Central precocious puberty occurs as a result of premature pituitary stimulation and increased secretion of gonadotropins. The aims of this study were to analyze MR imaging findings in the pituitary glands of children with central precocious puberty compared with matched control subjects, to define MR imaging-derived variables useful in the diagnosis of central precocious puberty, and to correlate MR imaging-derived variables with the hormonal profile and other imaging and clinical findings. Twenty-six children with central precocious puberty (two boys and 24 girls) were divided into two subgroups according to MR imaging findings: idiopathic (21 patients) and nonidiopathic (five patients: three hypothalamic hamartomas, one pineal tumor, one empty sella syndrome). The control group consisted of 17 normal age- and sex-matched children (two boys, 15 girls). Analyzed parameters included pituitary height, length, width, midsagittal cross-sectional area, calculated volume, and shape. The shape was assessed by a pituitary grading system and two other shape indexes (length-to-height and length-to-width ratios). Pituitary grade was defined by the concavity of the upper pituitary surface (grade 1 = marked concavity, grade 2 = mild concavity, grade 3 = flat, grade 4 = mild convexity, grade 5 = marked convexity). Pituitary grade showed a highly significant difference among groups (p < .001). Area, height, and length-to-height ratio were significantly different (p < .05), whereas length, width, length-to-width ratio, and volume were not. There was no significant difference in any of the variables compared between idiopathic and nonidiopathic groups. When selected variables (pituitary grade, area, height, length) in the central precocious puberty group were stratified by bone age and findings on pelvic sonograms, patients with advanced bone age had a significantly higher pituitary grade (p < .01) and had a tendency toward a greater pituitary length. Pituitary size and shape correlated with the hormonal profile. Change in pituitary grade is the most helpful variable for the diagnosis of central precocious puberty in a prepubertal child. A high pituitary grade (4 or above) is highly predictive of central precocious puberty, with the highest specificity and positive predictive value, but with low sensitivity. The use of combinations of high pituitary grade with two other positive findings (height and area greater than 1 SD from the respective means in the control group) improves the sensitivity, specificity, and predictive value of MR imaging in the diagnosis of central precocious puberty.