Spleen size was assessed in 73 patients with thrombocytosis and in 15 healthy subjects, comparing palpation with ultrasonography (US) measurement of longitudinal diameter and volume. Intraobserver and interobserver variability for volume on US, checked in 12 patients, was very low. Correlation between spleen volume measured by US and that measured by computed tomography was excellent. Splenomegaly was detected by palpation in 25% of patients, by US assessment of longitudinal diameter in 33%, and by US assessment of volume in 52%. After diagnostic work-up, 54 patients had a diagnosis of essential thrombocythemia (ET), 4 of idiopathic myelofibrosis (IMF), and 15 of secondary thrombocytosis (ST). Spleen volume in patients with ST was in the normal range (138 +/- 47 mL) and was significantly lower than that in patients with ET or IMF (370 +/- 210 mL; P <.001). Thus, US-measured volume was the most sensitive method for identifying nonpalpable splenomegaly in patients with primary myeloproliferative diseases, and it may help in distinguishing these diseases from reactive disorders.