怎麼樣讀懂一個典型的臨床試驗？

本文以解讀該文章為例進行說明：The Lancet Oncology. 2014. Afatinib versus cisplatin plus gemcitabine for first-linetreatment of Asian patients with advanced non-small-celllung cancer harbouring EGFR mutations (LUX-Lung 6): an open-label, randomised phase 3 trial.

怎麼樣最快速地簡練地提取臨床試驗的信息？

每一年全世界腫瘤學界大會（ASCO，ESMO等）都會產生許多學術進展，重要臨床試驗也會公布階段性結果。快速提取臨床試驗信息最直接明了的方法是：看Abstract部分。如果你想更快速，直接看Conclusions的結論部分。例文中的Interpretation部分相當於Conclusions部分：

First-line afatinib significantly improves progression-free survival with a tolerable and manageablesafety profile in Asian patients with EGFR mutation-positive advanced lung NSCLC. Afatinib should be consideredas a first-line treatment option for this patient population.

從Interpretation/Conclusions部分，我們可以粗略地得出這樣的印象：Afatinib（阿法替尼）在EGFR突變的NSCLC患者中效果不錯，一線治療可以考慮。這對於想非常粗略地了解該臨床試驗的人還是不錯的，畢竟我們通過一句話的閱讀時間就明確了該試驗的核心內容。

有時候我們非常想知道該臨床試驗的背景，為什麼要做這個臨床試驗呢？有何意義？這時候請看Abstract的Background部分。例文中Summary的Background部分：

Afatinib—an oral irreversible ErbB family blocker—improves progression-free survival compared withpemetrexed and cisplatin for first-line treatment of patients with EGFR mutation-positive advanced non-small-celllung cancer (NSCLC). We compared afatinib with gemcitabine and cisplatin—a chemotherapy regimen widely usedin Asia—for first-line treatment of Asian patients with EGFR mutation-positive advanced NSCLC.

從Background部分，我們知道：已有研究表明，在EGFR突變的NSCLC中，Afatinib（阿法替尼）一線治療好於培美曲塞 + 順鉑；本試驗是想看一看Afatinib（阿法替尼）是否也好於吉西他濱+順鉑。這樣一來，我們就對該臨床試驗的目的非常明確了。通過這段話，我可以推斷培美曲塞+順鉑，或者吉西他濱+順鉑，這兩種化療方案是治療晚期NSCLC目前的經典方案，因為只有和它們做對比才有意義。

此時，我們知道了該臨床試驗的結論與目的，那麼它是怎麼樣設計的呢？試驗可信度有多少呢？參加臨床試驗的患者都有多少呢？他們的代表性具有廣泛意義嗎？是所有晚期NSCLC的患者使用阿法替尼效果都要好於化療方案嗎，還是只是具有某一特徵類型的患者？要明確這些問題，就需要讀一讀Abstract的Methods部分。

This open-label, randomised phase 3 trial was done at 36 centres in China, Thailand, and South Korea.After central testing for EGFR mutations, treatment-naive patients (stage IIIB or IV cancer [American JointCommittee on Cancer version 6], performance status 0–1) were randomly assigned (2:1) to receive either oral afatinib(40 mg per day) or intravenous gemcitabine 1000 mg/m2 on day 1 and day 8 plus cisplatin 75 mg/m2 on day 1 of a3-week schedule for up to six cycles. Randomisation was done centrally with a random number-generating systemand an interactive internet and voice-response system. Randomisation was stratified by EGFR mutation (Leu858Arg,exon 19 deletions, or other; block size three). Clinicians and patients were not masked to treatment assignment, butthe independent central imaging review group were. Treatment continued until disease progression, intolerabletoxic effects, or withdrawal of consent. The primary endpoint was progression-free survival assessed by independentcentral review (intention-to-treat population). This study is registered with ClinicalTrials.gov, NCT01121393.

