AUTHORS
Mengyao Yuan, Jianyu Weng, MD, PhD2, Xin Du, MD3, Yu Zhu, MD4, Xiaodong Wang5, Zhenfang Liu, DO6, Hongmei Jing7, Lin Tan8, Yanli Zhang, MD9, Li Zhou10, Yan Wen, MD11, He-Bing Zhou I12, Congmeng Lin13, Yong Zhang14, Linhua Yang15, Yueqin Huang16, Cuicui Feng17, Liangding Hu18, Wenqing Hu19, Guohui Li20, Na Xu21, Zhuogang Liu22, Zhenling Li23, Chen Chunyan, MD24, Qian Jiang Jr, MD25
AFFILIATIONS
1 Peking University People's Hospital, Beijing, China
2 Guangdong Provincial People's Hospital, Guangzhou, China
3 The Second People’s Hospital of Shenzhen, The First Affiliated Hospital of Shenzhen University, Shenzhen, China
4 The First Affiliated Hospital of Nanjing Medical University, Jiangsu Province Hospital, Nanjing, China
5 Sichuan Academy of Medical Sciences Sichuan Provincial People's Hospital, Chengdu, China
6 The First Affiliated Hospital of Guangxi Medical University, Nanning, China
7 Peking University Third Hospital, Beijing, China
8 The First Affiliated Hospital of Kunming Medical University, Kunming, China
9 Affiliated Cancer Hospital of Zhengzhou University, Zhengzhou, China
10 Ruijin Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China
11 The First People's Hospital of Yunnan Province, The Affiliated Hospital of Kunming University of Science and Technology, Kunming, China
12 Beijing Luhe Hospital Affiliated to Capital Medical University, Beijing, China
13 Zhangzhou Municipal Hospital of Fujian Province, Zhangzhou, China
14 Civil Aviation General Hospital, Beijing, China
15 The Second Hospital of Shanxi Medical University, Taiyuan, China
16 The Second Affiliated Hospital of Fujian Medical University, Fujian, China
17 Beijing Anzhen Hospital,Capital Medical University, Beijing, China
18 The Fifth Medical Center, Chinese PLA General Hospital, Beijing, China
19 Peking University International Hospital, Beijing, China
20 The Affiliated Xi’an International Medical Center Hospital of Northwest University, Xian, China
21 Nanfang Hospital, Southern Medical University, Guangzhou, China
22 Shengjing Hospital of China Medical University, Shenyang, China
23 China-Japan Friendship Hospital, Beijing, China
24 Qilu Hospital of Shandong University, Jinan, China
25 Peking University People's Hospital, Beijing, China
ABSTRACT
Background Tyrosine kinase inhibitors (TKIs) have markedly improved outcomes in chronic-phase chronic myeloid leukemia (CML-CP), but TKI resistance remains a major challenge. More precise tools are needed to predict TKI resistance and inform frontline therapy choice.
Aims To develop and externally validate a genetic-clinical prediction model for TKI resistance and explore its potential role in optimizing frontline TKI therapy in newly diagnosed CML-CP patients.
Methods Targeted DNA sequencing was performed in newly-diagnosed CML-CP patients by 2022 WHO criteria. TKI-resistance was defined as 2025 ELN ‘unfavorable’ response, which indicates a high risk of developing resistance and a preference for treatment switch. Cox regression were used to identify prognostic co-variates associated with TKI resistance-free survival (RFS) in the imatinib cohort. Optimal cut-offs were determined using maximally selected rank statistics and kernel density estimation with 1,000 bootstrap resamples. The model was externally validated in an independent imatinib-treated cohort. Its prognostic performance was further evaluated in patients receiving frontline second-generation TKI (2G-TKI). Among pooled intermediate- and high-risk patients, RFS was compared between frontline imatinib and a 2G-TKI using inverse probability of treatment weighting (IPTW).
Results 804 CML-CP patients were included in this study, comprising an internal imatinib-treated training cohort (n = 345), an independent external imatinib-treated validation cohort (n = 109), and a frontline 2G-TKI cohort (n = 350) for further evaluating the prognostic performance of the model.
In the internal imatinib cohort, with a median follow-up of 57 months (IQR, 28-80), 56 of 345 (16%) patients experienced resistance at a median of 11 (IQR, 4-16) months. Multi-variable analyses showed that increasing splenomegaly, a higher a higher percentage of blood or bone marrow blasts, high-risk Ph+ ACAs and variants in ASXL1, RUNX1, NOTCH1 and FGFBP2 were significantly-associated with worse RFS. These genetic and clinical co-variates were integrated into an resistance risk score (GC), stratifying patients into low- (GC score < 1.189, n = 269, 78%), intermediate- (1.189 ≤ GC score < 1.683, n = 33, 10%) and high-risk (GC score ≥ 1.683, n = 43, 13%) subgroups with 6-year probabilities of RFS of 87% (95% CI, 82%, 92%), 70% (55%, 89%) and 39% (25%, 62%; p < 0.001). HRs with the low-risk subgroup as reference were 2.7 (1.2-5.9, p = 0.016) and 7.9 (4.4-14.2, p < 0.001; p-value for trend < 0.001).
The GC score was externally validated in an independent imatinib-treated cohort (n = 109), with a median follow-up of 22 months (IQR, 9-36). The 3-year RFS probabilities were 91% (84%, 99%) and 49% (30%, 79%) in the low- (n = 90, 83%) and intermediate-/high-risk (n = 19, 17%) subgroups, respectively (p < 0.001).
The prognostic performance of the GC score was further evaluated in a cohort of patients receiving frontline 2G-TKIs (n = 350), with a median follow-up of 31 months (IQR, 14–60). The 5-year RFS probabilities were 94% (90%, 98%) in the low-risk (n = 252, 72%) and 67% (56%, 79%) in the intermediate-/high-risk (n = 98, 28%) subgroups, respectively, demonstrating significant risk discrimination (p < 0.001).
Among patients classified as intermediate-/high-risk by the GC score, frontline 2G-TKI therapy was associated with superior RFS compared with imatinib. By IPTW analysis on well-balanced baseline characteristics of the imatinib (n = 95) and 2G-TKI (n = 98) cohorts, the 2G-TKI cohort had a higher 5-year RFS probability than the imatinib cohort (67% [54%, 78%] vs. 51% [36%, 63%]; HR, 0.6 [0.3-1.0]; p = 0.045).
Summary The genetic-clinical prediction model enables to identify the patients with newly diagnosed CML-CP at high risk of resistance and may help to optimize frontline TKI choice.
DISCLOSURES
Mengyao Yuan: No relevant conflicts of interest to declare.