Combination chemotherapy is able to induce an objective response in 60-80% and a pathological complete response in 25-30% of patients with advanced ovarian cancer [1]. The achievement of a pathological complete response has a favourable prognostic relevance, however about 50% of complete responders develop recurrent disease after a median time of 14-32 months from second-look [2,3]. Obviously, relapse of disease after a complete response is due to the regrowth of persistent tumour cells; in this setting, the prolongation of chemotherapy beyond six cycles might be of benefit for patients with advanced epithelial ovarian cancer. In Italy a national cooperative study has been recently launched with the aim to assess the potential role of maintenance chemotherapy with paclitaxel (Tax) in patients in complete or partial response after first-line platinum-Tax-based chemotherapy. In detail, complete responders will be randomised to receive Tax 175 mg/m every 21 days for six cycles versus no further treatment. Partial responders with microscopic residual disease will receive Tax 60 mg/m weekly for 21 week. Partial responders with macroscopic residual disease will be randomised to receive Tax 60 mg/m weekly until progression or toxicity versus other second-line chemotherapy (such as topotecan at standard dose) until progression or toxicity. Overall more than 80% of patients with advanced ovarian cancer developed recurrent disease after initial treatment [4]. The probability of response to second-line chemotherapy following platinum-based treatment has usually been related to the platinumfree interval, although other variables, such as tumour burden and histology, have been found to be predictors of response [5,6]. Patients can be classified as platinum resistant (progression under platinumbased therapy or relapse within six months), and platinum sensitive (relapse after six months) [5]. Platinum-based salvage therapy can be proposed for sensitive patients, whereas patients with resistant disease have an extremely unfavourable prognosis and should be encouraged to enter experimental clinical trials [7]. Several drugs have been employed in second-line therapy, such as paclitaxel, topotecan and anthracyclines, with response rates ranging from 14 to 34% [8-16]; however, the responses are generally short-lived and survival curves remain poor. The role of surgery in the management of recurrent ovarian cancer is still uncertain. The authors agree that the longer the disease-free interval after primary therapy the better is survival after secondary cytoreduction [17-19]. In the series of Eisenkop et al. [18], median survival following secondary debulking was 19, 17 and 43 months for disease-free intervals of 7-11, 12-36 and >36 months (p = 0.01), respectively. Debulking surgery gives no or little benefit to patients with platinum-resistant disease [20,21]. The clinical value of surgical cytoreduction at the time of recurrence has been debated [17-20,22-24]. In the series of Morris et al. [23], including 30 patients who developed recurrent disease after a period of clinical remission of six months at least, surgical debulking to less than 2 cm did not improve patient outcome significantly. Conversely, analysing 100 patients with recurrent or progressive disease after cisplatin based-chemotherapy, Segna et al. [19] observed that median survival after secondary surgery was 27.1 months for the 61 patients who were left with residual tumour less than 2 cm compared to 9 months for the 39 patients with greater residuum (p = 0.0001). Similarly, in a series of 38 patients who underwent laparotomy for recurrence, Vaccarello et al. [24] found that median survival was 23 months for the patients with residual disease greater than 0.5 cm, whereas patients with smaller residuum had a 75% probability of surviving 41 months. Other authors detected a survival advantage only for patients cytoreduced to no macroscopic residuum [17,18].