Abd El-Rehim et al., pp. 340–350 Breast cancer, a complex disease, cannot be completely assessed by currently available clinical methods. Novel molecular markers have been identified that may better address the molecular and genetic events underlying the disease. A panel of these new biomarkers has now been examined in a comprehensive British study of a large series of invasive breast cancer. Abd El-Rehim and colleagues tested the expression profiles of 25 candidate proteins by immunohistochemistry in tissue microarray sections of nearly 2,000 women with breast cancer. Using hierarchical clustering and artificial neural networks, the authors identified 6 main clusters of biomarkers. Hormone receptors, the epidermal growth factor c-erbB-2, cytokeratins 5/6 and 18, mucin1, the tumor suppressors p53 and BRCA1 and the cell adhesion molecule E-cadherin were the driving key markers and the most important discriminators among different clusters. The authors characterized the cancer cases into 6 groups. Whereas groups 1 and 2 were strongly positive for cytokeratin 18 and hormone receptors, group 3 showed strong expression of c-erbB-2 and mucin1. Group 4 was dominated by strong expression of c-erbB-2, p53 and nuclear BRCA1, whereas group 5 was characterized by strong p53 expression and the absence of hormone receptors and group 6 by the abundance of c-erbB-2 and E-cadherin. The authors found marked differences between the 6 groups with respect to tumor grade, size and lymph node stage. Importantly, the molecular clustering was highly significantly associated with overall survival and disease-free intervals. Mortality was highest in patients whose tumors were classified as group 5, followed by patients in groups 3 and 6. These results suggest a more complex classification system is needed for clinical and prognostic management of breast cancer. 1 Kaplan-Meier analysis of overall survival in different clusters of invasive breast cancer. Hoglund et al., pp. 401–406 Systematic analysis of of karyotypic evolution in epithelial tumors has identified a number of conserved features, including: (1) acquisition of chromosomal imbalance in a preferred order; (2) the presence of distinct cytogenetic pathways characterized either by gain or loss; and (3) conformity of the number of acquired chromosomal changes to a power law distribution. Hoglund and colleagues have proposed a model of stepwise acquisition of aberrations that faithfully reproduces the power law distribution to explain the observed distribution of the number of aberrations per tumor. In this report, the same group has analyzed the accumulated cytogenetic data on karyotypic evolution in Ewing tumors and synovial sarcomas. Both tumors are frequently characterized by the presence of the balanced translocations t(11;22) and t(X;18). In Ewing tumors, the authors identified +8, +12, +1q and 16q− as important secondary changes to t(11;22). By principal component analysis, a major karyotypic pathway characterized by gains and a minor one characterized by losses were identified. In synovial sarcomas, the karyotypic evolution was less clear. A power law distribution of the number of acquired aberrations was identified for both tumors. Ewing tumors and synovial sarcomas differ from other malignancies with balanced translocations resulting in fusion genes. These observations are consistent with the model that chromosomal changes observed in these two tumors may develop through mechanisms similar to those observed in epithelial tumors, in contrast to hematological malignancies, which are characterized by balanced translocations with gene fusions. Becher et al., pp. 451–457 Laryngeal cancer (mostly squamous cell carcinoma) is the second most frequent cancer of the respiratory system and generally carries a favorable prognosis. The two major risk factors identified are cigarette smoking and high alcohol consumption, particularly in combination. Some occupational risk factors have also been linked to laryngeal cancer, including exposure to asbestos, mineral coal products, mineral oils, fossil fuels, ionizing radiation and others. However, no conclusive evidence has been provided so far. In this issue, Becher and colleagues have analyzed occupational risk in a cohort of 257 cases of laryngeal cancer in Germany. Job-specific questionnaires for selected jobs known to be associated with carcinogenic exposure were used to ascertain exposure to potential carcinogens in all members of the cohort. A strong effect of polycyclic aromatic hydrocarbons was noted for laryngeal cancer risk after adjusting for smoking and alcohol and a clear dose-response effect was found for exposure duration. These findings are further supported by the risk associated with occupations in which exposure to polycyclic aromatic hydrocarbons is likely, such as road and building construction, where workers are exposed to asphalt. These observations firmly establish polycyclic aromatic hydrocarbons as an independent risk factor for laryngeal carcinoma.