Chondrosarcoma, a malignant cartilage-forming mesenchymal tumor, displays a wide range of clinical behavior that can be difficult to predict with histological analysis. Matrix metalloproteinases contribute to the processes of local invasion and metastasis by controlling the ability of a tumor to transverse tissue boundaries. The specificity of matrix metalloproteinase-1 (interstitial collagenase) for fibrillar collagen may be central to those processes. Matrix metalloproteinase-2 facilitates invasion by degradation of such basement-membrane structures as type-IV collagen. The balance between the activity of tissue inhibitors of metalloproteinase and the activity of matrix metalloproteinase determines the proteolytic activity and may, in part, determine the overall invasiveness and potential for metastasis. The measurement of the ratio of matrix metalloproteinase to tissue inhibitor of metalloproteinase may have prognostic value for determining whether individual chondrosarcomas are locally invasive or will metastasize. Furthermore, there may be a specific pattern of expression of matrix metalloproteinase and tissue inhibitor of metalloproteinase in chondrosarcomas that is related to local invasion and probability of metastasis. Sixteen paraffin-embedded archival specimens of tumors were examined. Six twenty-micrometer-thick sections were cut from each tumor, and the amounts of cDNA formed from the mRNA were determined with reverse transcription-polymerase chain reaction with use of novel primers for matrix metalloproteinase-1, matrix metalloproteinase-2, tissue inhibitor of metalloproteinase-1, and tissue inhibitor of metalloproteinase-2. The amounts of cDNA for the matrix metalloproteinases and their inhibitors were determined by chemiluminescence and band densitometry. The ratio of the amount of cDNA for matrix metalloproteinase-1 to that for its tissue inhibitor and the ratio of the amount of cDNA for matrix metalloproteinase-2 to that for its tissue inhibitor were calculated, and the results were compared with use of the Student t test, enabling log-rank analysis of Kaplan-Meier survival curves. These ratios as well as the age and gender of the patient; the grade, size, and location of the tumor; the type of adjuvant therapy; and the operative margins were examined for significance with use of stepwise logistic-regression analysis. The patients who had recurrent disease had a significantly higher (p < 0.003) ratio of matrix metalloproteinase-1 to tissue inhibitor of metalloproteinase-1 (mean, 0.939; range, 0.647 to 1.101) than the patients who were free of disease (mean, 0.703; range, 0.629 to 0.772). Moreover, there was a striking difference between the Kaplan-Meier survival curve associated with a high ratio (more than 0.8) and that associated with a low ratio (p = 0.0015). The mean ratio of matrix metalloproteinase-2 to tissue inhibitor of metalloproteinase-2 was 1.814 (range, 1.206 to 3.77) in the patients who had recurrent disease compared with 1.473 (range, 1.073 to 2.390) in those who were free of disease; this difference was not found to be significant, with the numbers available. Analysis of the survival curves indicated that a worse prognosis was associated with a high ratio, but again this relationship was not found to be significant. Regression analysis revealed that a high ratio of matrix metalloproteinase-1 to its tissue inhibitor was a moderately significant independent predictor of a poor outcome (alpha = 0.07).