Program DHC-Virtual
20 - 21 January 2021
Clinical Abstracts session 4
Conditional relative survival among long-term AML survivors
21 January
11:42 11:54
A. Dinmohamed

Age-specific conditional relative survival among long-term acute myeloid leukemia (AML) survivors

Avinash Dinmohamed (1,2,3), Otto Visser (4), Ward Posthuma (5,6), Gerwin Huls (7), Arjan van de Loosdrecht (3)
(1) Netherlands Comprehensive Cancer Organisation (IKNL), Research and Development, Utrecht, (2) Erasmus MC, Public Health, Rotterdam, (3) Amsterdam UMC, Hematology, Amsterdam, (4) Netherlands Comprehensive Cancer Organisation (IKNL), Registration, Utrecht, (5) Reinier de Graaf Gasthuis, Internal Medicine, Delft, (6) Leiden University Medical Center, Hematology, Leiden, (7) University Medical Center Groningen, Hematology, Groningen
No potential conflicts of interest

The connotation surrounding AML is often grim because of the heightened risk of mortality within the first few years postdiagnosis. Therefore, survival estimates measured from diagnosis might be somewhat pessimistic for long-term AML survivors and their caregivers who desire to obtain information on the prognosis after having survived some years post-diagnosis. Against this background, this population-based study aimed to estimate the conditional 5-year relative survival among AML patients up to fifteen years post-diagnosis.


All over 18-year-old AML patients diagnosed between 1989-2018—with follow-up for survival through December 31, 2019—were selected from the Netherlands Cancer Registry. We excluded acute promyelocytic leukemia patients. Relative survival (RS) was calculated to estimate disease-specific survival as the ratio of the observed to the expected survival, based on national life tables stratified by age, sex, and calendar year. We computed the conditional 5-year RS for each additional year survived up to fifteen years post-diagnosis. RS was estimated using a hybrid approach to predict up-to-date survival probabilities. This approach was applied for patients diagnosed between 1989-2018 who were alive during the follow-up interval 2008-2019. The produced estimates can be interpreted as the predicted survival probability of patients diagnosed during 2008-2019. Survival estimates were computed according to age at diagnosis (18-40, 41-60, 61-70, and >70 years). Only survival estimates with a standard error of ≤5% were presented. This choice relates to the precision of the survival estimates. Excess mortality is minimal when RS is ≥95%.


A total of 16,177 adult AML patients (median age, 68 years; and 55% males) were diagnosed in the Netherlands between 1989-2018, of whom 8,182 (50,6%) were alive at some point during the follow-up interval and contributed to the survival estimates. As expected, there was a prognostic effect of age on 5-year RS at diagnosis (Figure). This effect essentially persisted up to fifteen years post-diagnosis (Figure). Notwithstanding, excess mortality decreased with each additional year survived across all age groups (Figure). Encouragingly enough, patients aged 18-40 years at diagnosis hardly experienced excess mortality after four years post-diagnosis (Figure). In contrast, excess mortality remained substantial for patients aged >40 years at diagnosis, especially among those aged >60 years (Figure).




The prognosis of adult AML patients improved with each additional year survived. Reassuring is that patients aged 18-40 years can look forward to a normal life expectancy after four years post-diagnosis. This finding, unfortunately, does not hold for patients aged >40 years. Statistics on CRS in AML may offer patients and physicians with a better apprehension of the changing survival expectations over time. This information can also be used to tailor age-specific surveillance and follow-up activities.