Abstract
BACKGROUND: Currently, the most common and serious delayed complication of cerebral RT is radiation-induced leukoencephalopathy (RIL), roughly defined, in absence of strictly consensual criteria, by cognitive dysfunctions associated with diffuse FLAIR MRI white matter hyperintensity (HI). The objective of our study is to clearly define the RIL criteria. METHODS: Patients, addressed to our center between June 2010 and June 2012, with medical history of focal or whole brain RT who developed a persistent cognitive complain at least six months after the end of the brain RT, and presenting brain MRI FLAIR abnormalities were eligible. There was a clinical evaluation by a neurologist and a neuropsychological assessment (Edinburgh scale for lateralization, educational level, Mattis dementia rating scale (DRS) and Folsteins'MMSE, Grober and Buschke memory evaluation, forward and backward digit span, phonemic and semantic verbal fluencies, DO80). All patients were scanned by a standardized MRI protocol. White matter FLAIR lesions were evaluated with a 1 to 14 points score according to a modified Scheltens rating scale. Brain atrophy was visually assessed. Consensually defined signs of small vessels diseases (CSVD) were assessed for each patient. RESULT: Fourty patients were enrolled in our study (mostly stage 3 gliomas (27.5%)). The mean fraction dose was 2 Gy (range 1,5-3) and a mean total dose of 49 Gy (range 30-60). Most patient had gait difficulties, urinary trouble and cognitive complain. Memory difficulties were reported by 77% of patients. Neuropsychological assessment showed global and severe cognitive decline. Mattis DRS was abnormal in 75% of the patients. Forward and backward digit span were pathological in 80% patients. On brain MRI, an extensive FLAIR HI with anterior predominance was shown in 68,5% patients. Brain atrophy was present in 87% cases. T2*-weighted- MRI showed hypointensities like “black dots” probably corresponding to microbleeds or radiation induced cavernomas in 52,6% patients. We found no recent small cortical infarct. We found no correlation between FLAIR HI extension, brain atrophy, microbleeds numbers or location, perivascular spaces number, lacunes of presumed vascular origin and the cognitive compound score. We found no correlation between the delay from RT and one of the MRI lesions described above. The patients displayed a loss of executive functions responsible of severe cognitive disability and dementia as previously described in patients with CSVD. Our study give strong clinical and MRI arguments to consider that the RIL is a delayed progressive CSVD, with characteristic evolution and distribution pattern namely implicating distal brain superficial perforating arterioles that serve dorsolateral loops of white matter in the corona radiata and sparing basal ganglia perforating arterioles that spring from the cerebral medial artery.