BACKROUND: Hematoma volume is the most important predictor of outcome following intracerebral hemorrhage (ICH). A number of predictive models have been developed to identify patients with unfavorable prognosis after ICH. It is not known whether they are superior to predictions based solely on hematoma volume. In this study we assessed the predictive accuracy of hematoma volume and two widely used prognostic models - ICH scale and ICH grading scale (ICH-GS) - in determining in-hospital mortality in patients with ICH. METHODS: A consecutive series of patients with ICH and cranial computed tomography imaging within 24 hours of symptom onset were included into the study. Hematoma volume was calculated by using the ABC/2 method. ICH scale and ICH-GS scores of each patient were determined. The outcome variable was in-hospital mortality. The overall predictive ability of hematoma volume, ICH scale and ICH-GS were assessed by computing the area under the receiver-operating characteristic curves (AUC). RESULTS: Among a total of 193 (116 male, 77 female, mean age 66±13 years) patients included in the study, 75 (39%) died during hospitalization. The AUC was 0.761 (95%CI: 0.694-0.819) for hematoma volume, 0.797 (95%CI: 0.733-0.852) for ICH score and 0.784 (95%CI: 0.719-0.840) for ICH-GS (p>0.05). On the other hand the smallest hematoma volume at which the positive predictive value for mortality was 100% was 130 mL. CONCLUSION: Hematoma volume, by itself, is not inferior to other prognostic models in predicting in-hospital mortality in patients with ICH.