Day-21 bone marrow findings incorrectly designate residual leukaemia in FLT3-mutated acute myeloid leukaemia treated with intensive induction plus midostaurin: a morphology-focused study

Abstract

Early induction response assessment with day-21 bone marrow (D21-BM) is commonly performed in patients with FLT3-mutated acute myeloid leukaemia (AML), where detection of residual leukaemia (RL; blasts ≥5%) typically results in the administration of a second induction course. However, whether D21-BM results predict for RL at the end of first induction has not been systematically assessed. This study evaluates the predictive role of D21-BM morphology in detecting RL following first induction. Between August 2018 and March 2022, all patients with FLT3-AML receiving 7+3 plus midostaurin, with D21-BM performed, were identified. Correlation between D21-BM morphology vs D21-BM ancillary flow/molecular results, as well as vs D28-BM end of first induction response, were retrospectively reviewed. Subsequently, D21-BMs were subjected to anonymised morphological re-assessments by independent haematopathologists (total in triplicate per patient). Of nine patients included in this study, three (33%) were designated to have RL at D21-BM, all of whom entered complete remission at D28-BM. Furthermore, only low-level measurable residual disease was detected in all three cases by flow or molecular methods at D21-BM, hence none proceeded to a second induction. Independent re-evaluations of these cases failed to correctly reassign D21-BM responses, yielding a final false positive rate of 33%. In summary, based on morphology alone, D21-BM assessment following 7+3 intensive induction plus midostaurin for FLT3-AML incorrectly designates RL in some patients; thus correlating with associated flow and molecular results is essential before concluding RL following first induction. Where remission status is unclear, repeat D28-BMs should be performed.