In July 2023, a significant study led by Professor Xiaojun Huang from Peking University Institute of Hematology was published in Leukemia, focusing on chronic myeloid leukemia (CML) treatment with imatinib. This research introduced and validated a predictive scoring system to forecast molecular responses in CML patients. The study, involving an external validation on a cohort of 2,184 patients, demonstrated the efficacy of this model in stratifying patients into risk categories for more personalized treatment approaches. Through rigorous statistical analysis, including ROC curves and decision curve analyses, Huang and his team’s work stands as a pivotal contribution to optimizing CML management and enhancing patient care outcomes.
Chronic Myeloid Leukemia (CML) is a type of cancer that affects the blood and bone marrow and is characterized by the overproduction of myeloid cells. Imatinib, a tyrosine kinase inhibitor, has revolutionized the treatment of CML, offering patients a chance at a longer and healthier life. Predictive scoring systems that can forecast molecular response to imatinib are crucial for personalized patient management. However, the effectiveness and accuracy of these systems across diverse populations remain to be thoroughly validated. This study focuses on the external validation of a predictive scoring system, originally developed and internally validated using a Fine-Gray model, to predict molecular responses in patients with chronic phase CML receiving imatinib as their initial therapy. By applying this model to a larger, independent cohort, this research aims to confirm its utility and reliability in a real-world setting.
The study cohort consisted of 2184 patients with chronic phase CML, treated initially with imatinib. Each subject’s BCR::ABL1 transcript levels were meticulously quantified using quantitative real-time polymerase chain reaction (qRT-PCR), utilizing an ABL control gene and adjusting to international scales (BCR::ABL1IS) for values below 10%. This process ensured the accuracy of molecular response assessments, crucial for evaluating imatinib’s effectiveness. The demographic spread included a diverse range of ages and a balanced gender distribution, providing a robust dataset for external validation. The median follow-up period was 51 months, allowing for a comprehensive analysis of long-term treatment outcomes.
The predictive models successfully stratified patients into low-, intermediate-, and high-risk groups for major molecular response (MMR) and stable molecular response (MR4), with distinct outcomes observed across these cohorts. Specifically, 64% of subjects achieved MMR, while 43% reached a stable MR4, underscoring the effectiveness of imatinib in managing CML. Furthermore, the 7-year cumulative incidences of MMR were significantly different across risk groups, highlighting the model’s ability to discriminate between varying levels of response to treatment. Platelet concentration emerged as a significant predictor, with lower levels associated with reduced rates of MMR and MR4, suggesting its potential as a biomarker for treatment monitoring and adjustment.
Fig. Cumulative incidences of MMR and MR4 by the predictive scoring systems. A Cumulative incidences of MMR. B Cumulative incidences of MR4.
The study’s rigorous validation process, including ROC curves, calibration plots, and decision curve analyses, demonstrated the models’ high degree of accuracy in predicting molecular responses. However, the study’s retrospective nature and the specific ethnic composition of the cohort (exclusively Chinese patients younger than the European CML population) may limit the generalizability of the findings. The lack of differentiation between e14a2 and e13a2 BCR::ABL1 transcripts in the analysis could also affect the applicability of the results to all CML patients.
This external validation study confirms the predictive scoring system’s effectiveness for forecasting molecular responses in chronic phase CML patients undergoing initial imatinib therapy. The findings endorse the model’s utility in clinical settings, facilitating personalized treatment strategies that could significantly improve patient outcomes.
