BACKGROUND Osteochondral allograft (OCA) transplantation is an effective treatment for defects in the medial femoral condyle (MFC), but the procedure is limited by a shortage of grafts. Lateral femoral condyles (LFCs) differ in geometry from MFCs but may be a suitable graft source. The difference between articular surface locations of the knee can be evaluated with micro-computed tomography imaging and 3-dimensional image analysis. OBJECTIVE LFC OCAs inserted into MFC lesions can provide a cartilage surface match comparable with those provided by MFC allografts. METHODS Controlled laboratory study. METHODS Twenty MFCs and 10 LFCs were divided into 3 groups: 10 MFC recipients (MFCr), 10 MFC donors (MFCd), and 10 LFC donors (LFCd). A 20-mm defect was created in the weightbearing portion of the MFCr. Two grafts, 1 MFCd and 1 LFCd, were implanted sequentially into each MFCr. Micro-computed tomography (μCT) images of the MFCr were acquired and analyzed to compare the topography of the original recipient site with the MFCd- and LFCd-repaired sites. Three-dimensional transformations were defined to register the defect site in the 3 scans of each MFCr. Vertical deviations from each voxel of the graft cartilage surface, relative to the intact recipient cartilage surface, were calculated and assessed as root mean square deviation and percentage graft area that was proud, sunk, and within the "acceptable" distance (±1.00 mm). The effect of repair (with MFC vs with LFC) on each of the surface match parameters is presented as mean ± SD and was assessed by t test: height deviation over area (root mean square, mm), graft area acceptable (%), area unacceptably proud (%), area unacceptably sunk (%), step-off height over circumference (root mean square, mm), graft circumference acceptable (%), circumference unacceptably proud (%), and circumference unacceptably sunk (%). Percentage data were arcsin transformed before statistical testing. An alpha level of 0.05 was used to conclude if variations were statistically significant. RESULTS MFCr defects were filled with both orthotopic MFCd and nonorthotopic LFCd. Registered μCT images of the MFCr illustrate the cartilage surface contour in the sagittal and coronal planes, in the original intact condyle, as well as after OCA repairs. Specimen-specific surface color maps for the MFCr after implant of the MFCd and after implant of LFCd were generally similar, with some deviation near the edges. On average, the MFCr site exhibited a typical contour, and the MFCd and LFCd were slightly elevated. Both types of OCA-MFCd and LFCd-matched well, showing overall height deviations of 0.63 mm for area and 0.47 mm for step-off, with no significant difference between MFCd and LFCd (P = .92 and .57, respectively) and acceptable deviation based on area (87.6% overall) and step-off (96.7% overall), with no significant difference between MFCd and LFCd (P = .87 and .22, respectively). A small portion of the implant was proud (12.1% of area and 2.6% of circumference step-off height), with no significant difference between MFCd and LFCd (P = .26 and .27, respectively). A very small portion of the implant area and edge was sunk (0.3% of area and 0.6% of circumference), with no significant difference between MFCd and LFCd (P = .29 and .86, respectively). CONCLUSIONS The achievement of excellent OCA surface match with an MFCd or LFCd graft into the common MFCr site suggests that nonorthotopic LFC OCAs are acceptable graft options for MFC defects.