從Methods部分，我們可以了解很多細節，其實大多數情況下我們不大需要了解這麼多細節，只需選取自己認為重要的，或者對自己有用的信息就可以了。1，患者具有廣泛代表性，中國、泰國和韓國的36各中心；2，入組患者的分期，IIIB或者IV期；3，兩小組對比，阿法替尼 Vs 吉西他濱+順鉑；（用藥方式就是比較細節的事情了，如果不是醫生我認為沒太多必要關注這些細節，因為試驗設計自然而然是比較標準的用藥方案）；4，隨機化保障，這部分內容也沒太多必要去關注，如果隨機化這麼基礎的事情都不能保證那還做什麼臨床試驗呢？5，分子分層要關注，EGFR L858R，EGFR 19del，Others，這種分子分層往往都代表了會進行亞組分析，所以需要關注，而且通過該特徵我們能明確到底是哪些EGFR突變患者受益；6，醫生或者患者告知情況，這部分我們可以適當關注，畢竟他們知道用什麼葯其實心理作用可能會有較大影響的，醫生患者都知道自己在幹嘛。7，臨床試驗中止條件，這部分不是特別重要，因為它對結論影響不大，大多數臨床試驗中止的條件大同小異，疾病進展、無法耐受的毒性作用、患者退出臨床試驗。8，主要終點，這部分是比較重要的，它直接決定了評價藥物效果指標，就如評價深圳與成都遠不遠我們可以選擇寶安機場到雙流機場的飛行時間、深圳到寶安直線距離等。主要重點是根據Background的目的而直接設定的，有了它我們才可以達到目的。Progression-free survival (PFS)為例文的主要終點，它是指：開始用阿法替尼或者化療藥方案到腫瘤進展（影像學進展，國際標準評估）或者死亡的時間。這時我們可以推測，用阿法替尼的PFS肯定要大於用化療葯的。

現在我們知道該臨床試驗的結論是什麼，明白為什麼會做這樣的臨床試驗，也清楚臨床試驗的設計是怎麼樣的。那麼這兩種方案對患者影響到底有多大呢？程度是怎麼樣的？何以見得此種方案比彼種方案好呢？這時候我們就要看Abstract的Results部分了，例文中的Findings部分：

910 patients were screened and 364 were randomly assigned (242 to afatinib, 122 to gemcitabine andcisplatin). Median progression-free survival was significantly longer in the afatinib group (11·0 months, 95% CI9·7–13·7) than in the gemcitabine and cisplatin group (5·6 months, 5·1–6·7; hazard ratio 0·28, 95% CI 0·20–0·39;p<0·0001). The most common treatment-related grade 3 or 4 adverse events in the afatinib group were rash or acne(35 [14·6%] of 239 patients), diarrhoea (13 [5·4%]), and stomatitis or mucositis (13 [5·4%]), compared withneutropenia (30 [26·5%] of 113 patients), vomiting (22 [19·5%]), and leucopenia (17 [15·0%]) in the gemcitabine andcisplatin group. Treatment-related serious adverse events occurred in 15 (6·3%) patients in the afatinib group andnine (8·0%) patients in the gemcitabine and cisplatin group.

從Results部分，我們可以了解到結果的詳細信息。1，364個患者經過篩選後入組到兩個小組，242 to 阿法替尼，122 to 吉西他濱+順鉑。2，阿法替尼 Vs 吉西他濱+順鉑 （mPFS：11.0 months Vs 5.6 months）;這個信息應該是整個臨床試驗的核心，因為它從數據上告訴我們阿法替尼好，而且告訴我們好的程度，中位無進展生存期是一個非常有意義的指標。一般來說，我們看11.0 months與5.6 months這兩個數據就夠了，但是其他數據也有一定補充意義，尤其是公開展開的時候需要把它們寫出來以顯示科學性；95% CI，表明這個數據95%程度的靠譜；（9.7 - 13.7，5.1-6.7）表明了兩個小組PFS的具體範圍；Hazard Ratio （HR），風險比，這個可以理解為兩個比率之比，今天與公司同事討論，勿需理解太透徹，0.28的HR（HR = 1，說明兩組無明顯差異），HR為0.28說明有差異（Favors Afatinib or Chemotherapy regimen）；P value，P值，一般P值小於0.05就表明兩組的差異具有顯著性，若P值小於0.0001則說明差異非常顯著；3，副作用，副作用是非常重要的考慮指標，mPFS雖然是非常重要的考慮指標，但是副作用也是為患者苦惱，有時候副作用太強而不得不停止用藥。常見3或4級副作用對比：阿法替尼組（皮疹痤瘡14.6%，腹瀉5.4%，口腔炎黏膜炎5.4%），吉西他濱+順鉑組（嗜中性白血球減少症26.5%，嘔吐19.5%，白細胞減少15.0%）。副作用比較3級或者4級是非常有必要的，因為這才嚴重地足以停葯而影響治療。從（14.6%，5.4%，5.4% Vs 26.5%，19.5%，15.0%）這足以說明阿法替尼的副作用比例明顯少於化療組。最後再次總結出：Treatment-related serious adverse event （治療相關嚴重副作用）在阿法替尼組中為6.3%，在吉西他濱+順鉑組為8.0%。

至此，我們僅通過Abstract就明白了臨床試驗的目的、臨床試驗的設計方法、臨床試驗的結果、臨床試驗的結論。但是需要簡單明了的用PPT展示出來，還需對上述Abstract進行簡單整理與擴充。以下便用7張PPT來展示整個臨床試驗。

a 第一張PPT 題目與臨床編號

LUX-Lung 6 - Afatinib* vs Cisplatin + Gemcitabine in Asian Patients with Epidermal Growth Factor Receptor Mutation-positive Advanced Non-small Cell Lung Cancer

