The results of our retrospective study do no support the common belief that PLD is the first choice for recurrent or progressive OC treatment. Indeed, in our analysis, we were unable to demonstrate advantages in terms of PFS and OS for PLD administration at any time along the course of disease. However, platinum-sensitive patients took benefit from PLD more than the platinum-resistant group.
In our patient series, PLD given at any point indeed translated in a trend to worse outcome in term of both PFS and OS if compared to patients that never received this agent. While clinical trials indicated that PLD is better tolerated as compared to other drugs, no difference in toxicity was observed in the current study. In our series, 16 platinum-sensitive patients were treated according to CP schedule  and 6 patients received PLD. The latter group however, experienced a worse outcome. The common practice of PLD usage in the platinum-sensitive setting has been defined on subgroup analysis of major randomized trials. Indeed, some reports showed a 2 months advantage for PLD compared to topotecan in terms of PFS in platinum-sensitive subgroup only . Subsequently, the non-inferiority phase III Calypso trial compared carboplatin-PLD (CD) vs carboplatin-paclitaxel (CP) in patients with platinum-sensitive recurrent OC. In this trial the primary end-point (PFS) was met and a superiority for CD regimen was suggested. However, this conclusion is not appropriate in the light of the non-inferiority trial design. Indeed, reported results are not sufficiently strong to support the PLD use as first choice in clinical practice. The cornerstone of platinum-sensitive OC treatment is currently represented by re-challenge with platinum which is clearly effective, with a response rate of about 50-60%. Meaningfully, the anti-VEGF monoclonal antibody bevacizumab is the only agent able to produce an improvement of 4 months in terms of PFS in combination to platinum-containing second line treatment . In the platinum-resistant setting topotecan or gemcitabine may represent an adequate option despite a low response rate and a moderate bone marrow toxicity. Several trials compared PLD to gemcitabine in platinum-refractory patients and/or partially platinum-sensitive patients without showing advantage for PLD, but only a trend in terms of PFS for the subgroup of patients with a platinum-free interval (PFI) between 7 and 12 months .
In our experience, 18 patients with a partially platinum-sensitive status undergone a second line therapy. Of these, 13 patients were treated with a PLD-based schedule (80% of them in combination with trabectedine, TB) whereas 5 patients received other agents. Some authors demonstrated that TB in addition to PLD produces a 3 months increase in PFS . However, in our series, the 5 patients in PLD-free group had a better outcome in terms of both PFS and OS (the small sample size may account for the lack of statistical significance).
Our study presents some limitations such as the small sample of patients, the heterogeneous therapeutic strategies, the long enrollment time, with the possible occurrence of selection bias. Moreover, our medical records do not include data about quality of life, an important factor in palliative setting. However, our findings are in agreement with a small retrospective study which included 43 patients treated with PLD and indicated only a marginal benefit in terms of survival for PLD in the platinum-resistant/refractory setting despite a considerable toxicity . Furthermore it is important to underline that none of major trials reported a real advantage for PLD if compared to any other drug in terms of OS. Surprisingly, a recent phase III trial which compared PLD to patupilone showed a limited improvement in terms of OS for PLD in platinum-refractory OC . These findings are difficult to interpretate taking into account the limited experience on patupilone. Interestingly, the AURELIA trial, designed for the platinum-refractory setting, investigated the effect of bevacizumab addition to single agent chemotherapy including PLD. In this trial the anti-VEGF monoclonal antibody produced almost a doubling (3.4 vs 6.7 months) in PFS but this advantage was indeed more relevant for weekly paclitaxel if compared to PLD and topotecan .
Our findings need to be considered in the general scenario of anthracycline-based treatment of OC. Recently, a report demonstrated an advantage in terms of PFS for PLD vs Olaparib in platinum-refractory patients with identified BRCA mutations, . Moreover, others studies reported a correlation between functional homologous recombination deficiency (HRD) and clinical benefit from antracycline . We did not find, however, any advantage from the administration of PLD in patients not selected for BRCA1 mutational status. During the last years, several trials debunked anthracycline role in the management of OC for a no clear advantage in efficacy at cost of an increased toxicity [38, 39].
In particular, different trials evaluated the actual cost/effectiveness of chemotherapy beyond first line treatment. The effectiveness of chemotherapy is correlated to PFI: in platinum-sensitive patients PLD is cost/effective compared with paclitaxel or topotecan but less cost/effective if compared with carboplatin/paclitaxel combination; in platinum-refractory patients best supportive care could be cost/effective compared to PLD, topotecan or gemcitabine combination in unfit patients [40, 41].
In conclusion, our retrospective experience in a consecutive series of serous papillary OC challenges the current believes on PLD activity in OC. Further prospective studies are needed to confirm our results, and molecular analysis as well as investigations on the peritoneal microenvironment are eagerly awaited to shed light on the molecular basis of the response to PLD, in order to identify patients who would gain benefit from the treatment with this agent.