TO THE EDITOR:
We read with excitement "Association between Ratio of Matrix Metalloproteinase-1 to Tissue Inhibitor of Metalloproteinase-1 and Local Recurrence, Metastasis, and Survival in Human Chondrosarcoma" (80-A: 11—17, Jan. 1998), by Berend et al. We believe that this biological assessment was an important step in understanding the pathogenesis of local invasiveness and metastasis in human chondrosarcoma. However, the clinical data, especially those in Table I, concern us.
The authors did not differentiate between local recurrence and metastasis in identifying the patients who had a relapse. This distinction is important because the biology of local recurrence most likely is different from that of metastasis. Although the surgical margins were normal, local recurrence is often an operative or anatomical problem whereas metastasis probably reflects a different cellular phenomenon. The authors identified only one patient who had a local recurrence. We would like to know whether any of the nine patients who died of disease also had a local recurrence. If the answer is no, this paper probably could have focused solely on metastasis.
With regard to Table I, we believe that there were important errors in staging. Of particular concern are the columns regarding the duration of disease-free survival and the histological grade of the tumor, especially as they relate to the stage of the tumor. Two patients (Cases 3 and 8) had only sixteen and twenty-seven days of disease-free survival; these are very short durations. We would like to know if computerized tomography scans were made at two to four-week intervals to monitor the development of metastases. Second, we would like to know if the authors selected only tumors with soft-tissue extension. Many low-grade chondrosarcomas are intra-compartmental, either intraosseous or primarily extraosseous, because they are secondary to exostoses. Third, we are especially concerned with the case of one patient (Case 1) who allegedly had a grade-1 tumor that was stage IIIB and yet was reported to have 368 days of disease-free survival. We would like to know whether this patient had a thoracotomy and, if so, whether this was the reason he was considered disease-free. Two patients (Cases 12 and 16) had a grade-1 tumor that was stage IIB. The corresponding stage of these low-grade tumors should be IB.
Our last comment may be beyond the scope of this study, but we encourage the authors to investigate benign cartilage tumors with regard to their relative ratio of metalloproteinase-1 to tissue inhibitor of metalloproteinase-1; this information would finally give us a cellular marker with which to differentiate benign cartilage tumors from malignant cartilage tumors (that is, those capable of metastasis). It may lead to a better definition of chondrosarcoma.
We ask the authors to clarify the data on the basis of these comments in order for us to better understand this potentially important work.
Michael A. Simon, M.D.; Terrance D. Peabody, M.D.; Patrick J. Getty, M.D.: Section of Orthopaedic Surgery, The University of Chicago Musculoskeletal Oncology Group, 5841 South Maryland Avenue/MC 3097, Chicago, Illinois 60637
Dr. Berend, Dr. Toth, Dr. Harrelson, Dr. Layfield, Dr. Hey, and Dr. Scully reply:
Simon et al. pointed out inconsistencies in Table I. In response, we compared the existing computer database to the stages reported in Table I and confirmed that there is a discrepancy. It is not clear at what stage between the extraction from the database and the editing that the reformatting occurred. Certainly, it was perpetuated through the final publication, for which we apologize. We point out that, while the stage of each tumor is given in Table I, our results were based on grade and not on stage. We went back to the original patient files and ascertained that all of the grades reported in Table I are accurate. We have included an updated Table I that includes the corrected stages (Table I).
Simon et al. also asked about the difference between local recurrence and metastasis. Specifically, they asked if any of the nine patients who died of disease had a local recurrence. It is our contention that metalloproteinase facilitates the egress of cells from the cartilaginous tumor matrix. In this manner, it serves as one of the early components not only in the metastatic process but also in the establishment of satellite lesions within the reactive zone in the primary tumor. In the former case cell egress and access to the hematological compartment lead to metastasis, and in the latter case cell egress leads to the establishment of satellite lesions and the potential for local recurrance. We contend that the underlying mechanism is common to both processes.
Lastly, Simon et al. encouraged the comparison of metalloproteinase gene expression of malignant cartilage tumors with that of benign cartilage tumors. While we appreciate the importance of being able to distinguish between benign and malignant cartilaginous tumors, such differentiation is difficult histologically and was not the point of the current study. The purpose of the current study was to determine whether we can predict which patients who have a malignant chondrosarcoma are likely to have a recurrence. This information will allow us to focus novel therapies on these patients and to determine if interruptions of the early steps of the metastatic cascade are appropriate as a point of therapeutic intervention.
We are grateful to Simon et al. for pointing out the inconsistencies in Table I.
Keith R. Berend, M.D.; Alison P. Toth, M.D.; John M. Harrelson, M.D.; Lester J. Layfield, M.D.; Lloyd A. Hey, M.D.; Sean P. Scully, M.D., Ph.D.: Divisions of Orthopaedic Surgery (K. R. B. A.P.T., J. M. H., and S. P. S.) and Diagnostic Pathology (L. J. L.), and Center for Clinical Effectiveness (L. A. H.), Duke University Medical Center, Durham, North Carolina 27710