To examine the incidence of seed migration detected on chest X-ray and to identify the predictors associated with its occurrence. Between May 1998 and April 2000, 102 patients underwent permanent prostate brachytherapy at our institution and 100 were eligible for the study. Chest X-rays obtained at follow-up were examined for the number and location of seeds. The patient and treatment variables potentially associated with the occurrence and number of seed migrations were analyzed. One or more seeds were identified on the chest X-rays of 55 (55%) of 100 patients. The mean number of intrathoracic seeds in patients with migration was 2.2 (range, 1-10), and the proportion of seeds that migrated to the thorax was 0.98%. The rate of extraprostatic seeds planned was 43.9%, and postimplant CT identified 37.9% in such a location. The number of seeds planned for extraprostatic placement and below the apex were statistically significant (alpha = 0.05) predictors in univariate logistic analysis. Multivariate analysis revealed the planned number of extraprostatic seeds as the only statistically significant predictor (p = 0.04). Extraprostatic placement of loose seeds is associated with an increased likelihood for, and frequency of, seed migration to the thorax. Nonetheless, the small proportion of implanted seeds that migrated (
