研究生: |
李畤廣 Lee, Chih-Kuang |
---|---|
論文名稱: |
利用Cox時變變數模型分析癌症治療轉換之療效 Analysis the Efficacy of Cancer Treatment Switching Using Cox Time-varying Models |
指導教授: |
程毅豪
Chen, Yi-Hau 呂翠珊 Lu, Tsui-Shan |
口試委員: |
林惠文
Lin, Hui-Wen 呂翠珊 Lu, Tsui-Shan 程毅豪 Chen, Yi-Hau |
口試日期: | 2022/06/28 |
學位類別: |
碩士 Master |
系所名稱: |
數學系 Department of Mathematics |
論文出版年: | 2022 |
畢業學年度: | 110 |
語文別: | 中文 |
論文頁數: | 23 |
中文關鍵詞: | 存活分析 、治療轉換 、臨床試驗 、Cox model 、Time-varying |
英文關鍵詞: | Survival analysis, treatment switching, clinical trials, Cox model, Time-varying |
研究方法: | 次級資料分析 、 主題分析 |
DOI URL: | http://doi.org/10.6345/NTNU202200918 |
論文種類: | 學術論文 |
相關次數: | 點閱:110 下載:13 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
在過往臨床試驗之存活分析方式中,如:隨機對照試驗(RCT)、意向治療分析(ITT)及計畫族群分析(PP),皆沒有將患者轉換試驗組別的動作,然而隨著倫理意識及實踐層面等因素疊加,使治療轉換(Trestment Switching)在臨床試驗的地位日漸重要,有治療轉換發生時要如何能有效的估計試驗本身的治療效果及有效性,是學者們仍在努力的方向,亦為本篇論文的研究重點。
治療轉換的應用較常於腫瘤學實驗中出現,當患者隨機分配到臨床試驗對照組,在隨訪期間病情發生惡化或復發(統稱:疾病進展Disease progression)之時,便被允許轉換至實驗組持續治療,這時早期的ITT試驗方法會低估藥物治療效果,因此我們需透過在選模時置入特殊的影響參數,或是將權重估計的想法納入分析,才有辦法推算出其治療效果。而治療轉換亦可能由實驗組轉換至對照組,屆時ITT也會低估藥物治療效果,但現今在實際應用或計算療效尚無較完善被認可的方法,因此,本論文內容僅探討「對照組轉換至實驗組」的方式。
在本文中,資料由151名癌症患者包含治療轉換的臨床試驗所組成,其中患者大多數會有疾病復發的狀況,若有使用治療轉換便會在復發後立即轉換;少部分患者尚未觀察到復發狀況前便已死亡。我們先經過基本存活分析資料中各參數影響存活率的關係,再將不同存活分析方式應用到我們實際資料,並與我們題目主軸的方法進行比較。透過主題Cox 時變變數模型,分析治療轉換之療效,從分析結果足以顯示癌症病患使用治療轉換之療效及益處,此方法能使用的臨床資料型態較不受限制,雖未能估計出有轉換之患者若無進行轉換時原本的療效,但能有效估計含治療轉換的試驗之風險比。
從我們的研究執行及文獻探討中,皆顯示這些方法在治療轉換發生的情況下是必須做調整可以更好的。我們也持續在嘗試擷取各方之優點,延伸方法並更廣泛有效的應用在臨床試驗資料中。
Among the previous survival analysis methods of clinical trials, such as, randomized controlled trial (RCT), intention-to-treat (ITT), and Per-protocol Analysis (PP). None of them is about the analysis for the patients to switch their treatment groups. However, with the superposition of factors such as ethical awareness and practical level, the status of Treatment Switching in clinical trials has become increasingly important. How to effectively estimate the therapeutic effect and effectiveness of the trial itself when treatment switching occurs is the direction that scholars are still working on and this is also the research focus of this thesis.
The treatment switching often has application in oncology experiments, where patients are randomly assigned to a control group and allowed to switch to the experimental group when the disease worsened or recurred during follow-up (i.e.: Disease progression). At this time, the ITT test will underestimate the effect of drug treatment, so we need to insert special influence parameters in model selection, or incorporate the idea of weight estimation into the analysis, in order to calculate the treatment effe-ct. The treatment switching may also be switched from the experimental group to the control group. At that time,the ITT test may overestimate the effect of drug treatment. However, there is no more complete and recognized method for practical application or calculation of treatment effect. Therefore, the content of this thesis only discusses the method of "switching from the control group to the experimental group".
