陳為堅臺灣大學：流行病學研究所陳璿宇Chen, Hsuan-YuHsuan-YuChen2007-11-272018-06-292007-11-272018-06-292007http://ntur.lib.ntu.edu.tw//handle/246246/56148現行臨床分期系統無法準確地預測非小細胞肺癌病人的存活與復發情形，本研究利用微陣列技術以及即時反轉錄聚合酶連鎖反應兩種測量基因表現的方法，探討基因表現剖繪以及微小核醣核酸和非小細胞肺癌病人臨床結果的關係。本研究主要為兩部份，第一部份為利用基因表現預測非小細胞肺癌病人存活與復發，第二部份為微小核醣核酸 (microRNA)在非小細胞肺癌的預後上的應用。 研究一、利用基因表現剖繪預測非小細胞肺癌的臨床結果 背景：利用基因表現預測非小細胞肺癌病人的存活準確性較目前的臨床分期方法佳，在現有研究的結果中，由於需利用幾十或上百個基因才能達到準確預測的目的。本研究擬建立只利用少數基因表現的預測方法，將其應用至非小細胞肺癌病人的預後。方法：利用微陣列晶片 (microarray)以及即時反轉錄聚合酶連鎖 (real time RT-PCR)反應方法測量125個臨床肺癌檢體的基因表現，以及進行基因表現與存活資料的相關分析。最後利用風險分數 (risk score)與決策樹 (decision tree)分析方法建立基因表現預測模式，以此模式預測肺癌病人的存活與復發。除了原來的125檢體外，另外將此預測模式套用至另外的60個肺癌檢體以及來自國外已發表的資料中86個肺癌病人資料，以驗證此預測模式。結果：利用微陣列晶片與風險分數方法找出與臨床存活有關的16個基因，其中5個基因在RT-PCR與微陣列實驗結果相近，這5個基因分別為(DUSP6、MMD、STAT1、 ERBB3以及LCK)。接著利用此五個基因建立決策樹分析模式，以此模式能預測高風險的病人校正後的風險對比值為2.82倍 (95% 信賴區間: 1.38-5.78)。在另外的60個檢體的驗證資料中，預測的高風險的病人其風險對比值為3.36倍 (95% 信賴區間: 1.35-8.35)，在國外發表的資料中也有得到顯著的結果(風險對比值4.36倍，95% 信賴區間: 1.01-18.76)。結論：以此五個基因表現建構的預測模式，能準確地預測非小細胞肺癌病人的存活。 研究二、微小核甘核酸與非小細胞肺癌的臨床結果 背景：微小核甘核酸為一群新發現的核甘核酸，其不會轉譯成蛋白質而且能負向調構基因表現。本研究擬建立利用微小核甘核酸表現的預測方法，探討其與非小細胞肺癌病人預後的相關。方法：利用即時反轉錄聚合酶連鎖 (real time RT-PCR)反應方法測量112個臨床肺癌檢體的微小核甘核酸表現，以及進行基因表現與存活資料的相關分析。最後利用風險分數分析方法建立微小核甘核酸表現預測模式，以此模式預測肺癌病人的存活與復發。最後利用另一家醫學中心的62個肺癌檢體，驗證此風險分數預測模式。結果：5個微小核甘核酸與病人存活有關，分別為(hsa-let-7a、 hsa-miR-221、hsa-miR-137、hsa-miR-372和hsa-miR-182*)，利用此五個基因建構風險分數預測模式，以此模式預測肺癌病人的存活與復發的風險。以病人死亡情形來說，在原始資料中的訓練組56個病人，高風險的病人的風險對比值為10.31倍 (95% 信賴區間: 2.33-45.56)；驗證組56個病人中高風險的病人的校正風險對比值為3.65倍 (95% 信賴區間: 1.29-10.37)；在另一家醫學中心的62個檢體的驗證資料中，預測的高風險的病人其校正風險對比值為2.81倍 (95% 信賴區間: 1.13-7.01)。以病人復發情形來說，在原始的訓練組56個病人中，高風險的病人的校正風險對比值為3.29倍 (95% 信賴區間: 1.24-8.71)；驗證組56個病人中高風險的病人的校正風險對比值為2.86倍 (95% 信賴區間: 1.20-6.82)；在另一家醫學中心的62個檢體的驗證資料中，預測的高風險的病人其校正風險對比值為2.39倍 (95% 信賴區間: 1.12-5.10)。分層分析的結果顯示，不論是一期、二期、三期肺癌，或是肺癌癌與鱗狀細胞癌等不同細胞形態等分層中，此五微小核甘核酸預測模式能準確預測病人的存活與復發。結論：五個微小核甘核酸建構的預測模式，能準確地預測非小細胞肺癌病人的存活與復發，並能提供肺癌病人後續治療的參考。 本研究利用基因印記預測病人的存活與復發，包括五個基因表現所建構的模式以及五個微小核醣核酸的預測模式，皆能準確預測肺癌病人的存活與復發。未來需要前瞻性的大規模追蹤試驗，用足夠的樣本數目檢驗本研究的發現。Current clinical staging system can not accurately predict patients’ outcome. In this study, microarray technology and real time RT-PCR were carried out to assay the gene expressions, including gene expression profile and microRNA signature. This disser- tation included two studies of non-small-cell lung cancer. One is “Gene Expression Signature Predicts Clinical Outcomes in NSCLC” and another is “microRNA Expre- ssion Profile and Clinical Outcomes in NSCLC”. Study 1: Gene Expression Signature Predicts Clinical Outcomes in NSCLC Background: Using molecular profiling approach to predict patients’ outcome is better than current staging method. of non-small cell lung carcinoma (NSCLC). In this study, a model based on few number of gene will be established and predicted survival in NSCLC. Methods: Gene expression in surgical specimens of 125 samples of surgi- cally resected NSCLC was studied by microarray and real-time reverse transcriptase polymerase chain reaction (RT-PCR), and the results were compared with survival. We used the risk score and decision tree methods to develop a gene-expression model to predict the outcome of NSCLC. The results were validated in an independent cohort from 60 patients and published dataset from 86 samples. Results: Sixteen genes that correlated with survival in patients with NSCLC were identified using microarray and risk score analysis. We selected 5 genes (DUSP6, MMD, STAT1, ERBB3 and LCK) and developed a risk predictive model based on RT-PCR and a decision tree analysis. The 5-gene signature is an independent predictor of cancer recurrence and overall survival of NSCLC patients (hazard ratio [HR] =2.82, 95% CI= 1.38-5.78). We validated the model in an independent cohort of 60 NSCLC patients (HR=3.36, 95% CI= 1.35-8.35) and in a set of published microarray data of 86 patients (HR=4.36, 95% CI= 1.01-18.76). Conclusions: A 5-gene signature