Purpose: Serum beta(2) microglobulin (beta2M) is prognostic in other hematologic malignancies; therefore, we evaluated its prognostic significance in acute myeloid leukemia (AML).
Experimental design: Multivariate analyses were used to examine the effect of pretreatment serum beta2M levels on clinical outcomes in patients with AML. beta2M was associated with poorer survival in older but not younger patients. We thus fit separate Cox survival models in patients above and below age 60 years treated with remission induction therapy containing high-dose cytarabine (n = 1,280). In each age group, 50% of the patients were used to develop the model, which was tested in the other 50%. Resampling methods were also used to validate the independent prognostic significance of covariates.
Results: In patients 60 years or older (n = 591), poorer risk cytogenetics; poorer performance status; and higher levels of beta2M, uric acid, and lactate dehydrogenase were each found to independently predict shorter survival and formed the basis of a scoring system. A similar approach was used in patients younger than 60 years (n = 589), with poorer risk cytogenetics, poorer performance status, older age, higher hemoglobin level, and higher leukocyte count predicting a shorter survival and forming the basis of the scoring system. Higher beta2M levels were an adverse independent factor for response, survival, relapse-free survival, and event-free survival in older but not in younger patients.
Conclusions: Serum beta2M levels can help predict outcome in patients > or =60 years with untreated AML, and their use is strongly encouraged.