A randomised, open-label, Phase III study of afatinib vs chemotherapy as first-line treatment for patients with Stage IIIB or IV adenocarcinoma of the lung harboring an epidermal growth factor receptor (EGFR) activating mutation

Trial CTgov-Identifier: NCT01121393

b 第二張PPT 患者入組標準

Patients：1, Adenocarcinoma of the lung; 2, EGFR mutation positive; 3, Stage IIIB/IV; 4, No prior treatment with chemotherapy for advanced/metastatic disease; 5, No prior treatment with EGFR inhibitors; 6, Eastern Cooperative Oncology Group performance status 0-1;

N = 364;

c 第三張PPT 患者分組

Randomisation 2 : 1

Afatinib 40mg Oral once daily

Cisplatin + Gemcitabine: 75mg/㎡ + 1000mg/㎡, Intravenous, d1+d8, once every 3 weeks up to 6 cycles

d 第四張PPT 終點

Primary Outcome Measures: Progression-free survival (PFS), assessed by an independent central radiology review

Secondary Outcome Measures: Objective response rate, Disease control rate, Overall survival (OS), Time to and duration of objective response, Duration of disease control, Health-related quality of life (HRQoL), Pharmacokinetics

e 第五張PPT 結果

PFS, OS, HRQoL, etc; 配圖並以文字說明；

f 第六張PPT 結論

First-line afatinib for treatment of patients with advanced adenocarcinoma significantly (p<0.0001) improved PFS, vs treatment with cisplatin/gemcitabine chemotherapy. In addition, patients with EGFR del19 mutation showed >1 year overall survival.

Afatinib showed improved efficacy and clinical benefit compared with cisplatin/gemcitabine chemotherapy in patients with previously untreated advanced adenocarcinoma of the lung and EGFR mutations. Afatinib was also associated with better control of lung cancer-related symptoms and improvement in HRQoL.

g 第七張PPT 參考文獻

Wu YL, Zhou C, Hu CP, et al. Afatinib versus cisplatin plus gemcitabine for first-line treatment of Asian patients with advanced non-small-cell lung cancer harbouring EGFR mutations (LUX-Lung 6): an open-label, randomized phase 3 trial. Lancet Oncol 2014;15(2):213-222.

Geater SL, Xu CR, Zhou C, et al. Symptom and Quality of Life Improvement in LUX-Lung 6: An Open-Label Phase III Study of Afatinib Versus Cisplatin/Gemcitabine in Asian Patients With EGFR Mutation-Positive Advanced Non-small-cell Lung Cancer. J Thorac Oncol. 2005 Jun;10(6):883-889.

Yang JC, Wu YL, Schuler M, et al. Afatinib versus cisplatin-based chemotherapy for EGFR mutation-positive lung adenocarcinoma (LUX-Lung 3 and LUX-Lung 6): analysis of overall survival data from two randomized, phase 3 trials. Lancet Oncol 2015;16(2):141-151.

Clinicaltrials.gov. https://clinicaltrials.gov/ct2/show/study/NCT01121393 (Accessed: May 16).

以上七張PPT簡潔精鍊地描述了該臨床試驗的來龍去脈，在描述其它相似臨床試驗時可借鑒這七張PPT的描述方法。

通過Abstract的Conclusions, Background, Methods與Results，我們大致了解了該臨床試驗的來龍去脈，通過七張PPT的展示內容，我們較詳細的理順了該臨床試驗的主脈絡。通常來說，一般人了解到這樣的程度就可以了，就算是你給癌症研究者簡單地講解該臨床試驗也足夠了。那麼對於深入詳細了解該臨床試驗，那麼以上的總結還是不夠的，需要對臨床試驗的每個環節的細節進行深入閱讀與分析。那麼一下內容就是對該臨床試驗進行剖析。

Introduction

1，化療方案：順鉑+吉西他濱，順鉑+多西他賽，卡鉑+紫杉醇，順鉑+紫杉醇，治療NSCLC有8-10月的中位生存獲益；

In the past, four chemotherapeutic regimens—cisplatinand gemcitabine, cisplatin and docetaxel, carboplatinand paclitaxel, and cisplatin and paclitaxel—have beenused for treatment of advanced non-small-cell lungcancer (NSCLC), affording a median overall survivalof around 8–10 months.

