用次世代定序探討急性骨髓性白血病病患的風險評估

Applying Next-generation Sequencing to Explore the Risk Stratification in Acute Myeloid Leukemia Patients

蔡承宏(Cheng-Hong Tsai)

指導教授 ： 陳倩瑜

共同指導教授 ： 田蕙芬(Hwei-Fang Tien)

國立臺灣大學/生命科學院/基因體與系統生物學學位學程/博士(2021年)

https://doi.org/10.6342/NTU202100880

摘要

急性骨髓性白血病是成人最常見的白血病之一，在台灣癌症造成的死亡率中歷年排名皆約第十名。。急性骨髓性白血病是一種臨床和生物學上相當異質的血液惡性腫瘤，其特點是不成熟的造血細胞不受控制地增殖，同時喪失分化能力。為了改善患者的治療效果與提高醫療照護品質，結合精準醫學的醫療照顧是非常關鍵的，包括診斷時的風險分層和基於治療反應的治療策略調整。近來，學界已經發現了許多影響預後的因素，著眼於這幾年生物技術方法學的發展，急性骨髓性白血病病患的存活率有望更進一步的提升。本論文將利用次世代定序技術來解決這些與急性骨髓性白血病預後相關的關鍵問題，分為三個部分：病患初診斷時的基因突變、非編碼核糖核酸的表現量和接受化學治療之後的可測/最小殘餘疾病之監測。
第一部分是初診斷時的cohesin基因突變。在其他骨髓性惡性腫瘤中，cohesin基因突變的臨床使用與預後意義已有文獻闡述，但其在急性骨髓性白血病的臨床預後意義和在治療過程中的動態變化仍有待發掘。本研究針對391位新診斷的急性骨髓性白血病患者，發現有9.5%的患者存在cohesin基因突變，最常見的是RAD21（3.8%）和STAG2（3.1%）基因突變。這些cohesin基因突變是總體存活期和無復發存活期的獨立的良好預後因子。在病患接受治療的追蹤中，有兩名患者在過程中失去了原有的cohesin基因突變，而在初診斷沒有任何cohesin基因突變的患者在復發時都沒有獲得新的突變。在其功能方面的分析顯示，cohesin基因突變涉及造血系統、骨髓與血液細胞相關的功能。
第二部分是初診斷時的長鏈非編碼核糖核酸（lncRNAs）表現量。近年來有些文獻認為，長鏈非編碼核糖核酸的表現量可能可以作為急性骨髓性白血病患者的預後評估依據。然而，將長鏈非編碼核糖核酸表現量納入目前使用最廣的2017年歐洲白血病網（ELN）風險分類中，是否能進一步改善預後評估系統仍有待研究。在本論文中，我們將275位急性骨髓性白血病患者隨機分為訓練組（n = 183）和驗證組（n = 92）。在訓練組中，我們建構了一個由五個長鏈非編碼核糖核酸表現量組成的評分系統，在這個評分系統中，高分患者的總體存活期與無復發存活期皆比低分患者要來的短，這個發現在驗證組中得到進一步證實。多變項統計分析顯示，此長鏈非編碼核糖核酸表現量組成的評分系統是獨立於2017年歐洲白血病網分類的預後因子。
第三部分是治療之後可測/最小殘餘疾病（MRD）之監測。次世代定序（NGS）目前已知可用於急性骨髓病白血病患者的殘餘疾病的監測，但該檢測的最佳應用時間點仍不明確。在本論文中，我們收集了303位急性骨髓性白血病患者在診斷時、首次完全緩解後（第一個時間點）、首次鞏固化療後（第二個時間點），共三個時間點的骨髓細胞，進行了次世代定序的分析。在第一個和第二個時間點分別有44.6%和27.7%的患者檢測到殘餘疾病。在第一或第二個時間點帶有殘餘疾病的患者，其累計復發率較高，總體存活期與無復發存活期較短。第一個時間點帶有殘餘疾病但在第二個時間點沒有殘餘疾病的患者，其預後和第一與第二個時間點皆無殘餘疾病的患者一樣好。
總結本論文發現，cohesin基因突變對AML患者的預後有良好的影響。這些突變在疾病進展中可能並不關鍵，因為在追蹤過程中，沒有任何原本不帶有cohesin突變的患者獲得突變。在2017年歐洲白血病網分類中納入長鏈非編碼核糖核酸評分系統可以改善AML患者的風險預後評估。最後，使用次世代定序來監測可測/最小殘餘疾病，發現第二個時間點偵測得到的可測/最小殘餘疾病，可以幫助預測患者的臨床預後。希望藉由在診斷和追蹤階段納入這些預後指標，臨床醫師可以更進一步擬定最有勝算的治療計畫與提供精準到位的治療，提升急性骨髓性白血病患者的存活率。

