Forty-one ovarian cancer patients with less than 2 cm residual disease after systemic cisplatin-based chemotherapy received 4 courses of an ip regimen including cisplatin (75 mg/m2), mitoxantrone (20 mg/m2), and interferon-alpha 2b (30 mil IU/m2). The most important side effects were abdominal pain and fatigue. Overall 15/41 patients (37%) required narcotic analgesia for severe abdominal pain. In 1 case laparotomy was necessary due to bowel obstruction. Grade 3-4 myelotoxicity was observed in 18/41 patients (28 courses). No treatment-related death occurred. Pathological complete response (pCR) was achieved in 23/37 (62%) evaluable patients. Four-year disease-free survival was 50%, and no relapse occurred after 32 months. The estimated 4-year progression-free survival (PFS) and overall survival were 35 and 60%, respectively. Patients who achieved pCR showed significantly better survival than the others (P < 0.000). At multivariate Cox's analysis pCR achievement was the most important predictor of PFS (P < 0.005) and survival (P < 0.02). Age (< or = 60 vs > 60) and CA-125 serum levels at entry (normal vs increased) also showed independent predictive value. On the basis of multivariate analysis results we created a risk model for survival and PFS based on age and CA-125 at entry. We identified three subgroups of patients with significantly different outcomes. With this new ip combination long-term disease-free survival is achieved in a significant part of ovarian cancer patients with small tumor burden. A longer follow-up is needed to see whether it can cure some of these patients, and further comparisons with other ip or systemic regimens are needed to draw definitive conclusions about its role in these patients.