Cancer cells introduced into developing embryos can be committed to a complete reversion of their malignant phenotype. It is unlikely that such effects could be ascribed to only few molecular components interacting according to a simple linear-dynamics model, and they claim against the somatic mutation theory of cancer. Some 50 years ago, Needham and Waddington speculated that cancer represents an escape from morphogenetic field like those which guide embryonic development. Indeed, disruption of the morphogenetic field of a tissue can promote the onset as well as the progression of cancer. On the other hand, placing tumor cells into a "normal" morphogenetic field - like that of an embryonic tissue - one can reverse malignant phenotype, "reprogramming" tumor into normal cells. According to the theoretical framework provided by the thermodynamics of dissipative systems, morphogenetic fields could be considered as distinct attractors, to which cell behaviors are converging. Cancer-attractors are likely positioned somewhat close to embryonic-attractors. Indeed, tumors share several morphological and ultra-structural features with embryonic cells. The recovering of an "embryonic-like" cell shape might enable the gene regulatory network to reactivate embryonic programs, and consequently to express antigenic and biochemical embryonic characters. This condition confers to cancer an unusual sensitivity to embryonic regulatory cues. Thus, it is not surprising that cancer cells exposed to specific embryonic morphogenetic fields undergoes significant modifications, eventually leading to a complete phenotypic reversion.