2，EGFR TKI：EGFR TKI 對EGFR陽性突變患者有2年的中位生存獲益；

More recently, theidentification of lung tumours harbouring mutations inEGFRhasledtoafocusontargetedtreatments—EGFRtyrosine kinase inhibitors4—resulting in median overallsurvival of more than 2 years for patients with EGFRmutation-positive NSCLC.5 However, despite theseadvances, there remains considerable room forimprovement.

3，Afatinib（阿法替尼）：阿法替尼作用機理簡介，阿法替尼對比吉非替尼與厄洛替尼強效，阿法替尼與可逆EGFR TIK在細胞系與異種移植模型上的抗癌活性對比。

Afatinib is a novel, irreversible ErbB family blockerthat selectively and potently blocks signalling from ErbBfamily receptors (EGFR, HER2 [ErbB2], and ErbB4)6 andtransphosphorylation of ErbB3.7 Unlike reversible EGFRtyrosine kinase inhibitors (erlotinib and gefitinib),afatinib covalently binds to proteins of the ErbB receptornetwork, and irreversibly and completely abrogatessignalling, which causes a sustained and broad-spectrumanti-mitogenic activity. In preclinical studies, afatinibwas highly potent, with 50% inhibitory concentrations of0.5 nmol/L for EGFR, 14 nmol/L for HER2, and 1 nmol/Lfor ErbB4,6–8 compared with 0.1 μmol/L for gefitinibagainst EGFR,5 and 2 nmol/L for erlotinib against EGFR.9Afatinib has also shown greater anticancer activity than have reversible EGFR tyrosine kinase inhibitors, both inEGFR tyrosine kinase inhibitor-sensitive and inhibitor-resistant cell lines and xenograft models of NSCLC.6

4，已有臨床證據：EGFR TKI效果優於化療方案；介紹重要的臨床試驗內容LUX-Lung 3，闡述該臨床試驗的結論：阿法替尼優於化療方案。

Several randomised studies10–15 support the use ofEGFR tyrosine kinase inhibitors as the standard first-linetreatment for patients with activating EGFR mutations,showing high tumour response rates and longprogression-free survival compared with chemotherapy.Most of these trials were done in Asian populationsbecause EGFR mutations are more common in Asianpatients (47%) with lung adenocarcinoma than in non-Asian patients (13–15%).16 LUX-Lung 3 was the firstglobal trial to compare an irreversible ErbB familyblocker (afatinib) with chemotherapy and the first to usethe recently established best-in-class chemotherapytreatment—pemetrexed and cisplatin—as a comparator.17The investigators reported that patients taking afatinibhad significantly longer progression-free survival thanpatients taking the chemotherapy regimen.

5，吉西他濱+順鉑介紹：化療方案臨床應用地位介紹，吉西他濱與順鉑廣泛使用，在亞洲國家為批准的一線化療方案。

As a companion trial to LUX-Lung 3, we did LUX-Lung 6 tocompare afatinib with gemcitabine and cisplatin in Asianpatients. Gemcitabine and cisplatin is a widely used andapproved first-line chemotherapeutic regimen in Asiancountries (eg, China) where pemetrexed and cisplatinhas not been approved for first-line treatment of NSCLC.

Methods

1，試驗設計與患者：這部分內容為Abstract中Methods部分的擴充，大部分內容對於閱讀來說意義不大，就如同行之間的一個默認基礎，我們在這默認基礎之上談論事情，默認你這些小規範與細節都沒有問題。這部分內容包括：患者選取標準、腫瘤評價標準（RECIST）、EGFR基因突變檢測方法/試劑盒、患者同意書、倫理法規合宜等。對於已閱讀Abstract的讀者來說，這部分僅需稍微留意基因突變檢測試劑盒可檢測的位點：L858R，19del，其它。

We did this randomised, open-label, phase 3 trial at36 centres in China, Thailand, and South Korea. Eligiblepatients had pathologically confirmed and previouslyuntreated stage IIIB (with pleural effusion) or IV lungadenocarcinoma according to American JointCommittee on Cancer criteria,18 an Eastern CooperativeOncology Group (ECOG) performance status of 0 or 1,measurable disease according to Response EvaluationCriteria in Solid Tumors version 1.1 (RECIST),19 andadequate organ function. Tumour tissue had to beEGFR mutation-positive at the screening stage, asassessed at a central laboratory with a validated test kit(Therascreen EGFR 29; Qiagen, Manchester, UK). Thetest enabled us to identify 29 mutations, includingcommon (Leu858Arg, exon 19 deletions) and othermutations (see appendix for full eligibility criteria andEGFR mutations tested).

All patients provided written, informed consent forparticipation in the study and provision of tumoursamples. An independent data and safety monitoringcommittee monitored safety throughout the trial. Thestudy was done in accordance with the Declaration ofHelsinki, International Conference on Harmonisationgood clinical practice, local laws, and applicableregulatory requirements. The study was approved by theinstitutional review board or independent ethicscommittee of each centre.