關鍵字

急性骨髓性白血病 ； 風險分層 ； 次世代定序 ； cohesin基因突變 ； 長鏈非編碼核糖核酸 ； 可測/最小殘餘疾病

並列摘要

Acute myeloid leukemia (AML) is one of the most common forms of leukemia in adults and is associated with significant mortality in Taiwan. AML is a clinically and biologically heterogeneous hematologic malignancy characterized by uncontrolled proliferation of hematopoietic precursors along with loss of their ability to differentiate. To improve quality health care and patient outcomes, incorporation of precision medicine, including risk stratification at diagnosis and treatment strategy modifications based on the treatment response, is very critical. Recently, numerous risk factors have been identified using traditional techniques. With the advancement of biotechnology methodology, the prognosis of AML is expected to be more favorable. This thesis will address these crucial issues associated with AML prognosis using the technique of next-generation sequencing (NGS), which is divided into three parts: coding, non-coding, and measurable/minimal residual disease monitoring.
Mutations in the cohesin complex genes have been reported in myeloid malignancies, but their prognostic implications and dynamic changes during the clinical course of AML remain undefined. This thesis involved a study on 391 newly diagnosed de novo non-M3 AML patients. We found that 9.5% of patients had cohesin gene mutations, most prevalent in RAD21 (3.8%) and STAG2 (3.1%) genes. These cohesion gene mutations were independent favorable factors for both overall survival (OS) and relapse-free survival (RFS), irrespective of other prognostic factors. Serial analyses showed that two patients lost the original cohesin mutations during disease evolution, while none of the patients without any cohesin gene mutation acquired a novel mutation at relapse. Gene set enrichment analysis demonstrated that cohesin gene mutations were involved in functions related to hematopoiesis, myeloid, or blood cells.
The expression of long non-coding RNAs (lncRNAs) has recently been recognized as a potential prognostic marker for AML. However, it remains unclear whether incorporation of lncRNA expression in the 2017 European LeukemiaNet (ELN) risk classification can further improve prognostic prediction. In the study described in this thesis, 275 non-M3 AML patients were enrolled and randomly assigned to the training (n=183) and validation cohorts (n=92). In the training cohort, a lncRNA scoring system comprising five lncRNAs whose expression had a significant impact on treatment outcomes was constructed. Patients with higher scores were found to have shorter OS and RFS, which was further confirmed in both internal and external validation cohorts. Multivariate analyses revealed that the lncRNA score was an independent prognostic factor in AML, irrespective of the risk based on the 2017 ELN classification.
NGS has also been used for measurable/minimal residual disease (MRD) monitoring in AML patients, but the optimal time point for the test remains unclear. In our study described in this thesis, we performed targeted NGS for 54 genes of the bone marrow cells serially obtained at diagnosis, after first complete remission (1st time point), and after the first consolidation chemotherapy (2nd time point) from 303 de novo AML patients. MRD was detected in 44.6% and 27.7% of patients at the 1st (MRD1st) and 2nd (MRD2nd) time points, respectively. Patients with detectable NGS MRD at either time point had a significantly higher cumulative incidence of relapse and shorter OS and RFS. Patients with positive MRD1st but negative MRD2nd had a similar good prognosis to those with negative MRD at both time points.
This thesis concludes that cohesin gene mutations have a favorable prognostic impact in patients with AML. These mutations may not be critical in disease progression since none of the patients with wild-type cohesins acquired any mutation during follow-ups. Incorporation of the lncRNA scoring system in the 2017 ELN classification can improve the risk stratification of patients with AML. Finally, the use of NGS MRD, especially at the 2nd time point, can help predict the clinical outcomes of patients with AML. By incorporating these prognosticators at diagnosis and follow-up stages, physicians can additionally refine the current risk stratification system and further improve quality care and patient outcomes in AML cases.

並列關鍵字

Acute Myeloid Leukemia ； Risk Stratification ； Next-generation Sequencing ； Cohesin ； Long Non-coding RNA ； Measurable/minimal Residual Disease

參考文獻

1. Access, G.B.D.H., et al., Healthcare Access and Quality Index based on mortality from causes amenable to personal health care in 195 countries and territories, 1990-2015: a novel analysis from the Global Burden of Disease Study 2015. Lancet, 2017. 390(10091): p. 231-266.

2. 衛生福利部統計處, 107年國人死因統計結果. 2018.

3. 衛生福利部國民健康署, 105年癌症登記報告. 2018.

4. Arber, D.A., et al., The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood, 2016. 127(20): p. 2391-405.

5. Löwenberg, B., Acute myeloid leukemia: the challenge of capturing disease variety. Hematology Am Soc Hematol Educ Program, 2008: p. 1-11.