Defining the Optimal Total Number of Chemotherapy Courses in Younger Patients With Acute Myeloid Leukemia: A Comparison of Three Versus Four Courses

Research output: Contribution to journalArticlepeer-review


  • Alan K Burnett
  • Nigel H Russell
  • Robert K Hills
  • Stephen Knapper
  • Brian Huntly
  • Richard E Clark
  • Ian F Thomas
  • Lars Kjeldsen
  • Mary Frances McMullin
  • Mark Drummond
  • Jonathan Kell
  • Ruth Spearing

Colleges, School and Institutes

External organisations

  • Cardiff University
  • Nottingham University Hospitals NHS Trust
  • University of Oxford
  • University of Birmingham
  • University of Cambridge
  • The Royal Liverpool and Broadgreen University Hospital Trust
  • Copenhagen University Hospitals
  • Queen's University
  • Department of Haematology
  • Immunodeficiency Centre for Wales, University Hospital of Wales, Cardiff (UK)
  • Canterbury District Health Board


PURPOSE: The optimum number of treatment courses for younger patients with acute myeloid leukemia (AML) is uncertain. The United Kingdom National Cancer Research Institute AML17 trial randomly assigned patients who were not high risk to a total of three versus four courses.

PATIENTS AND METHODS: Patients received two induction courses based on daunorubicin and cytarabine (Ara-C), usually with gemtuzumab ozogamicin. Following remission, 1,017 patients were randomly assigned to a third course, MACE (amsacrine, etoposide, and Ara-C), plus a fourth course of MidAc (mitoxantrone and Ara-C) and following an amendment to one or two courses of high-dose Ara-C. Primary end points were cumulative incidence of relapse (CIR), relapse-free survival (RFS), and overall survival (OS). Outcomes were correlated with patient characteristics, mutations, cytogenetics, induction treatments, and measurable residual disease (MRD) postinduction.

RESULTS: In logrank analyses, CIR and RFS at 5 years were improved in recipients of four courses (50% v 58%: hazard ratio [HR] 0.81 [0.69-0.97], P = .02 and 43% v 36%: HR 0.83 [0.71-0.98], P = .03, respectively). While OS was not significantly better (63% v 57%: HR 0.84 [0.69-1.03], P = .09), the noninferiority of three courses to four courses was not established. The impact on relapse was only significant when the fourth course was Ara-C. In exploratory analyses, although MRD impacted survival, a fourth course had no effect in either MRD-positive or MRD-negative patients. A fourth course was beneficial in patients who lacked a mutation of FLT3 or NPM1, had < 3 mutations in other genes, or had a presenting WBC of < 10 × 109 L-1.

CONCLUSION: Although a fourth course of high-dose Ara-C reduced CIR and improved RFS, it did not result in a significant OS benefit. Subsets including those with favorable cytogenetics, those lacking a mutation of FLT3 or NPM1, or those with < 3 other mutations may derive survival benefit.


Original languageEnglish
Pages (from-to)JCO2001170
JournalJournal of Clinical Oncology
Publication statusE-pub ahead of print - 23 Dec 2020