CLINICAL AND MAGNETIC-RESONANCE IMAGE CORRELATION IN IDIOPATHIC CEREBELLAR-ATAXIA

Sixty-one patients who fulfilled the clinical criteria for idiopathic cerebellar ataxia and who had symptoms at least for 3 years were exami ned clinically and by magnetic resonance imaging (MRI). Based on the c linical signs, they were divided into patients with pure cerebellar si gns (Group 1), patients with additional mild rigidity and/or hyperrefl exia (Group 2) and patients with additional severe rigidity and hypoki nesia (Group 3). Patients in Group 1 had milder disability and better prognosis than patients in Group 2 or Group 3 (ataxic score: 14.9 vs. 28.6 and 36.0; annual progression ratio: 0.26 vs. 0.65 and 0.70, respe ctively). We measured the area of the cerebellar vermis, ventral pens and dorsal brainstem on midsagittal T1-weighted MR images for all pati ents and age-and sex-matched controls. The cerebellar vermis as well a s the ventral pens of patients were significantly smaller than corresp onding structures in controls (p < 0.001). The ventral pens of patient s in Group 2 and Group 3 was significantly smaller than that of patien ts in Group 1 (p < 0.0001, respectively), and the dorsal brainstem of patients in Group 2 and Group 3 was also significantly smaller than th at of patients in Group 1 (p < 0.001, respectively). The ventral pens of patients in Group 3 was significantly smaller than that of patients in Group 2 (p < 0.05) as well. There was a significant correlation be tween the area of the ventral pens and the annual progression ratio (p < 0.001). With MRI, slight but definite hyperintensities were demonst rated in the pontine base and the medulla of 22 patients on proton den sity images. In the longitudinal study, patients in Group 2 and Group 3 had atrophy of the ventral pens already at an early stage. The ventr al pens of patients in Group 3 was smaller at the initial MR examinati on than that of patients in Group 2. These observations suggest that p atients with smaller ventral pens may have rapid progression and poor prognosis. Thus, even a relatively simple quantitation of the area of the ventral pens may be useful to predict the prognosis of patients, i n addition to neurologic assessment at intervals.