To assess the prognostic factors and the outcome in patients with squamous-cell carcinoma of the penis, a retrospective review of 41 consecutive patients with non-metastatic invasive carcinoma of the penis, treated between 1962 and 1994, was performed. The median age was 59 years (range: 35-76 years). According to the International Union Against Cancer (UICC) 1997 classification, there were 12 (29%) T1, 24 (59%) T2, 4 (10%) T3 and 1 TX (2%) tumours. The N-classification was distributed as follows: 29 (71%) patients with N0, 8 (20%) with N1, 3 (7%) with N2 and 1 (2%) with N3. Forty-four per cent (n=18) of the patients underwent surgery: partial penectomy with (n=4) or without (n=12) lymph node dissection, or total penectomy with (n=1) or without (n=1) lymph node dissection. 23 patients were treated with radiation therapy alone, and all but 4 of the patients who were operated upon received postoperative radiation therapy (n=14). The median follow-up period was 70 months (range 20-331 months). In a median period of 12 months (range 5-139 months), 63% (n=26) of the patients relapsed (local in 18, locoregional in 2, regional in 3 and distant in 3). Local failure (stump in the operated patients, and the tumour bed in those treated with primary radiation therapy) was observed in 4 out of 16 (25%) patients treated with partial penectomy +/-postoperative radiotherapy versus 14 out of 23 (61%) treated with primary radiotherapy (P=0.06). 15 (83%) out of 18 local failures were successfully salvaged with surgery. In all patients, 5- and 10-year survival rates were 57% (95% confidence interval (CI), 41-73%) and 38% (95% CI, 21-55%), respectively. The 5-year local and locoregional rates were 57% (95% CI, 41-73%) and 48% (95% CI, 32-64%), respectively. In patients treated with primary radiotherapy, 5- and 10-year probabilities of surviving with penis preservation were 36% (95% CI, 22-50%) and 18% (95% CI, 2-34%), respectively. In multivariate analyses, survival was significantly influenced by the N-classification, and surgery was the only independent factor predicting the locoregional control. We conclude that, in patients with squamous-cell carcinoma of the penis, local control is better in patients treated with surgery. However, there seems to be no difference in terms of survival between patients treated by surgery and those treated by primary radiotherapy +/-salvage surgery, with 39% having organ preservation.