Thank you very much for your letter of 15 April 1993, and for the opportunity to comment on the correspondence from Dr. Pratt and his colleagues.
We would like to submit a reply as follows: We are grateful to Dr. Pratt and his colleagues for their comments. We applied the classical definition of multifocal synchronous osteosarcoma in our study, namely that multiple bony lesions should be present at diagnosis in the absence of pulmonary metastasis. This accounts for the small number of these tumours in our Registry as in other large series. If less stringent criteria are applied and patients with pulmonary metastases are included then the incidence of multifocal tumours will inevitably rise and a significant proportion will therefore have pulmonary metastases.
Of the 14 patients described by Dr. Pratt, 5 had pulmonary metastases on plain films at diagnosis. Applying our criteria rigidly, we would exclude these patients from the category of multifocal synchronous osteosarcoma. It is the 2 patients with negative chest radiographs and positive CT scans who would give rise to difficulty since they fulfill the criteria as originally proposed. It seems to us that if patients with pulmonary metastases are included in this category, then there is a risk that the margins of the entity will be blurred. We feel that the "multicentric" and "metastatic" hypotheses may be better tested by genetic studies on those cases without pulmonary metastases and would therefore urge that meantime strict criteria be applied. It may be that in the future these multiple tumours are shown to be of monoclonal origin and the metastatic theory proved correct. In these circumstances we would agree that the definition be expanded to include cases with pulmonary metastases at presentation.