AUTHOR=Wang Xiaoning , Wang Jie , Wei Suhua , Zhao Juan , Xin Beibei , Li Guoqing , Zhao Jing , Wu Di , Luo Minna , Zhao Sijie , Chen Ying , Liu Haibo , Zhang Hailing , Wang Jingcheng , Wang Wenjuan , Wang Huaiyu , Xiong Hui , He Pengcheng TITLE=The latest edition of WHO and ELN guidance and a new risk model for Chinese acute myeloid leukemia patients JOURNAL=Frontiers in Medicine VOLUME=10 YEAR=2023 URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2023.1165445 DOI=10.3389/fmed.2023.1165445 ISSN=2296-858X ABSTRACT=Objective

Diagnosis classification and risk stratification are crucial in the prognosis prediction and treatment selection of acute myeloid leukemia (AML). Here, we used a database of 536 AML patients to compare the 4th and 5th WHO classifications and the 2017 and 2022 versions of ELN guidance.

Methods

AML patients were classified according to the 4th and 5th WHO classifications, as well as the 2017 and 2022 versions of the European LeukemiaNet (ELN) guidance. Kaplan–Meier curves with log-rank tests were used for survival analysis.

Results

The biggest change was that 25 (5.2%), 8 (1.6%), and 1 (0.2%) patients in the AML, not otherwise specified (NOS) group according to the 4th WHO classification, were re-classified into the AML-MR (myelodysplasia-related), KMT2A rearrangement, and NUP98 rearrangement subgroups based on the 5th WHO classification. Referring to the ELN guidance, 16 patients in the favorable group, six patients in the adverse group, and 13 patients in the intermediate group based on the 2017 ELN guidance were re-classified to the intermediate and adverse groups based on the 2022 ELN guidance. Regrettably, the Kaplan–Meier curves showed that the survival of intermediate and adverse groups could not be distinguished well according to either the 2017 or 2022 ELN guidance. To this end, we constructed a risk model for Chinese AML patients, in which the clinical information (age and gender), gene mutations (NPM1, RUNX1, SH2B3, and TP53), and fusions (CBFB::MYH11 and RUNX1::RUNX1T1) were included, and our model could help divide the patients into favorable, intermediate, and adverse groups.

Conclusion

These results affirmed the clinical value of both WHO and ELN, but a more suitable prognosis model should be established in Chinese cohorts, such as the models we proposed.