2，隨機和盲法：這部分內容其實對理解這個臨床試驗意義不大，無非就是保證試驗設計的科學性，患者分組具有隨機性；盲法，醫生與患者有時可能不知道自己用的什麼方案，這樣保證心理作用不受影響，而有的時候二者有一個知道方案或者兩個都知道，總而言之這也是實驗設計科學性的重要一環；除此之外，第三方獨立保證機構對試驗的監督也是在這部分說明。

Eligible patients were randomly allocated to receiveafatinib or gemcitabine and cisplatin (2:1), stratified byEGFR mutation (Leu858Arg, exon 19 deletions, or other).A block size of three was used and randomisation wasdone centrally with a validated random number-generating system at Boehringer Ingelheim, verified by atrial-independent statistician, and implemented centrallyvia an interactive internet and voice-response system.Access to the randomisation code was supervised by theclinical trial support group; those directly involved in theconduct and analysis of the trial had no access to therandomisation schedule.

Clinicians and patients were not masked to treatmentassignment. The study investigators who did assessmentsof patient-reported outcomes and safety, along withsupportive assessments of tumour response (used forsensitivity analyses), were not masked to treatmentassignment. The independent central imaging reviewgroup who assessed tumour response (used for primary and key secondary efficacy analyses) were masked totreatment assignment. During the study, employees ofthe sponsor were masked to treatment assignment untilthe database was locked and ready for statistical analyses.

3，臨床試驗程序/過程：這部分交代了用藥方案、定期監測方案、預期不同情況的特殊處理及各標準。

Pre-treatment testing of fresh or archived tumoursamples for EGFR mutations was done by standardisedallele-specific quantitative real-time PCR at centrallaboratories. Patients with inadequate tumour tissuewere not entered into the screening phase. Patientsreceived either oral continuous afatinib (40 mg per day)or intravenous gemcitabine (1000 mg/m2, on day 1 andday 8) plus cisplatin (75 mg/m2, on day 1), in a 3-weekschedule until disease progression, intolerable toxiceffects, or withdrawal of consent. Gemcitabine andcisplatin was given for a maximum of six cycles.

Adverse events were assessed throughout the studyand documented with the National Cancer InstituteCommon Terminology Criteria for Adverse Eventsversion 3.0. Severity of all adverse events was graded withthese criteria. Regular physical (including assessment ofsymptoms) and laboratory assessments, as well as 12-leadelectrocardiograms and echocardiogram or multigatedacquisition scans, were also used to monitor safety.

Patients treated with afatinib 40 mg per day could havetheir dose increased to 50 mg per day from the secondcycle to account for interpatient variability in afatinibexposure and to tailor dosing to individual tolerability.Dose escalation to 50 mg per day was allowed in theabsence of predefined levels of toxic effects—ie, rash,diarrhoea, mucositis, or any other treatment-relatedadverse event greater than grade 1 in the first 21 days oftreatment. As per protocol, if the patient had anygrade 3 or higher treatment-related adverse event,prolonged grade 2 diarrhoea (≥48 h), grade 2 nausea orvomiting for 7 days or more consecutively despiteappropriate supportive care, or grade 2 or moreworsening renal function, afatinib was withheld for up to14 days until the severity fell to grade 1 or less or tobaseline levels. Afatinib could then be resumed at a lowerdose (10 mg reductions to a minimum dose of 20 mg).

In the gemcitabine and cisplatin group, patientsreceived six treatment courses unless they had diseaseprogression or unacceptable adverse events, or if thepatient or investigator requested permanent dis-continuation of study drug. For patients who had adverseevents related to gemcitabine and cisplatin, treatmentwas delayed or the dose was reduced (by 50% for non-haematological toxic effects or 75% for haematologicaltoxic effects as judged by the treating physicians) on thebasis of the patient』s tolerability and abnormal laboratorymeasurements, in accordance with the guidance in thecurrent summary of product characteristics andinstitutional guidelines.

Tumours were assessed by CT scan or MRI every6 weeks for the first 48 weeks, then subsequently every 12 weeks until objective disease progression or start offurther cancer treatment. Brain imaging and bone scanswere done if clinically indicated. All scans were reviewedby an independent central imaging review group thatconsisted of radiologists and oncologists.