In this article, the data consisted of 151 clinical trials involving treatment switching in cancer patients, most of whom had a disease relapse condition, and patients with treatment switching who switched immediately after relapse; a minority of patients have not been observed died before relapse. Firstly, we analyzed the relationship between the parameters in the data affecting the survival rate through basic survival analysis, applied different survival analysis methods to our real data, and compared with the main method of our topic. Secondly, through the Cox time-varying variable model, the efficacy of treatment switching is analyzed, and the analysis results are sufficient to show the efficacy and benefit of cancer patients using treatment switching. The type of clinical data that can be used with this method is relatively unrestricted. Although it is not possible to estimate the original effect of switching patients without switching, it is useful to estimate the hazard ratios of trials with treatment switching.
Both our study execution and literature review suggest that these methods must be better adjusted in the event of treatment switching. We also continue trying to capture the advantages of each method, then extend the method and apply it more widely and effectively in clinical trial data.
[1] Enwu Liu, Karen Lim (2018). “Using the Weibull accelerated failure time regression model to predict time to health events”, bioRxiv 362186.
[2] Rinku Saikia, Manash Pratim Barman (2017). “A Review on Accelerated Failure Time Models”, ISSN 0973-2675 Volume 12.
[3] Latimer NR, Abrams KR (2014). “NICE DSU Technical Support Document 16: Adjusting Survival Time Estimates in the Presence of Treatment Switching”, NBK310374, PMID: 27466662. National Institute for Health and Care Excellence (NICE).
[4] Zhang Z, Reinikainen J, Adeleke KA, Pieterse ME, Groothuis-Oudshoorn CGM (2018). “Time-varying covariates and coefficients in Cox regression models”, PMID: 29955581; PMCID: PMC6015946.
[5] Yuqing Xu, Meijing Wu, Weili He, Qiming Liao, Yabing Mai (2021). “Teasing out the overall survival benefit with adjustment for treatment switching to other therapies.” Department of Statistics, University of Wisconsin-Madison, Madison, WI 53705, USA.
[6] Nicholas R. Latimer, hris Henshall, Uwe Siebert, Helen Bell (2016). “TREATMENT SWITCHING: STATISTICAL AND DECISION-MAKING CHALLENGES AND APPROACHES.” nternational Journal of Technology Assessment in Health Care, 32:3 (2016), 160–166
[7] Nicholas R Latimer, Keith R Abrams, Paul C Lambert, James P Morden and Michael J Crowther (2018). “Assessing methods for dealing with treatment switching in clinical trials: A follow-up simulation.” Statistical Methods in Medical Research 2018, Vol. 27(3) 765–784
[8] Thomas R. Sullivan, PhD, Nicholas R. Latimer, PhD, Jodi Gray, MPH, Michael J. Sorich, PhD, Amy B. Salter, PhD, Jonathan Karnon, PhD (2019). “Adjusting for Treatment Switching in Oncology Trials: A Systematic Review and Recommendations for Reporting.” VALUE HEALTH. 2020; 23(3):388–396
[9] N. R. Latimer, K. R. Abrams, U. Siebert (2019). “Two-stage estimation to adjust for treatment switching in randomised trials: a simulation study investigating the use of inverse probability weighting instead of re-censoring.” Latimer et al. BMC Medical Research Methodology (2019) 19:69.
[10] Annabel Allison, Ian R White, Simon Bond (2018). “rpsftm: An R Package for Rank Preserving Structural Failure Time Models.” Author manuscript; available in PMC. R J. ; 9(2): 342–353.
[11] Marshall M. Joffe, Wei Peter Yang and Harold Feldman (2012). “G-Estimation and Artificial Censoring: Problems, Challenges, and Applications.” Biometrics 68, 275-286.
[12] Nathalie Grafféo, Aurélien Latouche, Christophe Le Tourneau, Sylvie Chevret (2019). “ipcwswitch: An R package for inverse probability of censoring weighting with an application to switches in clinical trials.” Computers in Biology and Medicine.Volume 111, 103339. 0010-4825/ © 2019 Elsevier Ltd.
[13] Kaufman M, Mehrotra B, Limaye S, White S, Fuchs A, Lebowicz Y, Nissel-Horowitz S, Thomas A. “EGFR Expression in Gallbladder Carcinoma in North America.” Int J Med Sci 2008; 5(5):285-291