A dynamic model for the estimation of optimum timing of computed tomography scan for dose evaluation of I-125 or Pd-103 seed implant of prostate

Purpose/Objective: The dosimetric evaluation of permanent I-125 or Pd-103 p rostate implant is based on the assumption that both prostate and seeds are static throughout the entire treatment time which lasts months. However, t he prostate is often edematous after the surgical implantation of seeds. Th erefore, both the volume of the prostate and the seed locations change dyna mically as the edema resolves. This effect has impact on the validity of po stimplant analysis based upon a CT scan. If a CT scan is taken too early af ter implantation while there is edema in the prostate, the dose delivered b y the implant may be underestimated. If the imaging is delayed too long, th e dose may be overestimated. The magnitude of this effect depends on both o f the half-life of the isotope used and the half-life and magnitude of the edema. This study describes a dynamic biomathematical model which takes ede ma into account in calculating the dose delivered by the implant and is use d to investigate the optimum time to obtain the postimplant CT scan. Materials and Methods: The dynamic biomathematical model is a numerical int egration of the accumulated dose in which the prostate dimensions, the seed locations, and the source strength are all functions of time. The function which describes the change in prostate dimensions and seed locations as a function of time was determined in a separate study by analysis of serial p ostimplant CT scans. Dose-volume histograms (DVH) of the prostate for the t otal dose generated by the dynamic model are compared to DVHs generated by CT scans simulated for postimplant intervals ranging from 0 to 300 days aft er the implantation for 30 different combinations of the magnitude and dura tion of edema. Results: DVHs of the prostate calculated by taking edema into account show that the time of obtaining a CT scan for postimplant analysis is critical t o the accuracy of dose evaluations. The comparison of the DVHs generated by the dynamic model to those generated by the CT scans simulated for a range of postimplant intervals show that obtaining the CT scan too early tends t o underestimate the total dose while obtaining the CT scan after the edema is resolved tends to overestimate it. The results show that the optimum tim ing of the CT scan depends upon the duration of the edema and the half-life of the radiorsotope used. It is almost independent of the magnitude of the edema. Thus, a unique optimum time window for the imaging study cannot be defined for either I-125 or 103Pd implants. However, an optimum time window can be identified for which the calculated dose, on the average, will gene rally differ from the actual dose by less than 5%, with a maximum error not exceeding 15%. Such a window is 4 to 10 weeks after the implantation for a n I-125 implant, and 2 to 4 weeks for a Pd-103 implant. Conclusions: A dynamic biomathematical model to correct for the effects of edema in calculating the total dose delivered by an I-125 or Pd-103 Seed im plant has been developed. The model has been used to investigate the optimu m time window during which the postimplant CT scans for analysis should be obtained. (C) 1999 Elsevier Science Inc.