The axioms defining stochastic processes are generally simple. However, estimation of the parameters of a process from data is extremely difficult if customary techniques are used. This is due to the complexities involved in obtaining closed forms of likelihoods and evaluating them. The authors develop an estimation technique which selects those parameters which produce simulations that best mimic the data. SIMEST makes stochastic process modelling in oncology (and other fields) an attractive alternative to such currently popular alternatives as ad hoc regression models.