Performance of prognostic scoring systems in elderly patients with acute myeloid leukaemia on intensive chemotherapy: A PETHEMA registry study

Selection of elderly patients (aged ≥60 years) for intensive chemotherapy treatment of acute myeloblastic leukaemia (AML) remains challenging. Several cooperative groups such as Acute Leukaemia French Association (ALFA), Haematological Oncology Clinical Studies Group (HOCSG) and MD Anderson Cancer Center (MDACC) have developed predictive models to select those patients who can benefit from intensive chemotherapy. Our purpose is to validate and compare these three models in a cohort of patients treated in real-life setting. For this, a total of 1724 elderly AML patients and treated with intensive chemotherapy regimens were identified in the PETHEMA registry. Median age was 67.2 years (range, 60–84,9) and median overall survival [OS] 9 months (95 % confidence interval [CI], 8.2–9.7). Taking into account the ALFA group's model, patients likely to benefit from intensive chemotherapy had longer OS (14 months, 95 % CI 12.3–15.7) than those unlikely to benefit (5 months, 95 % CI 4.1–5.9; p < 0.001). Significant differences in OS were observed between patients with favourable risk (17 months, 95 % CI 13.2–20.7), intermediate risk (11 months, 95 % CI 9.3–12.6) and adverse risk (6 months, 95 % CI 5.1–6.4; p < 0.001) according to the HOCSG model. No significant differences in OS were observed between patients with 0, 1, 2 or ≥3 points according to the MDACC model. However, when patients with ≥1 point were compared with those with 0 points, median OS was significantly longer in the latter [15 months (95 % CI 12.1–17.8) vs 7 (95 % CI 5.7–8.5)]. This retrospective study validates predictive models proposed by the ALFA, HOCSG and MDACC groups in this real-life cohort.