A new combination of carboplatin, high-dose cytarabine and cross-over mitoxantrone or idarubicin for refractory and relapsed acute myeloid leukemia.

BACKGROUND AND OBJECTIVE: High-dose cytarabine (HIDAC) and new anthracycline-type drugs (mitoxantrone, idarubicin) are the mainstay of several active regimens against relapsed and refractory acute myeloid leukemia (AML). The present study was undertaken to assess the feasibility, toxicity, and antileukemic activity of carboplatin (CBDCA) added to a combination of the two former agents. DESIGN AND METHODS: Two regimens (R) of CBDCA plus HIDAC and either mitoxantrone or idarubicin (crossover) were sequentially evaluated. R-1 consisted of CBDCA 300 mg/m2/d (24-hour infusion) on days 1-4, HIDAC 1 g/m2/bd on days 1-5, and mitoxantrone/idarubicin 12/6 mg/m2/d on days 1-3, followed by granulocyte colony-stimulating factor (G-CSF). R-2, an attenuated-toxicity regimen, consisted of CBDCA and G-CSF as above, HIDAC on alternate days (1, 3, 5), and mitoxantrone/idarubicin 8/5 mg/m2/dose. Intended post-remission therapy included a similar, lower intensity course and a myeloablative phase supported by an allogeneic or autologous blood cell transplant. RESULTS: Twenty-nine patients (median age 53 years, one child) formed the study group: 10 (34%) had a primary refractory disease (8 to idarubicin-cytarabine-etoposide, ICE), 6 (21%) were at second or subsequent relapse, and 5 (17%) had a first remission lasting < 12 months. In addition, 4 patients (14%) had received prior HIDAC and 10 (34%) were relapsing after a bone marrow/blood cell transplant. Twelve patients were treated with R-1 and 17 with R-2. The complete response rate was 25% with R-1 and 53% with R-2, due to a significantly lower death rate by pancytopenic complications (p = 0.023). The probability of response by risk class was: primary refractory 30% (43% with R-2), > 2nd relapse 33% (50% with R-2), 1st relapse < 12 months 40% (50% with R-2), 1st relapse > 12 months 50% (75% with R-2), prior HIDAC 75%, and prior transplant 30% (33% with R-2). Seven patients could undergo an autologous (n = 5) or allogeneic (n = 2) bone marrow/peripheral blood cell transplant after one consolidation cycle. Overall survival was 4.2 months, significantly longer in responders (complete and partial: median 11 months) than non-responders (p < 0.001). Median duration of complete remission was 10 months and 2-year probability 0.31, but no patient remained disease-free at 3 years. INTERPRETATION AND CONCLUSIONS: R-2 was well tolerated, exerted a significant activity in high-risk AML, and is amenable to further improvements. However, the lack of long-term disease-free survivors indicates the need for innovative post-remission strategies.