Background and Purpose— Patients with intracerebral hemorrhage (ICH) are often subject to rapid deterioration due to hematoma expansion. Current prognostic scores are largely based on the assessment of baseline radiographic characteristics and do not account for subsequent changes. We propose that calculation of prognostic scores using delayed imaging will have better predictive values for long-term mortality compared with baseline assessments. Methods— We analyzed prospectively collected data from the multicenter PREDICT study (Prediction of Hematoma Growth and Outcome in Patients With Intracerebral Hemorrhage Using the CT-Angiography Spot Sign). We calculated the ICH Score, Functional Outcome in Patients With Primary Intracerebral Hemorrhage (FUNC) Score, and modified ICH Score using imaging data at initial presentation and at 24 hours. The primary outcome was mortality at 90 days. We generated receiver operating characteristic curves for all 3 scores, both at baseline and at 24 hours, and assessed predictive accuracy for 90-day mortality with their respective area under the curve. Competing curves were assessed with nonparametric methods. Results— The analysis included 280 patients, with a 90-day mortality rate of 25.4%. All 3 prognostic scores calculated using 24-hour imaging were more predictive of mortality as compared with baseline: the area under the curve was 0.82 at 24 hours (95% CI, 0.76–0.87) compared with 0.78 at baseline (95% CI, 0.72–0.84) for ICH Score, 0.84 at 24 hours (95% CI, 0.79–0.89) compared with 0.76 at baseline (95% CI, 0.70–0.83) for FUNC, and 0.82 at 24 hours (95% CI, 0.76–0.88) compared with 0.74 at baseline (95% CI, 0.67–0.81) for modified ICH Score. Conclusions— Calculation of the ICH Score, FUNC Score, and modified ICH Score using 24-hour imaging demonstrated better prognostic value in predicting 90-day mortality compared with those calculated at presentation.