PURPOSE: To elucidate the pitfalls of Tc-99m RBC liver SPECT in the diagnosis of hepatic hemangioma, and to compare this modality with two-phase dynmic CT. MATERIALS AND METHODS: Forty patients with 48 liver masses, suspected on ultrasonography to be hemangiomas, underwent two-phase dynamic CT scanning and SPECT within a two week period. All masses were diagnosed through pathologic and follow up radiologic studies. The final diagnoses were hemangioma (n=42), metastasis (n=2), abscess (n=2), hepatocellular carcinoma (n=1), and cholangiocarcinoma (n=1). Sensitivities and specificities of CT and Spect for the diagnosis of hemangioma and the relationship between false positives or false negatives seen on SPECT and the Pattern of contrast enhancement seen on CT were investigated. RESULTS: The sensitivities of CT and SPECT for the diagnosis of hemangioma were 95.2 (40/42) and 76.2% (32/42), respectively. The false-negative rate of SPECT was significantly higher in the early enhancing (54.5%, 6/11) than in the late enhancing type (13.8%, 4/29) and in the 'less than 1cm' group, false-negatives (70%, 7/10) were more common than in the 'more than 1cm' group (9.4%, 3/32). for the two lesions with false-positive findings on SPECT, the final diagnosis was metastasis, and two false-negative lesion, seen on CT, were misread as metastases. Four other lesions were negative in both studies. CONCLUSION: For the detection of hepatic hemangioma, two-phase dynamic CT is a better modality than SPECT. False positives on SPECT occurred in metastasis, and false negatives are more common in the small lesion and early enhancing group than in the late enhancing group. Between the two groups, there is a difference in hemodynamics, and considerable further pathological investigation is needed.