Risk of neoplastic transformation from cellular DNA: calculations using the oncogene model.

Based on a number of assumptions about oncogene size, frequency, biological integrity, and in vitro as well as in vivo transformation efficiency, estimates are made of the risk that the residual cellular DNA (rcDNA) contaminant in a biological product will cause a neoplastic transformation event. Using a statistical Poisson distribution approach, the probability of such an event is calculated to be at most 10(-6) assuming optimal in vitro conditions with 100 oncogene copies per cell and a 10 pg contaminant. More realistic assumptions using in vivo data suggest that the probability of a transformation event is at most 10(-9) assuming 100 oncogene copies per cell and a contaminant of 1 ng. Imperfections of the model and specific considerations of the human in vivo case are discussed.