This study was terminated due to limited enrollment.

In this study, the efficacy and safety of nilotinib 400 mg twice daily, will be compared with imatinib 400 mg twice daily in patients with a suboptimal response to imatinib for their Philadelphia chromosome-positive (Ph+) Chronic Myelogenous Leukemia in the chronic phase (CML-CP).

This trial was to evaluate the CCyR rate at 12 months of nilotinib therapy when compared to imatinib treatment in patients with suboptimal response to imatinib. The patients were stratified by prior duration of initial imatinib treatment, and were randomized to receive either 400 mg/twice daily of continuous nilotinib or imatinib treatment. The first stratum patients were treated with imatinib = 6 to < 12 months and having at least a minimal cytogenetic, but no partial cytogenetic response; and the second stratum patients were treated with imatinib = 12 months to < 18 months and having partial cytogenetic response (PCyR), but no CCyR.

Inclusion criteria:

Diagnosis of Philadelphia chromosome positive chronic myelogenous leukemia in the chronic phase.

Patients with suboptimal cytogenetic response to a dose of 400 mg imatinib (first line therapy) defined as:

• 6 to < 12 months of treatment and -have 36 - 95% Ph+ metaphases, or
• 12 to <18 months of treatment and have 1 - 35% Ph+ metaphases (Standard cytogenetics, no FISH [fluorescence in situ hybridization] analysis was allowed).

Exclusion criteria:

• Patient who have received more than 18 months of imatinib therapy
• Patients who have achieved partial or complete cytogenetic response and lost that response prior to entering the study.
• Prior treatment with greater than 400 mg/day imatinib.
• Uncontrolled or significant cardiovascular disease.
• Severe or uncontrolled medical conditions (i.e. uncontrolled diabetes, active or uncontrolled infection).
• Use of therapeutic coumarin derivatives (i.e., warfarin, acenocoumarol, phenprocoumon)
• Currently taking certain medications that could affect the rhythm of your heart.

Other protocol-defined inclusion/exclusion criteria may apply