The treatment of prostate cancer by 125I interstitial implants has been extensively studied with mixed results by one institution or another. A recent study from Hahnemann [Int. J. Radiat. Oncol., Biol., Phys. 21,955-960 (1991)] reported results that were extremely poor compared to those reported in an earlier study at Yale [Int. J. Radiat. Oncol., Biol., Phys. 14, 1153-1157 (1988)] or those in an Eastern Virginia Study [Cancer 63, 2415-2420 (1989)]; differences in 5-yr survival rates being more than a factor of 2. Such large discrepancies from institution to institution led us to a reexamination of the dosimetry. This study analyzed quantitatively three-dimensional dosimetric parameters of 110 prostate cancer patients treated with 125I interstitial implants. The study searched for "cutoff" values in each parameter that divided the patients into two groups with statistically significant differences in the local recurrence-free survival rates. A comparison of the three-dimensional isodose surfaces of patients with favorable values in all of the parameters to those patients with all unfavorable parameters show how these characteristics translated into poor dose coverage and much inhomogeneity within the implant even for cases that met the traditional criteria for adequacy (160 Gy to the tumor volume). Patients in the favorable group had 10-yr survival rates higher by a factor of up to 2 compared to those in the unfavorable group. The strong correlation of three-dimensional volume-dose parameters to the local control rate observed in this study further emphasizes how important it is to assess the three-dimensional dosimetric adequacy of interstitial implants before deciding on their clinical efficacy. If implants are performed with appropriate attention to dosimetry parameters, excellent clinical results are obtained. On the other hand, if dosimetry parameters are not correct, the implant results can be poor.