The purpose of this study was to analyze the in vivo dimensions of each tibial plateau for planning of unicompartmental knee arthroplasty (UKA), and to compare the morphometric data to the dimensions of nine current designs of UKA tibial components. Thirty-seven knees (31 females and 6 males) operated on with UKA were studied. All patients were examined postoperatively using computed tomography (CT). There were 18 lateral and 19 medial UKAs. On the CT scan, each operated tibial plateau was measured in the transverse plane at the resection level, just below the full polyethylene tibial component. We measured the length of the anteroposterior (AP) cut as well as the maximal mediolateral dimension of the resected plateau (perpendicular to the AP cut). We compared the measurements with nine current UKA systems: Accuris (Smith and Nephew), Advance (Wright Medical), HLS Uni Evolution (Tornier), Miller-Galante and "ZUK" (Zimmer), Oxford and Oxford alpha (Biomet), Preservation (DePuy) and Unix (Stryker). There was good correlation between patient height and mediolateral dimension (r=0.6), and between patient height and area of total tibial plateau (r=0.7). The anteroposterior dimension was greater for the medial plateau (mean 50.8 mm, SD 3.3) than for the lateral plateau (mean 47.2 mm, SD 3.3). This difference was statistically significant (P=0.0016). Some UKA implants are designed with an asymmetric femoral component, but none have an asymmetric tibial component. The present study suggests, however, that the shape of the medial tibial plateau differs from that of the lateral plateau. This difference can lead to mediolateral overhang for medial UKA, if the surgeon aims for optimal anteroposterior coverage.