Results of recent studies have greatly changed our understanding of ovarian carcinogenesis. The time-honoured notion that ovarian cancer originates from epithelium on the surface of the ovaries that invaginates into the underlying stroma, undergoes malignant transformation, and then spreads from the ovary to distant pelvic and intra-abdominal sites has been abandoned. Instead, a dualistic model was proposed for a group of rather indolent and genetically stable tumours (type I) and a large group of highly aggressive and genetically unstable tumours (type II). Type I tumours often are diagnosed in early stage, confined to the ovary, and develop from precursors such as borderline ovarian tumours, whereas type II cancers present at an advanced stage and originate from a putative precursor lesion in the fallopian tube. On the molecular level, type II tumours are characterised by a high rate of TP53 mutations and often have deficient homologous recombination and repair of double-strand DNA breaks. This deficiency has led to promising new treatment approaches with PARP inhibitors, both as a single agent and in combination with cytotoxic drugs. So far, analogue-targeted approaches for type I ovarian carcinomas have not been established, although this subgroup is also characterised by distinct molecular alterations. In many of them, there is a constitutive activation of the MAP-kinase signalling pathway because of mutations in ERBB2, KRAS, or BRAF.