BACKGROUND: Among 40 patients with primary malignant tumors of the knee joint who underwent reconstruction of the affected limb with tumor prosthesis, revision was required in 7 due to stem breakage or loosening. SUBJECTS AND METHODS: In the 7 cases undergoing revision, conditions and background factors at the time of breakage, the breakage site, time of revision, models of previous and new prostheses, stem diameters before and after revision, details of the revision (blood loss, operative time), and the presence or absence of adjuvant therapy were determined. RESULTS: The replacement site was the distal femur in 5 and proximal tibia in 2. Revision was performed 6 years and 2 months after the previous prosthesis placement on average. The broken prosthesis model was KMFTR in 4 and HMRS and the physio-hinge type in one each. Revision due to loosening was performed in a case requiring replacement with Growing Kotz prosthesis. The model was switched to HMRS in 3, and the stem diameter was changed to 12 mm in 3 KMFTR breakage cases. The mean stem diameters were 11.2 and 10.2 mm in the non-revision and revision groups. The respective resection rates were 36 and 45%. The mean functional evaluation was 70.1% before and 76.2% after revision. CONCLUSION: To reduce the risk of tumor prosthesis breakage, the amount of bone resection should be limited to 30% or less in the affected bone, the stem diameter should be at least 12 mm, and the stem shape should be fitted to the anatomical shape of the femur.