The objective of this study was to analyze retrospective populations with recurrent ovarian cancer to assess differences in CA-125 patterns during chemotherapy. The populations included all patients treated between January 1994 and January 2004, who received liposomal doxorubicin and topotecan, and all patients treated between July 1997 and June 2001, who received carboplatin. Prognostic variables were abstracted from the medical records. Eighty-nine patients received liposomal doxorubicin and topotecan therapy and 21 received carboplatin; of these, 59 (liposomal doxorubicin), 60 (topotecan), and 17 (carboplatin) patients had evaluable CA-125 patterns. Patients given liposomal doxorubicin were more likely to have received only one or two cycles of therapy (37/89 [42%]) than patients receiving either carboplatin (5/21 [24%]) or topotecan (20/89[22%]). In cycle 1, CA-125 increases in patients were carboplatin, 4/17 (24%); liposomal doxorubicin, 41/59 (69%); and topotecan, 11/60 (18%). In cycle 2, CA-125 increases were carboplatin, 2/16 (13%); liposomal doxorubicin, 19/37 (51%); and topotecan, 9/50 (18%). In cycle 3, CA-125 increases were carboplatin, 0/12 (0%); liposomal doxorubicin, 7/23 (30%); and topotecan, 6/38 (16%). Of patients having any CA-125 decrease and given two or more cycles, fewer declines were seen in those given liposomal doxorubicin precycle 2 (18/35[51%]) than in those given carboplatin (13/16[81%]) or topotecan (49/56[88%]). The most prominent delay in CA-125 decline was in patients given liposomal doxorubicin compared with those given topotecan or carboplatin. In the entire population, only 3 of 107 (2.8%) patients demonstrated first CA-125 decline precycle 4. Discontinuation of therapy solely on the basis of early CA-125 increase (precycle 3), particularly with liposomal doxorubicin chemotherapy, may exclude some patients who will benefit from continued therapy.