Prostatic Intraepithelial Neoplasia (PIN) is widely considered to be a precursor lesion for adenocarcinoma of the prostate. No information is available, however, on the sensitivity of PIN to irradiation or the distribution of residual PIN after radiotherapy. We studied a series of forty-six totally embedded, whole mounted, serially sectioned prostates removed by salvage radical retropubic prostatectomy following irradiation failure in which no hormonal therapy/ablation had been undertaken. The mean age of patients was 65 (56-74) years, the mean dose of radiotherapy was 7,266 (6,000-9,000) cGy (15 external beam, 27 external beam plus iridium or gold seed, 3 iodine, and 1 unknown) and the mean interval for irradiation therapy to prostatectomy was 60 (16-145) months. Thirty-two (70%) of the patients had high grade PIN within the prostatectomy specimen. The pattern of PIN was recorded as described by Bostwick and co workers. The frequency of the different patterns per positive case paralleled in rank those in Bostwick's series of non-irradiated prostates, with the most common to least common per patient being: tufting (78.1%), micropapillary (59.3%), cribriform (34.4%) and flat (15.6%). However, the mean number of foci of PIN per prostate was less than in Bostwick's series (7.1 foci vs. 17 foci). There was no statistically significant difference between groups with or without high grade PIN with regard to various clinical factors (last preoperative serum PSA, age, dose of radiotherapy, interval from irradiation therapy to prostatectomy, Kaplan-Meier survival), or pathologic factors (presence of confined tumor, extracapsular extension, positive surgical margins, seminal vesicle invasion, or positive lymph nodes on permanent sections). We conclude that high grade PIN is common in the prostates of patients who have failed irradiation therapy and that, theoretically, not all recurrent tumors derive from regrowth of the initial, incompletely eradicated tumor. However, because there was no significant difference between the two groups with regard to the clinical and pathologic parameters listed, we consider it likely that most recurrent tumors derive from the initial incompletely eradicated tumor.