Pb1968: impacts on outcomes and a predictive model of severe thrombocytopenia in patients with chronic phase chronic myeloid leukemia receiving initial nilotinib-therapy

Topic: 8. Chronic myeloid leukemia - Clinical Background: Although nilotinib as initial therapy led to faster and deeper responses and reduced risk of disease progression in subjects with chronic phase chronic myeloid leukemia (CML), 8-26% subjects developed severe (≥ grade 3) thrombocytopenia. Rare data reported the impacts of severe thrombocytopenia on treatment outcomes, and no robust scoring system can predict severe thrombocytopenia at early stage of initial nilotinib-therapy. Aims: To explore the impacts of severe thrombocytopenia on treatment outcomes and to develop a predictive scoring system for severe thrombocytopenia in patients with chronic phase CML receiving initial nilotinib-therapy. Methods: Data from consecutive subjects ≥ 18 years with newly diagnosed chronic phase CML receiving initial nilotinib-therapy at Peking University People’s Hospital were interrogated. Fine-gray and Cox regression models were applied to identify impacts of severe thrombocytopenia on treatment outcomes. Co-variates with p < 0.1 in multi-variable logistic analysis were included to establish a predictive scoring system for severe thrombocytopenia. The system was subsequently validated in subjects from 62 other Chinese centers. Receiver-operator characteristic curve, calibration plots and decision curve analyses were applied to evaluate the system. Results: Data from 288 consecutive subjects receiving initial nilotinib-therapy was interrogated as training dataset. In the training dataset, 46 (16%) subjects developed grade 3 (n = 21, 7%) or grade 4 (n = 25, 9%) thrombocytopenia at a median of 29 (interquartile range [IQR], 24-48) days after starting nilotinib-therapy, with a median duration of 32 (IQR, 19-49) days. In multi-variable analyses severe thrombocytopenia was significantly-associated with lower cumulative incidences of complete cytogenetic response (HR = 0.5 [0.3, 0.7], p < 0.001), major molecular response (HR = 0.4 [0.3, 0.7], p < 0.001) and molecular response 4 (HR = 0.5 [0.2, 0.9], p = 0.025) and inferior failure-free survival (FFS, HR = 2.7 [1.5, 5.0], p = 0.001). Larger spleen size below costal margin, higher WBC counts and higher percentage of blood blasts were associated with severe thrombocytopenia and were included in the predictive scoring system. K-means clustering algorithm stratified subjects into low-, intermediate- and high-risk groups. The system was validated in 621 subjects from 62 other Chinese centers. Area under receiver-operator characteristic curve values of the scoring system was 0.73 (0.66, 0.81) in the training dataset and 0.73 (0.66, 0.79) in the validation dataset. Calibration plots indicated good concordance between predicted and observed incidences. Decision curve analyses indicated subjects could benefit from the system in clinical practice. Incidences of severe thrombocytopenia among the 3 risk cohorts in training or validation dataset were significantly-different (Figure 1). Summary/Conclusion: Nilotinib-related severe thrombocytopenia was a common hematologic adverse event and had negative impact on cytogenetic and molecular responses as well as FFS in patients with chronic phase CML. We developed a scoring system integrating WBC count, percentage of blood blasts and spleen size below costal margin to predict the development of severe thrombocytopenia during initial nilotinib-therapy.Figure 1. Observed incidences of severe thrombocytopenia by the predictive scoring system in the training and validation datasets. Keywords: Chronic myeloid leukemia, Thrombocytopenia, Nilotinib