We evaluated the medium to long-term results of treatment with a custom prosthetic knee replacement after wide resection of a primary malignant tumor of the distal part of the femur in forty consecutive patients. The duration of follow-up ranged from five to seventeen years (median, eight years). At the time of the latest follow-up, thirty-five (88 per cent) of the forty patients were free of disease and five (13 per cent) were alive with metastatic disease. No local recurrence was observed. Twenty early complications occurred in eighteen patients (45 per cent). Aseptic loosening of the femoral component, which necessitated a revision in eleven patients at an average of fifty-one months, was the most frequent mode of failure. The rate of prosthetic survival, as estimated with use of the Kaplan-Meier method, was 85, 67, and 48 per cent at three, five, and ten years. Univariate analysis demonstrated that the rate of prosthetic survival was significantly worse for male patients, for those in whom at least 40 per cent of the femur had been resected, for those who had had total resection of the quadriceps muscles or subtotal resection (preservation of only the rectus femoris muscle), and for those in whom a straight femoral stem had been used (p < 0.05 for all comparisons). Multivariate analysis showed that the independent adverse prognostic factors for prosthetic survival were male gender, resection of at least 40 per cent of the femur, and fixation of the femoral stem with cement. The rate of limb salvage was calculated, with use of the Kaplan-Meier method, to be 93 per cent at three years and 90 per cent at five and ten years. At the latest follow-up examination, the functional scores according to the classification system of the Musculoskeletal Tumor Society ranged from 14 to 29 points; the mean was 24 points, which represents function that is 80 per cent that of normal. The mean scores in the categories of walking supports and gait were better for the patients in whom the quadriceps muscles had been preserved than for those who had had total or subtotal resection of those muscles. Although advances in imaging and local therapy narrow the indications for an extra-articular resection of a tumor, the implant that was used in the present study continues to be used in approximately 15 per cent of patients who have a fracture or an intra-articular extension of the tumor that necessitates extensive extra-articular resection.