associated hippocampal response during fractionated RT. The laterality-specific findings further support an association between local radiation exposure and ADC change and support hippocampal ADC as a candidate treatment-phase imaging phenotype for prospective validation.