Patient-reported outcomes were assessed atrandomisation and every 3 weeks until disease pro-gression or start of new cancer treatment with the self-administered cancer-specific European Organisation forResearch and Treatment of Cancer (EORTC) quality oflife core questionnaire QLQ-C30,20 and the lung cancer-specific module QLQ-LC13.21

4，研究終點：主要研究終點：PFS；核心次要研究終點：完全應答（CR，complete response）、部分應答（PR，partial response），疾病控制率（CR + PR + PD）；其它次要研究終點：反應持續時間（duration of response），疾病控制時間（duration of disease control），患者報告結局評價（patient-reported outcomes），安全性（safty），阿法替尼葯代動力學（pharmacokinetics of afatinib）

The primary endpoint was progression-free survival byindependent review (time from randomisation to diseaseprogression or death from any cause, whichever occurredfirst). Key secondary endpoints were the proportion ofpatients who achieved an overall response14 (ie, thepercentage of patients with complete response [CR] orpartial response [PR]), the proportion of patients whoachieved disease control (ie, the percentage of patientswith the best overall response of CR, PR, or stable disease[SD]) both by independent review, and overall survival.Other secondary endpoints included duration of response and disease control, patient-reported outcomes,safety, and pharmacokinetics of afatinib.

5，統計分析：統計方法需要科學，這部分介紹臨床試驗各個指標所使用的統計學方法，交代處理過程。

We calculated that at least 330 patients would need to beenrolled with a minimum of 217 progression-free survivalevents to detect a 57% or greater improvement inprogression-free survival with afatinib versusgemcitabine and cisplatin, with 90% power and a two-sided significance level of 5% with a log-rank test. Thiscalculation assumed a hazard ratio (HR)5 of 0·64 and anexpected progression-free survival of 11 months withafatinib and 7 months with gemcitabine and cisplatin.12–15No interim analysis was planned.

Efficacy endpoints and patient characteristics wereassessed for the intention-to-treat population, includingall randomly assigned patients. Safety was assessed forall randomised patients who received at least one dose ofstudy medication. Primary and key secondary endpointswere analysed following a hierarchical testing strategy tominimise the overall risk of type I error. Progression-freesurvival was compared between groups with a stratified(by mutation type) log-rank test. Cox proportional hazardmodels and Kaplan-Meier estimates were also used tocompare progression-free survival between treatmentgroups. Logistic regression models were used to compareobjective response rates and disease control ratesbetween treatment groups. The primary analysis ofoverall survival is planned for when the data aresufficiently mature, after roughly 237 deaths.

We did pre-specified sensitivity analyses withinvestigator assessment (rather than independentassessment) of efficacy outcomes to test the robustnessof the results. We also did pre-specified subgroupanalyses by sex, age (<65 years vs ≥65 years), EGFRmutation type (exon 19 deletions vs Leu858Arg vs others),ECOG performance status (0 vs 1), and smoking history.

Pre-specified analysis of patient-reported outcomesfocused on the NSCLC-related symptoms of cough(question 1 of QLQ-LC13), dyspnoea (questions 3–5 ofQLQ-LC13 and question 8 of QLQ-C30), and pain(questions 9 and 19 of QLQ-C30 and questions 10–12 ofQLQ-LC13). Three pre-specified analyses were done,comparing treatment groups in terms of: the distributionof patients whose symptoms had improved (≥10-pointincrease from baseline score), remained stable, orworsened (≥10-point decrease from baseline score); thetime to deterioration of symptoms; and the meandifference in symptom scores over time (longitudinalanalysis). Scales and items were scored according to the EORTC algorithm.22 For each scale or item, a lineartransformation was applied to standardise the raw scoreto a range of 0–100,22 and a 10-point change was consideredto be clinically meaningful.23 The longitudinal analysiswas done with a mixed-effects growth curve model, withthe average profile over time for each endpoint describedby a piecewise linear model (as reported previously24).Statistical analyses were done with SAS (version 9.2).

This study is registered with ClinicalTrials.gov, numberNCT01121393.

該部分涉及到的統計學工具有：時序檢驗（log-rank test），Cox比例風險模型（Cox proportional hazard model），卡普蘭-邁耶（Kaplan-Meier），EORTC algorithm，線性變換（Linear transformation），縱向分析（Longitudinal analysis）。在這部內容里，除了可以看到統計學分析方法，還可以看到做了哪些亞組分析：性別、年齡、EGFR突變狀態、ECOG PS，吸煙史。

6，經費來源：這部分都不想解釋

The sponsor provided the study drug, and wasresponsible for trial design, the collection, analysis, andinterpretation of data, and coordination of articlepreparation. The corresponding author had full access toall the data and final responsibility to submit forpublication.

