No definite objective guidelines for selecting patients for pelvic nodal or seminal-vesicle irradiation exist in radical radiotherapy for prostate cancer. Developing such guidelines will aid in decreasing unnecessary irradiation of normal tissues in patients in whom involvement of pelvic nodes and seminal vesicles is unlikely. The development of objective criteria will also be of help in successful planning of dose-escalation studies with 3D conformal radiotherapy. Some recent advances are useful in the development of such guidelines. Based on whole-mount, morphometric studies from radical prostatectomy specimens, a new understanding of the zonal anatomy of the prostate has emerged. The three glandular zones are the transitional, central, and peripheral zones; the fourth zone is a nonglandular, fibromuscular one situated anteriorly. The majority of stage A tumors arise in the transitional zone, whereas stage B tumors arise from the peripheral zone. A tumor arising from the transitional or nontransitional zone usually does not cross the boundary between zones unless the tumor volume exceeds about 4 cm3. The tumor volume influences the following: (a) formation of pelvic nodal me-tastases; (b) seminal-vesicle involvement; (c) capsular extension, and (d) the extent of the Gleason pattern 4-5 component. In general, tumors smaller than 4 cm3 tend not to involve regional nodes or seminal vesicles; they have a negligible Gleason pattern 4-5 component, and this is a favorable prognostic factor. Tumor volume estimates can be obtained from transrectal ultrasound. In addition, transrectal ultrasound has significant potential in the identification of patients with gross seminal-vesicle involvement. Transrectal ultrasound slightly underestimates tumor volumes, and about 20% of tumors may not be visible sonographically, but these shortcomings can be remedied by measurement of the prostate specific antigen level. The level of this antigen, a very useful tumor marker in prostate cancer, depends on the tumor cell burden. Studies show that nodal and seminal-vesicle involvement is unlikely if the prostate specific antigen level (determined by a monoclonal antibody method) is less than 10 ng/ml, and that it is highly likely with levels above 20 ng/ml. In patients with prostate specific antigen levels between 10 and 20 ng/ml, ultrasonographic findings as well as the stage and grade of the tumor can help the radiotherapist to determine the extent of the target volume. For treatment planning, decision trees are developed based on stage, grade, tumor volume determination, and prostate specific antigen levels. A significant difference in the volumes of irradiation to the rectum and bladder with or without pelvic nodal/seminal vesicle irradiation is illustrated with dose-volume histograms. The margin around the target in the definition of the planning target (i.e., the margin from prostate to block) due to day-to-day set-up variations and physiologic movement must be about 1-2 cm. Prospective use of the objective criteria developed here ca...