This study was designed to serially analyze the magnetic resonance (MR) appearance of patella tendon grafts during the first year of implantation, and to determine if the sagittal location of the tibial tunnel affects the MR appearance of the graft. An additional goal was to analyze the effect of the sagittal placement of the tibial tunnel placement on knee extension and stability. Two groups were defined by comparing the sagittal relationship of the tibial tunnel to the slope of the intercondylar roof from a lateral roentgenogram of the knee in full extension. The roof impinged group consisted of nine patients who had the tibial tunnel placed anterior to the tibial intersection of the slope of the intercondylar roof. The unimpinged group was composed of eleven patients who had the tibial tunnel placed posterior and parallel to the slope of the intercondylar roof. MR scans were obtained at 0-2, 12, 24, 36, and greater than 48 weeks postoperatively. The signal intensities of grafts in both groups were identical at 1 week, but increased in the impinged group by 3 months and remained increased at 12 months postoperatively (p less than 0.01). In contrast, the MR signal of grafts in the unimpinged group remained unchanged during the first year. Knees with impinged grafts had extension deficits but remained stable. Roof impingement was impossible to view directly with the knee in full extension.(ABSTRACT TRUNCATED AT 250 WORDS)