Malignant lumors are characterized by their ability to invade adjacent tissues and to metastasize. A few tumor types show invasive behaviour and usually fail to metastasize during their entire life histories; examples are chordoma, craniopharyngioma, glioma and basal cell carcinoma. Of these, basal cell carcinoma is by far the commonest, and the rare event of its metastasis has been recorded to date in almost 100 cases. It is important to consider whether analysis of these cases can provide any broad clues as to how this invasive tumor has, on occasion, gained metastatic potential. In recent years many studies on tumor cell mobility, enzyme production and cell surface characteristics^-^ have revealed the complexity of the metastatic process but so far no clear pattern has emerged from the detail. It is not the intention here to add to detailed knowledge of the metastatic process, nor to contribute new experimental data, rather to take a broad viewpoint in the hope of identifying a direction or area in which this basic research might profitably be applied. Metastasis of Basal Cell CarcinomaReview of 38 cases described in 24 papers^"^^ reveals the following information. The primary tumor is almost always situated on the face or scalp and the histological diagnosis is beyond reasonable doubt in all but a very few cases. In most accounts metastasis is not associated with any particular or unusual histological pattern of basal cell carcinoma, nor do the metastases differ his-tologically from the primary tumors. The primary tumor almost invariably extends over an unusually large area and deeply erodes underlying structures. There is a long clinical course including repeated recurrences and attempts at treatment, which has consisted of both surgery and radiotherapy. The age of the patient at onset of the tumor, when known, ranges from 17-68 years (mean 44 years). The interval between onset of the tumor and detection of metastasis ranges from 4-25 years (mean 11 years). The site of metastasis in most cases is local lymph nodes. Blood-borne spread is less common and tends to affect the lungs, liver and bone. Metastasis in the spleen, pancreas, dura mater, pericardium and kidney has also been recorded. Metastasis may be restricted to lymph nodes, lung, bone or liver, but multiple organs are often involved. (A capricious pattern of metastasis is also seen in other metastasizing tumors.) When blood-borne metastasis has occurred, the tumor behaves with a high degree of malignancy and prognosis is poor. DiscussionLarge tumor size is an outstanding feature of metastasizing basal cell carcinomas. This is consistent with the long period before metastasis and the multiple treatments and recurrences. The face is an area in which treatment might err on the conservative side; furthermore, the subcutaneous tissue here is thin and underlying muscle and bone are vulnerable to invasion.This relationship between tumor size and metastatic potential is seen in most tumors, giving rise to good prospects of cure if the tumor is excised before...

Immunofluorescence techniques failed to reveal evidence of anti-tumour antibody in the sera of patients with basal cell carcinima. Although the presence of such antibodies has previously been associated with the absence of metastasis in malignant melanoma, other explanations for the low metaststic potential of basal cell carcinoma should be sought.

An average human contains about 500g lymphocytes, equivalent to an organ one third the size of the liver, but far more likely to undergo neoplastic transformation. Lymphocytes are the most extensively studied and complex cell type in the body, and this is equally true for their tumours. Enormous advances have been achieved in the epidemiology, cytogenetics, and classification of lymphoid tumours. In these papers we shall see how remarkably close we are to understanding the cause of some types of lymphoid cancer. Part I will concentrate on epidemiology and the role of viruses and radiation. The contribution to lymphoid neoplasia of genetic mutations will be reviewed in Part II, and the new classification will be presented in a subsequent part.

In a sense we know the cause of cancer; it is genetic mutations. These operate by progressively damaging the mechanism for control of the cell mitotic cycle. For mutations to do this several condition must apply: • the mutations must somehow escape the cell's DNA repair mechanism • the mutant cell must be able to enter the mitotic cycle by evading normal mechanisms which prevent this • the mutant cell must survive the effects of the mutation and be able to pass on its defective DNA to succeeding generations of its progeny. When sufficient mutations have accumulated in its mitotic control genes, a cell can break free of mitotic control and become a cancer cell.