Editorial

Screening for cancer can be very effective in reducing incidence and mortality rates for some cancers and thus avoiding substantial suffering and distress for affected individuals. However, it can also be quiet futile if the wrong screening techniques are being used in the wrong target populations or can be even harmful, for example, if too many false-positive tests cause unnecessary feelings of insecurity and fear for patients tested in vain. False-negative test results can, moreover, lend individuals a false sense of security. As Anthony Miller points out in his mini review in this issue, the organization of well-designed screening programs is thus of pivotal importance to achieve beneficial results by cancer screening programs (pp. 1039–1046, this issue.).

Colorectal cancer is one of the major causes of death worldwide. Despite intensive basic research, which has provided a substantial amount of knowledge on molecular mechanisms that contribute to carcinogenesis and tumor progression, mortality rates are still high.1 As for many other cancers, early detection and timely surgical removal of precancerous or early invasive lesions is clearly the best strategy to prevent disease-related death. Hence, tremendous efforts have been undertaken to establish mass screening programs on a population-wide scale. However, these have relied, and still do rely, on screening assays that are far from being perfect.1 Procedures for detecting occult blood or specific blood components in fecal samples have been extensively used; however, the sensitivity and specificity profiles of these tests are not yet satisfactory.

More invasive techniques such as endoscopy are expensive, hampered by a substantial rate of severe side effects, and quite unpleasant for the patients. Hence, simple tests that indicate the presence or absence of colorectal lesions would have very high medical impact -- as long as they indeed can detect the respective lesions soon enough and with sufficiently high sensitivity and specificity profiles. It is commonly believed that proteins produced and released by tumor cells would represent good tumor markers. Similarly, molecular alterations of the DNA of tumor cells detected in DNA fragments released by disintegrating tumor cells could serve as potential biomarkers for neoplastic lesions (see Fig. 1). However, despite substantial efforts to identify such markers, no satisfactory clinical parameter could be identified yet.2, 3 This may be due, at least in part, to the fact that small, still resectable lesions may simply not produce and release enough marker molecules that can be detected in stool or plasma/serum samples.

Marshall et al. employed the conceptually novel approach of using subtle alterations of the transcriptome of peripheral blood-derived mononuclear cells (PBMNCs) as indicators of systemic conditions such as outgrowing cancers.3, 4 They isolated mRNA from PBMNCs of both cancer patients and normal healthy controls. By analyzing the gene expression patterns of both samples using mRNA array technology, they compared in training sets 112 patient-derived samples with 120 control samples and identified seven potential marker genes (ANXA3, CLEC4D, LMNB1, PRRG4, TNFAIP6, VNN1, IL2RB) that were differentially expressed in the respective PBMNC fractions. In a validation set of 202 patients versus 208 control samples, they confirmed their differential expression pattern. Although nothing is known about the underlying molecular mechanisms that trigger these different transcriptome patterns, the data described in this article suggest that the gene expression profile of PBMNCs may indeed be substantially influenced by the absence or presence of a neoplastic lesion. If these data can be further validated and if these signatures prove to be robust enough to rule out other conditions that may potentially impact these findings, this approach may be a substantial step ahead in cancer screening technologies.

If these data can be confirmed and the results eventually be validated in other cancers or even for other diseases, they clearly will open up a new avenue of clinical research using the tools of molecular and genomic medicine. Although the concept is fascinating, the evidence for its reliability is still rather vague. From the present data (pp. 1177–1186, this issue), for example, it is not clear why and in which cells transcriptome shifts occur. The number of patients investigated so far is small; the approach has not yet been reproduced by other independent groups. Other medical conditions may substantially influence the gene expression signature and thus interfere with the sensitivity and specificity profile of the assay. Alterations of changes in the PBMNC transcriptomes by inflammatory, septic, metabolic, genetic, iatrogenic, and many other conditions need to be carefully ruled out as a source of false-positive or false-negative results. Furthermore, the authors of the current paper all were part of a commercial organization producing and purchasing diagnostic products related to the content of the paper, which may represent a certain bias. Clinicians and scientist must remain skeptical and await confirmatory data, but the concept of using specific shifts of the transcriptomes of a distinct fraction of cells that can easily be collected as indirect biomarkers for the presence or absence of neoplasms4, 5 in the early detection and diagnosis of cancer and in surveillance programs is clearly challenging and worth exploring in substantially more depth and detail.

Magnus von Knebel Doeberitz

Associate Editor