AUTHORS
Guanghui Chang, Hongyu Zhang, PhD2, Yajing Xu3, Jinhai Ren, MD4, Yu Zhang5, Pengcheng He6, Liping Dou, MD7, Zhi Guo, PhD8, Yongqian Jia, MD9, Haitao Gao10, Xiaosu Zhao11, Guorui Ruan, PhD12, Qian Jiang Jr, MD13, Hao Jiang14, Lanping Xu, MD15, Xiaohui Zhang, MD, PhD16, Xiaojun Huang, MD, PhD17, Feifei Tang, MD, PhD18
AFFILIATIONS
1 Peking University People's Hospital, Beijing, Beijing, China
2 Peking University Shenzhen Hospital, shenzhen, China
3 Xiangya Hospital, Changsha, China
4 The Second Hospital of Hebei Medical University, Shijiazhuang, China
5 Nanfang Hospital, Southern Medical University, Guangzhou, China
6 The First Affiliated Hospital of Xi’an Jiaotong University, Xi'an, China
7 Chinese PLA General Hospital, Beijing, China
8 Huazhong University of Science and Technology Union Shenzhen Hospital, Shenzhen, China
9 West China Hospital, Chengdu, China
10 Peking University People's Hospital, Beijing, Beijing, China
11 Peking University People's Hospital, Beijing, Beijing, China
12 Peking University People's Hospital, Beijing, Beijing, China
13 Peking University People's Hospital, Beijing, Beijing, China
14 Peking University People's Hospital, Beijing, Beijing, China
15 Peking University People's Hospital, Beijing, Beijing, China
16 Peking University People's Hospital, Beijing, Beijing, China
17 Peking University People's Hospital, Beijing, Beijing, China, Peking-Tsinghua Center for Life Sciences, Academy for Advanced Interdisciplinary Studie, Beijing, China
18 Peking University People's Hospital, Beijing, Beijing, China
ABSTRACT
Background:Acute myeloid leukemia (AML) with RUNX1::RUNX1T1 is recognized as a distinct core-binding factor AML subtype and is generally associated with favorable clinical outcomes. However, therapy-related AML is usually characterized by adverse genetic features, lower response rates, and inferior survival compared with de novo AML. Whether the favorable biological and prognostic effects of RUNX1::RUNX1T1 can offset the adverse impact of a therapy-related background remains unclear.
Aims:This study aimed to determine whether the RUNX1::RUNX1T1 can offset the adverse impact of a therapy-related background by comparing the differences between therapy-related RUNX1::RUNX1T1-positive AML (t-ETO AML), RUNX1::RUNX1T1-negative therapy-related AML (t-AML) and de novo RUNX1::RUNX1T1-positive AML (dn-ETO AML).
Methods:This study compared the genetics, clinical baseline and outcomes of patients with t-ETO AML, with t-AML, and with dn-ETO AML. All patients were newly diagnosed AML and received treatment at 9 centers in China from January 2015 to December 2024.
Results: A total of 713 patients were included: 38 with t-ETO AML, 227 with t-AML and 448 with dn-ETO AML. Patients with t-ETO AML were older than those with dn-ETO AML but younger than those with t-AML (median age: 47, 35 and 53 years, respectively; P<0.001 and P=0.031). Both t-ETO and dn-ETO AML were characterized by higher KIT mutation frequencies (39.1% vs. 43.6% vs. 6.5%, P<0.001) and lower FLT3-ITD frequencies (0% vs. 6.8% vs. 17.8%, P<0.001) than t-AML. The composite complete remission rate of t-ETO was comparable to that in dn-ETO AML (86.8% vs 94.2%, P=0.084) but higher than that in t-AML (86.8% vs 66.1%, P=0.013). t-ETO AML and dn-ETO AML showed comparable survival and better outcomes than t-AML, with 3-year OS rates of 66.5%, 80.1%, and 48.2%, respectively (t-ETO vs dn-ETO, P=0.111; t-ETO vs t-AML, P=0.007), and 3-year DFS rates of 67.8%, 71.8%, and 55.9% (P=0.793 and P=0.040, respectively). Within therapy-related AML, RUNX1::RUNX1T1 remained associated with superior OS after multivariate analysis (MVA) (HR=0.336, P=0.021). However, among RUNX1::RUNX1T1-positive patients, therapy-related status didn’t impact OS or DFS after MVA.
The rates of Major Molecular Response (MMR) were similar between t-ETO AML and de novo ETO AML. After the first, second, and third treatment courses, the MMR rates were 14.3% (2/14) versus 22.5% (71/316; P=0.743), 52.9% (9/17) versus 54.2% (195/360; P=1.000), and 57.9% (11/19) versus 71.3% (258/362; P=0.208), respectively. In de novo ETO AML, course-1 MMR did not consistently predict survival (MMR vs non-MMR: 3-year OS, 94.2% vs 83.2%, P=0.061; DFS, 88.6% vs 72.8%, P=0.014), whereas MMR after course 2 (OS, 89.0% vs 77.2%, P=0.001; DFS, 81.2% vs 68.0%, P=0.004) and course 3 (OS, 88.7% vs 68.5%, P<0.001; DFS, 79.1% vs 63.7%, P<0.001) was associated with improved OS and DFS. In t-ETO AML, only course-2 MMR was associated with improved survival (OS, 88.9% vs 37.5%, P=0.049; DFS, 88.9% vs 25.0%, P=0.008). Delayed MMR was defined as failure to achieve MMR after course 2 followed by achievement of MMR after course 3. In t-ETO AML, patients achieving MMR by course 2 had superior survival compared with those with persistent non-MMR through course 3 (3-year OS, 88.9% vs 33.3%, P=0.044; 3-year DFS, 88.9% vs 16.7%, P=0.005). In de novo ETO AML, patients achieving MMR by course 2 had significantly superior survival compared with both the delayed-MMR and persistent non-MMR groups (3-year OS, 89.7% vs 77.0% and 71.3%; 3-year DFS, 82.1% vs 69.1% and 66.0%; course-2 MMR vs delayed MMR: OS, P=0.009 and DFS, P=0.033; course-2 MMR vs persistent non-MMR: both P<0.001). MVA showed that course-2 MMR was the principal favorable prognostic marker in both RUNX1::RUNX1T1-positive subgroups (t-ETO, DFS, HR=0.098, P=0.032; dn-ETO, OS, HR=0.421, P=0.002, DFS, HR=0.414, P<0.001).
Summary: Therapy-related RUNX1::RUNX1T1-positive AML showed similar genetic features, remission rates, molecular response kinetics, and survival outcomes to de novo RUNX1::RUNX1T1-positive AML. Therapy-related status was not independently associated with inferior outcomes among RUNX1::RUNX1T1-positive patients. MMR after two treatment courses was the earliest and most informative prognostic milestone.
DISCLOSURES
Guanghui Chang: No relevant conflicts of interest to declare.