Defects in pipelines may corrode and grow larger over time. Eventually the size of such defects may threaten the integrity of pipelines and result in expensive repairs replacements or even abandonment. Predicting the growth of defects allows operators to assess how long pipelines may be safely operated and to actively manage integrity. Current assessment methodologies commonly apply a single linear corrosion growth rate to all defects in a pipeline in their depth-based dimension only. The acceptability of defects at a given time interval is assessed and used to prioritise mitigations such as in-line inspection frequencies or repair schedules. The method presented in this paper describes an alternative way of predicting the growth of corrosion defects in three dimensions by the application of statistical modelling. This method may be used to predict the increase in axial size circumferential size and depth of a defect for any given time interval; this enables the acceptability of groups of defects that are predicted to interact to be assessed.