Purpose/Objective: The ability to deliver a radiation dose with high accuracy using external beam radiotherapy for the treatment of prostate carcinoma has enabled dose escalation, which in turn has impacted on the local control. Moreover, the exposed volume effect for late rectal complications has supported the reductions of CTV-PTV margin size. These margins however are limited by internal motion of the prostate. In this study, inter-fraction deformation of the rectum is modeled from data presented in the literature. The aim is to investigate the ramifications (in terms of rectal NTCP) from inter-fraction rectal deformation both for individual patients with set systematic errors and also across a patient population. The relative importance of margin size (i.e. internal margin + setup margin) against rectal deformation is also examined. Materials/Methods: A Monte Carlo model is constructed to simulate rectal deformation as inter-fraction changes of rectal wall radius. The parameters for mean rectal radius, inter-fraction and inter-patient standard deviations in rectal radius were taken from the literature (Stroom et al 1999). Geometrical information is taken from a local CT data set and normalized according to the published population mean. For each patient a systematic difference between the rectal size depicted in the planning CT and the treatment mean is presumed. The systematic errors range from an empty rectum at planning to a full rectum at planning. Inter-fraction deviations are then sampled for each patient of the population. Preplanned dose distributions with uniform margins of 0.5, 1.0 and 1.5 cm around the prostate clinical target volume were used to test the importance of dose variation against the margin size. Patient NTCP is calculated (using the Kallman s-model) with each modeled patients accumulated biological effective dose. Results: The dose delivered to the rectal wall depends on the status of the rectum (full/empty) in the preliminary planning CT scan. Deviations from the planned dose were larger if the rectum was empty in the planning CT scan (ΔD = ±25%) than if it was full (ΔD = ±15%). If the planning CT scan demonstrated the rectum in the mean treatment position, the intended planned dose will have been delivered. Residual dose variation in this case (ΔD = ±10%) is due to inter-fraction displacements only. Consequently, the NTCP will vary depending on the initial rectum status. Due to the symmetrical nature of the function used to model rectal deformation across a patient population, the mean NTCP did not vary greatly from that expected (planned position). The mean NTCP is governed by the margin size and average dose distribution. Inter-patient variation in radiosensitivity of the rectum (modeled with standard deviation 15% of mean rectal radiosensitivity) dominated the complication rate above the very large dose fluctuations and the margin size. Accordingly, despite the calculated population mean NTCP equalling the intended NTCP, inter-patient variation in NTCP is substantial. Conclusions: The study concludes that correct prescription of margin size is more important in reducing NTCP than daily fluctuations in the rectal wall. The study also supports protocols that seek to limit variability in rectal volume from planning to treatment delivery. Reference Stroom, J. C., Koper, P. C. M., Korevaaar, G.A., van Os, M., Janssen, M., de Boer, H. C. J., Levendag, P. C., Heijmen, B. J. M. Internal Organ Motion in Prostate Cancer Patients Treated in Prone and Supine Treatment Position. Radiother. Oncol. 51:237-248, 1999.