Results

這部分內容一般以圖或者表進行展示，對於熟悉該領域的人只需直接看圖表就可以了，其它文字說明一般不需要去細看，除非想了解某一個方面的細節。在做展示的時候，可選取一部分重要圖表進行展示。

1，分組情況與隨訪時間交代：242：122分組到阿法替尼組與化療組；中位隨訪時間為16.6個月。

910 patients were screened between April 27, 2010, andNov 16, 2011. 364 eligible patients with EGFR mutationswere assigned to afatinib (n=242) or gemcitabine andcisplatin (n=122; figure 1). Of these patients, 352 receivedat least one dose of study drug (figure 1). Data cutoff datefor the primary analysis was Oct 29, 2012. The primaryanalysis was done after 221 progression events hadoccurred as assessed by independent review. At that time,57 (15·7%) of 364 patients (all in the afatinib group) werestill taking study treatment. Median duration of follow-up for progression-free survival was 16·6 months (IQR4·7–19·4).

2，樣本自身特徵分析：樣本特徵在兩組中均衡，但是有部分特徵例外：PS得分。

Baseline demographics and patient characteristicswere generally balanced between treatment groups(table 1), with the exception of performance score: ahigher proportion of patients had a score of 0 in thegemcitabine and cisplatin group than in the afatinibgroup (table 1). EGFR mutations were mainly exon19 deletions and Leu858Arg mutations (table 1);uncommon mutation types were not balanced betweentreatment groups (appendix).

3，隨試驗進行靈活調整：Afatinib組的患者用藥劑量調整，化療組的用藥方案周期告知。

Median duration of treatment with afatinib was398 days (IQR 173–537). At the end of treatment, 122 of182 (67·0%) patients were still receiving the starting doseof afatinib 40 mg. After the first cycle of treatment, 38 of239 (15·9%) patients in the afatinib group had their doseescalated to 50 mg per day. 67 of 239 (28·0%) patients inthe afatinib group had their dose reduced to 30 mg, andten (4·2%) had further reductions to 20 mg. Medianduration of gemcitabine and cisplatin treatment was89 days (IQR 60–119), with 40 of 113 (35·4%) patientscompleting six cycles. Overall, 62 of 101 (61·4%) patientsreceiving more than one cycle of gemcitabine andcisplatin required dose delay. The median number oftreatment cycles with gemcitabine and cisplatin was four.

4，主要研究終點結果：PFS及亞組分析。第一個主要結果在這裡給出，mPFS（11.0m vs 5.6m）； EGFR L858R突變組與EGFR 19del組的PFS無顯著差異；

Median independently assessed progression-freesurvival was 11·0 months (95% CI 9·7–13·7) in the afatinib group versus 5·6 months (5·1–6·7) in thegemcitabine and cisplatin group (HR 0·28, 95% CI0·20–0·39; p<0·0001; figure 2A). Investigator assessmentgave much the same result: median progression-freesurvival was 13·7 months (95% CI 11·5–13·9) versus5·6 months (5·1–6·8; HR 0·26, 95% CI 0·19–0·36;p<0·0001; figure 2B).

Progression-free survival was numerically longer in theafatinib group than in the gemcitabine and cisplatingroup for all subgroups; except for the two smallestsubgroups, the differences were statistically significant(figure 2C). Additionally, progression-free survival wasmuch the same for the overall population compared withpatients with the two common mutations (exon 19deletions and Leu858Arg; by independent assessment11·0 months, [95% CI 9·7–13·7] vs 5·6 months [4·5–6·2],HR 0·25, 95% CI 0·18–0·35; p<0·0001; by investigatorassessment 13·8 months [95% CI 12·5–14·4] vs5·6 months [4·7–6·7] months, HR 0·21, 95% CI0·15–0·30; p<0·0001). The appendix shows progression-free survival over time.

5，次要研究終點：OR, mDOR, DC 等的描繪。

A significantly greater proportion of patients in theafatinib group had an objective response than in thegemcitabine and cisplatin group (162 of 242 [66·9%] vs28 of 122 [23·0%]) according to independent review (oddsratio [OR] 7·28, 95% CI 4·36?12·18; p<0·0001).Investigator assessment gave much the same results (OR6·53, 95% CI 4·02?10·60; p<0·0001; table 2). By week 6,119 of 242 (49·2%) patients in the afatinib group versus16 of 122 (13·1%) in the gemcitabine and cisplatin grouphad had an objective response. Median duration ofresponse according to independent review was 9·7 months(95% CI 8·3–12·5) for afatinib and 4·3 months (2·8–5·8)for gemcitabine and cisplatin. Disease control according toindependent review was also significantly more commonin the afatinib group than in the gemcitabine and cisplatingroup (OR3·84, 95%CI 2·04?7·24; p<0·0001; table2 ),with a median duration of disease control of 11·1 months (95% CI 9·7–13·8) versus 5·7 months (5·5–6·9).Furthermore, objective response and disease control bothoccurred in much the same proportions of patients in theoverall population as in patients with common mutations (appendix). Of three patients with Thr790Met mutations(two in the afatinib group, one in the gemcitabine andcisplatin group; appendix), one patient in each group had apartial response.