can predict survival and relapse of NSCLC patients. Study 2: microRNA Expression Profile and Clinical Outcomes in NSCLC Background: MicroRNAs are a new class of small non-protein-coding RNAs that function as endogenous negative gene-regulators and can act as oncogenes or tumor- suppressors. An microRNA signature will be developed and significantly associated with survival of NSCLC patients. Methods: Using real-time reverse transcriptase polymerase chain reaction (RT-PCR), we studied microRNA expression in tumor-specimens of 112 patients who had undergone surgical resection of NSCLC. Results were correlated with patients’ survival. We used Cox regression and risk-score analysis to develop a microRNA signature for the prediction of treatment outcome of NSCLC. Results: We identified a 5- microRNA signature (hsa-let-7a, hsa-miR-221, hsa-miR-137, hsa-miR-372 and hsa-miR -182*) associated with survival of 56 NSCLC patients each in the training and testing sets. We reconfirmed the findings in an independent cohort of 62 NSCLC patients. NSCLC pa- tients with high-expression of the 5-microRNA signature had reduced overall survival (adjusted HR=2.81, 95%CI=1.13-7.01, p=0.026) and disease-free survival (adjusted HR= 2.39, 95%CI=1.12-5.10, p=0.024) compared to low-expression patients, even after stratify- ing patients by stage I, II, III, adenocarcinoma or squamous cell carcinoma subgroups. The 5-microRNA signature was more effective to predict survival of NSCLC patients compared to less-than-five or single-microRNA signatures (p<0.05, log-rank tests).Conclusions: An unique microRNA signature can predict cancer relapse and survival of NSCLC patients. MicroRNAs may have implications in molecular-pathogenesis of NSCLC, selection of high-risk patients for adjuvant chemotherapy or development of new targeted-therapy for NSCLC. In conclusion, our results indicate that using a gene signature composed of relatively small number of genes, either a five-gene or 5-microRNA signature, can predict the recurrence as well as overall survival of NSCLC and further validation of these findings in a prospective cohort of large sample size is warranted.中文摘要 …………………………………………………………… I ABSTACT …………………………………………………………… IV LIST OF FIGURES …………………………………………………………… IX LIST OF TABLES …………………………………………………………… X CHAPTER 1 Introduction……………………………………………… 1 CHAPTER 2 Study 1: Gene Expression Signature Predicts Clinical Outcomes in NSCLC ………………………………. 5 2.1 Introduction …………………………………………………………… 5 2.2 Methods …………………………………………………………… 6 2.2.1 Patients and Tissue Specimens …………………………. 6 2.2.2 cDNA Microarray Analysis ……………………………... 6 2.2.3 Real-Time RT-PCR Analysis …………………………… 7 2.2.4 Statistical Analysis ……………………………………… 7 2.3 Results …………………………………………………………… 11 2.3.1 A 16-Gene Microarray Signature and Survival and Metastasis in NSCLC …………………………………… 11 2.3.2 5-Gene Signature by Real-Time RT-PCR Correlates with Survivals of 101 NSCLC Patients ……………………… 12 2.3.3 Validation of the 5-Gene Signature ……………………... 13 2.3.4 Validation of the 5-Gene Signature in a Set of published Microarray Data ………………………………………… 14 2.4 Discussion …………………………………………………………… 14 CHAPTER 3 Study 2: microRNA Expression Profile and Clinical Outcomes in NSCLC ………………………………….. 27 3.1 Introduction …………………………………………………………… 27 3.2 Methods …………………………………………………………… 28 3.2.1 Patients and Tissue Specimens …………………………. 28 3.2.2 MicroRNA Profiling ……………………………………. 28 3.2.3 Statistical Analysis ……………………………………… 29 3.3 Results …………………………………………………………… 30 3.4 Discussion …………………………………………………………… 32 CHAPTER 4 Conclusion and Future Perspective ……………………... 49 REFERENCES …………………………………………………………… 51 APPENDIX …………………………………………………………… 62en-US非小細胞肺癌基因印記微小核甘核酸微陣列晶片即時反轉錄聚合酶連鎖反應風險分數決策樹non-small-cell lung cancergene signaturemicroRNAmicroarrayreverse- transcription polymerase chain reaction (RT-PCR)risk scoredecision tree[SDGs]SDG3thesis