A new model to predict remission status in AML patients based on day 14 bone marrow biopsy

Although bone marrow evaluation on day 14 after initiation of induction chemotherapy (D14 BM) is a widely accepted practice in patients with acute myeloid leukemia (AML), it has suboptimal predictive value for predicting complete remission. We retrospectively analyzed pretreatment characteristics and post-induction response in a cohort of AML patients to determine if adding clinical and laboratory characteristics can improve the predictive value of the D14 BM evaluation. Among 297 patients treated for AML at the single institution 183 patients (61%) had leukemia-positive D14 BM. Of those, 94 were given reinduction chemotherapy and 89 were not. Of the 89 patients who did not receive reinduction, 32 (36%) subsequently achieved complete remission (CR) or complete remission with incomplete count recovery (CRi), and 57 (64%) had persistent disease. Persistent disease after positive D14 BM was more likely associated with higher percentage of D14 myeloblasts, a history of relapsed disease before induction, and higher risk disease compared to patients who subsequently achieved CR. Age, diagnostic white blood cell count, and the D14 BM cellularity did not influence the subsequent likelihood of achieving remission in patients with a positive D14 BM. A new mathematical equation was created and resulted in a positive predictive value of 83%, negative predictive value 90% and accuracy 88% for correctly identifying remission status after positive D14 BM in AML. The accuracy of predicting response using these additional parameters was significantly higher than without (0.88 vs. 0.80, P=0.002). Our new model provides better accuracy for predicting the likelihood of achieving remission and if validated in future studies may be useful for managing AML patients.