There is still considerable debate about the optimal diagnostic imaging modality for acute pulmonary embolism. If imaging is deemed necessary from an initial clinical evaluation such as d-dimer testing, options include nuclear medicine scanning, catheter pulmonary angiography, and spiral CT. In many institutions, spiral CT is becoming established as the first-line imaging test in daily clinical practice. With spiral CT, thrombus is directly visualized, and both mediastinal and parenchymal structures are evaluated, which may provide important alternative or additional diagnoses. However, limitations for the accurate diagnosis of small peripheral emboli, with a reported miss rate of up to 30% with single-slice spiral CT so far, have prevented the unanimous embrace of spiral CT as the new standard of reference for imaging pulmonary embolism. The clinical significance of the detection and treatment of isolated peripheral pulmonary emboli is uncertain. Evidence is accumulating that it is safe practice to withhold anticoagulation in patients with suspected pulmonary embolism on the basis of a negative spiral CT study. Remaining concerns about the accuracy of spiral CT for pulmonary embolism detection may be overcome by the introduction of multidetector-row spiral CT. This widely available technology has improved visualization of peripheral pulmonary arteries and detection of small emboli. The most recent generation of multidetector-row spiral CT scanners appears to outperform competing imaging modalities for the accurate detection of central and peripheral pulmonary embolism. In this review, we assess the current role and future potential of CT in the diagnostic algorithm of acute pulmonary embolism.