Summary.-The paper presents a detailed comparison of the anatomical distribution and frequency of clinically evident metastases in 152 cases of osteosarcoma, and autopsy findings in 43 cases. The behaviour of long bone tumours is contrasted with those arising elsewhere, which tend to metastasize less widely because of early death from effects of the primary tumour. In both clinical and autopsy series long bone tumours produced lung metastases (LM) in over 90% of patients dying with metastases, but the terminal frequency of extra-pulmonary metastases (EPM) rises from a clinical level of 33% to 83% at autopsy.There was little difference between tumours of the major long bones in the frequency of either LM or EPM, but EPM from the humerus tended to be fewer and sited above the diaphragm and from the femur below it. EPM most often involved other bones, notably vertebrae and pelvis. Not more than 10% of tumours invaded regional lymph nodes but terminally a quarter of the long bone tumours had metastasized to heart and abdomen. The infrequency of metastases in muscle was confirmed.The median time for LM was 5-6 months after starting treatment, for EPM 9-10. months. First metastases after 24 months were infrequent, especially in children. With delay in the appearance of metastases, whether LM or EPM, post-metastatic survival lengthened. Neither age, sex nor mode of treatment of the primary notably affected metastatic frequency, although recurrences were much more numerous when radiotherapy, even with high dosage, was the definitive treatment. Local recurrence usually appeared within 6-8 months and was shown to lead to increased frequency of osseous metastases. It is suggested that terminal dissemination may often be tertiary but not always from a pulmonary secondary. THE PAST 30 years' experience of the treatment of osteosarcoma shows universally bad results for any large series of cases. This is due both to the intrinsic limitations of current therapeutic methods and to concentration upon control or destruction of the primary tumour only, although it becomes increasingly obvious that the overwhelming lethal factor is distant metastatic growth, particularly in the lungs. With tumours of long bones, which comprise the great majority of osteosarcomata presenting in young people, lung secondaries are almost invariably the cause of death, even though metastases in other sites are not infrequent and are also potentially lethal.The explosion in diversity and scope of chemotherapy in the last decade is already producing better prognoses for some solid tumours. It is hoped that the use of cytotoxic drugs in the treatment of osteosarcoma may prolong the disease-free survival in this tumour also. Planned treatment along these lines requires a precise knowledge of tumour behaviour, especially of the dominating metastatic activity. We present an analytical study of the metastatic patterns of osteosarcoma arising in otherwise normal bones, as found both clinically

No abstract

A study is presented of all cases of primary sarcoma of bone registered during the period 1946 to 1974 for a specified population resident in south-western England. Ninety-six per cent of the 365 cases were histologically and radiologically verified and are separated into 8 categories of sarcoma. The number of tumours presenting during each hemi-decade did not markedly diverge from the 5-year mean for the period, nor was any significant change found in tumour incidence during the last 20 years of the survey. The age, sex and site distributions correspond with those reported elsewhere. Age-specific incidence rates are compared with those published for Sweden. For osteosarcoma and Ewing's tumour, both commoner in young people, the two series agree closely up to age 55 years, after which the Swedish incidence rates rise and are not exceeded when, for the present cases, Paget's osteosarcomas are included. Whilst Paget's disease may change the age incidence of some types of bone sarcoma, it is uncertain whether it increases the total number which occur. Differences in tumour incidence between males and females, whether for a specific type or for all bone sarcomas, are seldom statistically significant, but the patterns appear to be consistent. Images Fig. 2

Summary.-A study is presented of the rate of metastatic spread of osteosarcoma. The series consists of 123 tumours in long bones and 26 elsewhere in the skeleton. All tumours occurred in otherwise normal bones and were histologically proven. With a few stated exceptions all the cases were consecutively registered.Both the mean disease-free interval from the time of starting treatment and the crude survival curves are given. The long bone cases are analysed by groups according to the method of treatment, the patient's sex, age and tumour site. There were too few tumours of all other sites to warrant this discriminative treatment. Whilst the results of surgical treatment are better than for radiotherapy or a combined technique, the differences are not statistically significant and the information is recorded primarily to assist the evaluation of new forms of treatment of occult and overt metastases. Some problems in connection with such clinical trials are discussed briefly.

1. Alkaline and acid phosphatase, non-specific esterase and beta-glucuronidase have been estimated and demonstrated histochemically in a series of bone tumours and allied lesions, of which ten were osteogenic sarcomata, ten were giant-cell lesions, eleven were fibroblastic lesions and seven were tumours of cartilage. 2. Osteogenic sarcoma was found to be characterised by high levels of alkaline phosphatase, with rich staining for this enzyme in the tumour cells. Similar high levels of alkaline phosphatase were found in other bone-forming lesions, such as fibrous dysplasia, a giant-cell sarcoma with osteogenic matrix, and fracture callus. 3. Giant-cell lesions were characterised by high levels of acid phosphatase, and intense staining for this enzyme in the osteoclasts. These cells were also found to be rich in non-specific esterase (as shown by the alpha-naphthyl acetate method) and in beta-glucuronidase, but almost or entirely lacking in alkaline phosphatase. High levels of alkaline phosphatase were not found in giant-cell lesions except in relation to osteogenic matrix. 4. Fibroblastic tumours were characterised by moderate levels of all four enzymes, with little or no staining for phosphatases in the tumour cells; non-specific esterase was generally present in a proportion of the cells. 5. In certain lesions intermediate stages in the differentiation of fibroblasts to osteoblasts were found, notably in fibrous dysplasia, in which the biochemical change preceded the histological. In such lesions high total levels of alkaline phosphatase were found. 6. Cartilaginous tumours were characterised by low levels of all four enzymes, and little histochemical staining except in hypertrophied cells in areas of ossification. 7. It was found in general that the enzyme distributions in these neoplasms and other lesions reflected the findings in comparable reactive and growing normal tissues.