The aim of this study was to compare the summing method (A) with the complement method (B) for calculating the cumulative lifetime-attributable-risk (LAR(tot)) of tumor incidence and mortality of multiple CT exposures. Method A defines LAR(tot) as the summation of the risk of each separate exposure. Method B was defined as the complement of the probability of inducing no cancer in N separate exposures. The risk of each separate exposure was estimated using dose, gender, and age at exposure (BEIR VII phase 2). Both methods were compared in a simulation and applied to a database of 11,884 patients exposed to multiple CTs. The relative difference between the methods was defined as ΔP%. Simulation confirmed that Method A always overestimates LAR(tot). ΔP% was proportional to the dose per exposure and the number of exposures. The differences between Methods A and B were small. Average LAR(tot) of tumor incidence was 0.140% (Method A) and 0.139% (Method B) with maxima of 5.70% and 5.56%, respectively. Average LAR(tot) of mortality was 0.085% for both methods, with maxima of 2.20% and 2.18%, respectively. ΔP% was highest (2.43%) for a female patient (3-y old) exposed to eight recurrent scans and a cumulative dose of 144 mSv. Although Method B is more accurate, both methods can be used to estimate the cumulative risk of multiple CT exposures. These results have to be interpreted, however, in the perspective of the uncertainties in the cancer risk model, which have been estimated at a factor of 2 or 3.