Editorial: Special section on "site-specific metastasis".

It seems evident that truly major strides in preventing cancer deaths will come only from the universal application of strategies, which can limit the initial occurrence of neoplasia or reduce the actual development of cancers. The difficulty of implementing such a strategy, however, is highlighted when one considers how successful our efforts are in limiting the devastating impact of smoking and other tobaccorelated diseases. Despite the plethora of data available, which indicate unequivocally the major contributions these habits make to the development of a range of tumour types, we are faced, in 2011, with the unpalatable fact that around 20% of the World’s 6.6 billion people are smokers. In Europe, for example, an average of 30% of adults admit to being smokers or tobacco users. With such ‘‘intransigence’’, in the face of overwhelming evidence to indicate that pursuit of one’s habit may well lead to development of a debilitating disease, it is no surprise that many researchers and clinicians take the view that, being unable to wait for a Utopia where cancer may be seen as a preventable disease, we must do more now to limit the severe effects that inevitably will occur from the continuing incidence of a spectrum of malignancies. With metastatic spread probably accounting for around 90% of all cancer-associated deaths, it is little wonder that study of the biology of metastatic processes, formerly something of a ‘‘Cinderella’’ or neglected discipline, has undergone a massive burgeoning in the past few years as we attempt to ‘‘do something’’ about this devastating aspect of oncologic disease. As the various contributions to this special section indicate, we now know many of the right notes that make up the chilling symphony of malignant dissemination. Perhaps, though, it is a case of knowing ‘‘the right notes’’, but not necessarily knowing whether they are ‘‘in the right order’’. The four reviews, ‘‘Do all roads lead to Rome? Routes to metastatic development’’ by Sleeman et al.; ‘‘Leukocytes as paracrine regulators of metastasis and determinants of organspecific colonization’’ by Erez and Coussens; ‘‘The seed and soil hypothesis revisited—the role of tumour-stroma interactions in metastasis to different organs’’ by Langley and Fidler and ‘‘Steps in prostate cancer progression that lead to bone metastasis’’ by Jin et al., all serve to indicate the vast amounts we now know about the molecular players involved in the intricate interactions between disseminating neoplastic cells and normal stromal components. Paradoxically though, this sheer wealth of recent data serves to highlight the ‘‘elephant in the room’’ or obvious truth which often is being ignored; that is, as indeed all of our contributors note, much of these data are based on animal model systems whose fidelity to the process they are meant to mimic may be, at best, somewhat questionable. Equally, the very heterogeneity of the conditions lumped together under the broad umbrella of ‘‘Cancer’’ mitigates against the development of simple generalisations or unifying hypotheses. Thus, for example, in their elegant dissection of the role that immunocytes play in regulating tumour progression, Erez and Coussens tend to focus on the stimulatory or enhancing role that such cell populations have upon tumour progression. There now are many studies in humans which provide strong supportive evidence for the occurrence of such a phenomenon. The complete picture though is less than clear-cut when we look at various tumours. How to explain the observation that cutaneous squamous cell carcinomas in organ transplant recipients are more invasive in the immunosuppressed population than they are in normal cohorts? Is it that most animal studies have looked at visceral metastases as an end point, and tended to ignore the primary tumour, or is it that squamous cell carcinomas of the skin are different from, say, adenocarcinomas of the colon? What is one to make of the animal studies indicating that circulating tumour cells are able to return to the primary tumour? If this is occurring on any scale in human cancer, would not an inevitable sequela be that the morphological appearance of the cancer changes towards a less structured, higher grade appearance over extended periods of cancer development? How to explain then the observation in breast cancer, for example, that when a Grade I or Grade II cancer reappears, often decades later, it maintains its original characteristic appearance rather than progressing to a Grade III tumour? If metastatic dissemination to specific premetastatic niches is dependent upon modifying effects from the primary tumour, are such mechanisms as operative in hitherto unsuspected early dissemination from small, initial lesions or from primary tumours of unknown origin as they are during late stage disease where the masses may have progressed in a conventional linear fashion and had more time to ‘‘prepare’’ the niche? We still appear to lack the essential model systems with which to tease out the mechanisms behind such conundrums. ‘‘Do all roads lead to Rome?’’ is the modern interpretation of ‘‘Mille viae ducunt homines per secula Romam’’ (A thousand roads lead men forever to Rome). In its day, a reflection of the hegemony of Rome this saying, indicative of the radiating web of roads emanating from the Forum, now is taken to mean that different paths can take a person, or organisms, to the same goal. As a metaphor for tumour dissemination, the difficulty is that in the pathophysiology of cancer there is no single goal. Different cancers spread to different sites, apparently often using mechanisms which differ according to their original anatomical location and/or cellular lineage. Like E di to ri al