Senescence was originally identified by the finite lifespan of normal cells that is a consequence of telomere shortening with each cycle of DNA replication. Cells undergoing replicative senescence display pronounced morphological and biochemical changes such as flattening and/or enlargement, increases in p21(WAF1) and/or p16(INK4A), a senescence-associated secretory phenotype, and often senescence-associated heterochromatic foci. Senescence also occurs in tumor cells in response to various forms of chemotherapy or radiation (therapy-induced senescence), which could be the basis for prolonged or (ideally) permanent growth arrest. Alternatively, therapy-induced senescence could represent a means whereby tumor cells evade the potential toxicity of chemotherapy and radiation, allowing for the eventual re-emergence or escape from senescence that could lead to disease recurrence. This review discusses the experimental data in the literature that support the premise that senescence is potentially reversible through the inactivation of p53, p16(INK4A) and/or Rb, over-expression of Cdc2/cdk1 and survivin, the development of polyploidy, the survival of cancer stem cells and/or restoration of the nuclear landscape. If senescence is truly reversible, then the re-emergence of tumor cells from senescent arrest induced by chemotherapy or radiation could represent a barrier to the development of effective and curative cancer therapies.