Background: Early intracerebral hemorrhage (ICH) expansion is a major determinant of poor clinical outcome. We previously reported baseline hematoma volume was a predictor of hematoma expansion (HE), and that hematomas <3mL may represent a subgroup with good prognosis. Our objective was to validate our previous findings in a multi-centre prospective observational cohort, and to assess the relationship between baseline hematoma size and the CTA spot-sign. We hypothesized that small hematomas are less likely to expand, and have low spot-sign prevalence. Methods: The PREDICT study is a prospective, observational cohort study of consecutive patients with acute ICH. Inclusion criteria are age>18, symptom onset <6 hours, and baseline CT and CTA; exclusions are baseline ICH >100ml, planned ICH surgery within 24 hours, known secondary cause of ICH, known renal impairment, GCS<6, or premorbid disability or terminal illness. Scans were reviewed for spot sign presence/absence by a neuroradiologist blinded to outcomes and follow-up imaging. Volumes were measured by planimetry by a neurologist blinded to CTA images and outcomes. The predictor of interest was baseline hematoma volume which was stratified as <3mL, 3-9mL, 10-19mL, 20-29mL and >30mL based on our prior study. Primary outcome was significant HE defined as ≥6mL. We used multivariable models to calculate adjusted odds ratios (aOR) for HE. Findings: Two-hundred and sixty-eight patients were enrolled from 11 centers in 6 countries: HE analysis was limited to 228 patients with follow-up CT before rFVIIa or surgical intervention. Median baseline hematoma volume was 12.4ml, spot-sign was present in 26.8% of patients, and 25% of patients had HE of ≥6ml. HE and spot sign prevalence increased with increasing baseline hematoma volume (see table ) . Only one patient with volume <3ml had HE; the patient was on warfarin (INR 2.2) but spot negative. Two patients with volumes <3ml were spot positive, but neither had HE. When compared to hematomas >30ml, the aOR for HE was 0.09 for <3ml hematomas, 0.14 for 3-9ml, 0.49 for 10-20ml, and 1.83 for 20-30ml (p<0.001). Associations between baseline hematoma volume and clinical outcomes will be presented. Discussion: Our results validate baseline hematoma volume as a predictor of HE. Furthermore, spot sign prevalence is associated with baseline hematoma volume. These results can inform ICH trial design and clinical prognostication at the bedside: small hematomas have a low spot sign prevalence and are unlikely to expand ≥6 ml, even when spot positive. Conversely, half of hematomas >30ml are spot positive and will expand.