The development of highly conformal dose distributions in the treatment of prostate cancer has heightened the deleterious influence of inter-fraction displacements of the prostate gland. Two alternative image-guidance technologies are being evaluated for their relative merits in compensating for inter-fraction displacements. A well-established on-line guidance program based upon localization portal images and implanted fiducial markers is being compared to a novel system that employs volumetric kilovoltage (kV) cone-beam computed tomography (CT) and a contour-based alignment system. A relative comparison of the limiting precision and accuracy; setup corrections; volumetric coverage; imaging dose; and, clinical workflow issues for these two approaches is reported. An investigation of the relative merits of the marker and cone-beam CT approaches in 30 patients undergoing conformal radiation therapy (42 fx) of the prostate has been initiated. The marker-based method employs 2 orthogonal portal images (Au 5mm x 0.8mm, 6MV, 4–8 MU, 10cm x 10cm) and manual alignment with plan-based DRR for estimation of the positioning error. All corrections are based on the marker-based estimate. Immediately prior to the MV localization and correction, a volumetric cone-beam CT set (120 kVp, 1280 mAs, 360 deg., 0.1 cm resolution, 40cm/25cm axial/longitudinal FOV) of the patient is acquired. All corrections in patient position are derived from the marker data and are applied by manual adjustment to the treatment couch. A patient-specific imaging technique is identified and image quality is assessed for detection of the prostate and surrounding structures. Retrospective analysis of the geometric displacement of the prostate gland using the cone-beam CT datasets is performed using a manual contour alignment tool. Preliminary investigations of the limiting accuracy and precision demonstrate that both methods can achieve sub-mm precision and accuracy for field placement with respect to high contrast test objects (markers or acrylic sphere in the case of cone-beam CT). The elapsed time for the two approaches in phantom studies indicate that both approaches can be performed in comparable time intervals. However, subtleties in the interpretation of the two different image modalities may influence the overall guidance time significantly (see Figure). The distribution of the imaging dose is considerably different in the two approaches: 4–8 cGy nominal for the marker-based approach, and 3.3 cGy peripheral/2.1 cGy central for the cone-beam CT approach. Overall, image quality is high with clear visualization of the bladder and rectum. Artifacts induced in the cone-beam CT images by the presence of the Au markers interfere with the visualization of the gland itself. Projection-based processing is being explored to correct for the presence of the markers and permit independent comparison of the two methodologies. A clinical process for cone-beam CT based imaging of the prostate has been developed and is being compared with the institutional standard of marker-based on-line correction. The two approaches have comparable limiting precision and accuracy. Image quality for the cone-beam CT approach shows excellent promise. The comparison of marker and cone-beam CT-derived corrections in a large patient population will permit the relative merits of these two approaches to be quantified and will direct suitable application of these two techniques.