To compare intensity-modulated radiotherapy (IMRT) with three-dimensional conformal radiotherapy (3D-CRT) in terms of carcinogenic risk for actual clinical scenarios.Clinically equivalent IMRT plans were generated for prostate, breast, and head-and-neck cases treated with 3D-CRT. Two possible dose–response models for radiocarcinogenesis were generated based on A-bomb survivor data corrected for fractionation. Dose–volume histogram analysis was used to determine dose and its distribution to nontargeted tissues within the planning CT scan volume and thermoluminescent dosimetry for the rest of the body. Carcinogenic estimates were calculated with and without a correction factor accounting for cancer patients' advanced age and reduced longevity.For the model assuming a plateau in risk above 2-Gy single-fraction-equivalent (SFE), IMRT and 3D-CRT produced risks of 1.7% and 2.1%, respectively, for prostate; 1.9% and 1.8%, respectively, for nasopharynx; 1% each for tonsil; and 1.4–2.2% and 1.5–1.6%, respectively, depending on technique, for breast. Assuming a reduction in risk above 2-Gy SFE, risks for IMRT and 3D-CRT were 1.1% and 1.5%, respectively, for prostate; 1.4% and 1.2%, respectively, for nasopharynx; 1% each for tonsil; and 1.3–1.8% vs. 1.3–1.6%, respectively, for breast. Applying a correction factor of 0.5 for cancer patients halved these risks and their relative differences.Carcinogenic risks were comparable in absolute terms between modalities. Risks are dependant on technique used. Risks with IMRT are influenced by monitor unit demand and are therefore software/hardware dependant. The dose–response model accounting for cell killing at higher doses fitted best with actual observed risks.