Development and Validation of an Individualized Metabolism-Related Prognostic Model for Adult Acute Myeloid Leukemia Patients

Cong Wei; Lijuan Ding; Qian Luo; Xiaoqing Li; Xiangjun Zeng; Delin Kong; Xiaohong Yu; Jingjing Feng; Yishan Ye; Limengmeng Wang; He Huang

doi:10.3389/fonc.2022.829007

Development and Validation of an Individualized Metabolism-Related Prognostic Model for Adult Acute Myeloid Leukemia Patients

Front Oncol. 2022 Jun 17:12:829007. doi: 10.3389/fonc.2022.829007. eCollection 2022.

Authors

Cong Wei^{1

2

3}, Lijuan Ding^{1

2

3}, Qian Luo^{1

2

3}, Xiaoqing Li^{1

2

3}, Xiangjun Zeng^{1

2

3}, Delin Kong^{1

2

3}, Xiaohong Yu^{1

3}, Jingjing Feng^{1

2

3}, Yishan Ye^{1

2

3}, Limengmeng Wang^{1

2

3

4}, He Huang^{1

2

3

4}

Affiliations

¹ Bone Marrow Transplantation Center, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China.
² Liangzhu Laboratory, Zhejiang University Medical Center, Hangzhou, China.
³ Institute of Hematology, Zhejiang University, Hangzhou, China.
⁴ Zhejiang Province Engineering Laboratory for Stem Cell and Immunity Therapy, Hangzhou, China.

Abstract

Objectives: Acute myeloid leukemia (AML) is a highly heterogeneous hematologic malignancy with widely variable prognosis. For this reason, a more tailored-stratified approach for prognosis is urgently needed to improve the treatment success rates of AML patients.

Methods: In the investigation of metabolic pattern in AML patients, we developed a metabolism-related prognostic model, which was consisted of metabolism-related gene pairs (MRGPs) identified by pairwise comparison. Furthermore, we analyzed the predictive ability and clinical significance of the prognostic model.

Results: Given the significant differences in metabolic pathways between AML patients and healthy donors, we proposed a metabolism-related prognostic signature index (MRPSI) consisting of three MRGPs, which were remarkedly related with the overall survival of AML patients in the training set. The association of MRPSI with prognosis was also validated in two other independent cohorts, suggesting that high MRPSI score can identify patients with poor prognosis. The MRPSI and age were confirmed to be independent prognostic factors via multivariate Cox regression analysis. Furthermore, we combined MRPSI with age and constructed a composite metabolism-clinical prognostic model index (MCPMI), which demonstrated better prognostic accuracy in all cohorts. Stratification analysis and multivariate Cox regression analysis revealed that the MCPMI was an independent prognostic factor. By estimating the sensitivity of anti-cancer drugs in different AML patients, we selected five drugs that were more sensitive to patients in MCPMI-high group than those in MCPMI-low group.

Conclusion: Our study provided an individualized metabolism-related prognostic model that identified high-risk patients and revealed new potential therapeutic drugs for AML patients with poor prognosis.

Keywords: MCPMI; MRPSI; acute myeloid leukemia; drug response; prognosis.