We have used confocal laser microscopy and a novel “voxel”-based imaging software to study the dynamics of platelet aggregation and thrombus formation when anticoagulated blood was perfused over collagen-coated surfaces at shear rates simulating arterial flow. The objective was to evaluate the three-dimensional growth of platelet thrombi over time (“4-D” imaging). Blood from healthy donors, anticoagulated with either PPACK (80 μM) or, depending on the type of experiment, with trisodium citrate (11 mM), was incubated with mepacrine (10 μM) to render platelets fluorescent. Blood was aspirated with a syringe pump through a rectangular perfusion chamber (flow path height of 80 μm) at a flow rate of 160 or 480 μl per min to provide initial shear rates of 500 or 1,500 sec−1, respectively. Prior to perfusion, glass coverslips were coated with fibrillar type I collagen (Roche Diagnostics, Mannheim, Germany) prepared in 0.5 M acedic acid, pH 2.8, and blocked with 2 % BSA. The chamber was mounted on a Zeiss Axiovert 100M/LSM 510 invert laser scanning confocal microscope (Carl Zeiss, Oberkochem, Germany). Upon perfusion, a series of stacks, i.e. 30 confocal optical sections, from the bottom to the apex of the forming platelet aggregate or thrombus, were obtained every 25 sec with a 488-nm laser and a scanning time of < 500 msec on an area of 26,450 μm2. Images corresponding to an area of 0.202 μm2 were analyzed by a “voxel”-based procedure, whereby a voxel is defined by a volume of 0.202 μm3 (0.45 μm x 0.45 μm x 1 μm). For calibration, fluorescent beads (Invitrogen, Carlsbad, CA, USA) were used, and the volume coresponding to a 1.0 μm thick stack was calculated pursuant to the voxel technique. A threshold was applied to distinguish adherent platelets from the background. Using these procedures, a uniform profile of thrombus formation and volume was observed (n=7). With citrate anticoagulated blood at an initial shear rate of 500 sec−1, thrombus growth begun after a lag phase of 220 sec, and, after 420 sec, thrombus volume reached a maximum (mean ± SD, 5x104 ± 4.9x103 μm3). Thrombus progression occurred in a two-step way with an apical growth (height extension) at the interval of 220 and 300 sec, and a further growth in the plane section at the interval of 300 and 420 sec after perfusion. Prolonged perfusion resulted in markedly abnormal flow pattern due to thrombus growth and increased shear rates. Again at an initial shear rate of 500 sec−1, platelet aggregate formation and thrombus progression were completely suppressed in the presence of anti-αIIbβ3 antibody (abciximab, 4 μg/ml). Interestingly, the polymorphism (HPA-1, PlA) of the β subunit of αIIbβ3 had a dramatic effect on thrombus growth. Thus, when comparing blood from homozygous carriers of HPA-1b (n=8) and HPA-1a (n=8), thrombus formation and progression occurred more rapidly with HPA-1b than with HPA-1a platelets, resulting in significantly larger thrombi from HPA-1b than from HPA-1a individuals (p=0.001). In conclusion, the voxel-based determination of thrombus formation and progression in vitro provides an appropriate technique to assess volumina of thrombi. Moreover, this technique can detect phenotypic differences related to an αIIbβ3 polymorphism which is postulated to modulate platelet thrombogenicity.