Liver tumor volume measurements are clinically useful in patients undergoing cancer treatment. The techniques of planimetry and stereology were applied for this purpose on magnetic resonance (MR) imaging. Fifty-eight malignant liver lesions were depicted on MR images in 20 consecutive patients. The volume of all lesions was estimated using stereology technique, based on point counting. Stereological tumor volume estimations were compared with those determined by manual planimetry. The repeatability of both techniques was assessed. Tumor volumes estimated by the two techniques were highly correlated (r = 0.98, p < 0.0001). The 95% limits of agreement showed that the stereological volume estimations may differ from the planimetric assessments by less than 23%. Both techniques presented comparable intra- and interobserver variability. The planimetry was 1.5 times faster than the stereology. Both volumetric techniques may provide reliable and reproducible liver tumor volume estimations. The planimetry may be the method of choice because of its superior speed.