Abstract
AIMS: Measurements of apparent water exchange time (τex) have been linked to progression of brain metastases in studies using dynamic-contrast-enhanced MRI. We used an alternative, time dependent diffusion-weighted MRI (dMRI), to infer τex without contrast agents by characterizing the diffusion time dependence of the kurtosis (K) of endogenous water molecules diffusing in brain tissue. In this study, we hypothesize that τex estimates from diffusional kurtosis time dependence may have the same sensitivity to tumour progression after treatment as the ‘Response assessment in neuro-oncology brain metastases (RANO-BM)’ criteria and examine the relation- ship between changes of τex in brain metastasis after stereotactic radiosurgery (SRS) and long term treatment response in a cohort of patients with different primary cancers. METHODS: 16 patients with brain metastases were scanned at baseline, 1-month and 3-month timepoints. Patients received SRS between baseline and 1-month. τex was calculated for each patient at each timepoint. Treatment response was assessed using RANO-BM criteria. RESULTS: In the good treatment response group (RANO-BM=1,2), the percentage change in τex from baseline (Δτex(%)) at 1 and 3 months was negative for all but one patient. In the poor treatment response group (RANO-BM=3,4), Δτex(%) at 1 and 3 months was positive, except for one patient. All patients whose cause of death was not “Intracranial progression” had a negative Δτex(%) at 1 or 3 months except for one patient. All the others had a positive Δτex(%) at 1 or 3 months CONCLUSION: Except for two patients with different clinical factors (necrosis/immunotherapy), Δτex(%) was associated with the RANO-BM criteria. Long term treatment response was associated with Δτex(%) at 1 month. This suggests τex could help predict response to SRS. Our cohort was small and heterogeneous; future work will apply this analysis to larger groups.