Ten patients with multiple non-ossifying fibromata are reported. All had associated extraskeletal congenital anomalies such as cafe-au-lait spots, mental retardation, hypogonadism or cryptorchidism, ocular anomalies or cardiovascular malformations. The radiographic picture and the distribution of the skeletal lesions are characteristic and constant. There are lucent areas in the shaft with a sclerotic margin; these areas narrow the medullary canal or may completely fill it. It is suggested that these features characterise a new malformation syndrome, possibly allied to neurofibromatosis. After skeletal maturation is complete, the skeletal lesions may (like non-ossifying fibromata) regress and undergo spontaneous healing.

This review of 27 cases serves to emphasis that periosteal chondrosarcoma and periosteal osteosarcoma are two distinct entities. Clinically, periosteal chondrosarcoma is less painful than periosteal osteosarcoma and runs a slower course. Radiographically, periosteal chondrosarcoma tends to affect the metaphysis and contains granular or "popcorn" opacities; while periosteal osteosarcoma more often affects the mid-diaphysis and shows lytic lesions with some spicules of reactive bone perpendicular to the underlying cortex. Histologically, periosteal chondrosarcoma shows lobular well-differentiated cartilage with Grade I or II (rarely Grade III) malignancy; periosteal osteosarcoma has a chondroid matrix with some osteoid component and Grade II or III malignancy. The prognosis in periosteal chondrosarcoma is good; conservative surgery is usually effective and metastases are very uncommon. In periosteal osteosarcoma the prognosis is less satisfactory but is better than that of other osteosarcomata; wide surgical excision is, however, needed and the incidence of metastases is about 15 per cent.

Two hundred and forty-eight high-grade central osteosarcomata were treated by amputation or disarticulation; in 5.2 per cent the tumour recurred at the amputation site. The following causes may be responsible for local recurrence: the level of the amputation is too close to the tumour; there is an unrecognised intramedullary extension of the tumour; during a previous block resection tumour cells may have been seeded in the soft tissues; the primary tumour was too extensive even for radical surgery; "skip" metastases may have been present; iatrogenic tumour implantation may have occurred while a biopsy was being performed during the course of an amputation. Treatment of the primary osteosarcoma should take all these possibilities into account. In our experience adjuvant chemotherapy has not significantly changed the frequency of local recurrences which should be treated by radical operation or, if this is not possible, by irradiation; chemotherapy may be used as an adjuvant. The prognosis of local recurrences is bad.