6，總生存分析：此數據不成熟。

Overall survival data were immature at the time of theprimary analysis of progression-free survival: 155 of 364(42·6%) patients had died. Median overall survival was22·1 months (95% CI 20·0–not estimable) for afatinibversus 22·2 months (18·0–not estimable) for gemcitabineand cisplatin (HR 0·95, 95% CI 0·68?1·33; p=0·76;appendix). The final OS analysis will be presented oncesufficient follow-up has been achieved.

7，一些補充說明：按照臨床試驗設計，之後運用的治療方案。阿法替尼組之後使用吉西他濱+順鉑，吉西他濱+順鉑組之後使用EGFR TKI。

At the time of analysis, almost 60% of patients whodiscontinued study drug went on to receive at least onesubsequent cancer treatment (108 of 185 [58·4%] patientsin the afatinib group vs 74 of 122 [60·7%] in thegemcitabine and cisplatin group). Of these patients,101 (54·6%) in the afatinib group were subsequentlytreated with gemcitabine and cisplatin in any line oftreatment, whereas in the gemcitabine and cisplatingroup, 59 of 122 (48·4%) patients were subsequentlytreated with EGFR tyrosine kinase inhibitors. Full datafor post-progression treatment are not yet available.

8，併發症的比較：兩個小組惡化概率的比較，阿法替尼組惡化風險小於化療組。

Approximately 85% of patients who were alive andprogression-free during the study had completed andreturned patient-reported questionnaires (appendix). Agreater proportion of patients in the afatinib group thanin the gemcitabine and cisplatin group had improvementsin cough, dyspnoea, and pain (figure 3A). Time todeterioration was significantly longer in the afatinibgroup than in the gemcitabine and cisplatin group forcough, dyspnoea, and pain (figure 4). Additionally,afatinib significantly improved mean scores over timecompared with gemcitabine and cisplatin for cough(mean treatment difference –6·34, 95% CI –9·10 to–3·58; p<0·0001), dyspnoea (–9·89, –12·13 to –7·66;p<0·0001), and pain (–5·89, –8·50 to –3·27; p<0·0001).For overall health status and quality of life, a higherproportion of patients in the afatinib group than in thegemcitabine and cisplatin group had improvement(143 of 228 [62·7%] vs 33 of 101 [32·7%]; p<0·0001), hadsignificantly longer time to deterioration (HR 0·56,95% CI 0·41?0·77; p=0·0002), and had greaterimprovement in mean scores over time (mean treatmentdifference ?8·78, 95% CI ?11·19 to ?6·36; p<0·0001).Figure 3B shows self-reported outcomes for which amore than 10% difference existed in the percentage ofpatients whose symptoms improved or worsened.

9，副作用總結：治療相關副作用，皮疹痤瘡、腹瀉、疲勞等，其實是對Table 3 的一個詳盡的描述。

Table 3 summarises the most common treatment-related adverse events. The appendix shows the mostcommon adverse events irrespective of relation to studydrug. Based on maximum Common Toxicity Criteriagrade for each patient, treatment-related adverse eventsof grade 1–2 were reported by 150 of 239 (62·8%) patientsreceiving afatinib and 44 of 113 (38·9%) patients receivinggemcitabine and cisplatin. Treatment-related adverseevents of grade 3 or greater occurred in 86 of 239 (36·0%)patients receiving afatinib and 68 of 113 (60·2%)receiving gemcitabine and cisplatin. Diarrhoea, rash oracne, and stomatitis or mucositis were the most commonadverse events in the afatinib group, whereas vomiting, nausea, neutropenia, and leucopenia were the mostcommon in the gemcitabine and cisplatin group (table 3).No ECG abnormalities were reported. 14 of 239 (5·9%)patients treated with afatinib versus 45 of 113 (39·8%)treated with gemcitabine and cisplatin had treatment-related adverse events leading to permanent treatmentdiscontinuation. Treatment-related serious adverseevents were reported by 15 of 239 (6·3%) patients in theafatinib group and nine of 113 (8·0%) in the gemcitabineand cisplatin group. The most common treatment-related serious adverse events were rash or acne (three, 1·3%) and diarrhoea (two, 0·8%) in the afatinibgroup and thrombocytopenia (two, 1·8%) in thegemcitabine and cisplatin group. No patients takingafatinib permanently discontinued treatment because ofdiarrhoea only, five (2·1%) afatinib-treated patientsdiscontinued treatment because of rash or acne. In thegemcitabine and cisplatin group, the most commontreatment-related adverse events resulting in treatmentdiscontinuation were vomiting (16, 14·2%), nausea (11,9·7%), neutropenia (ten, 8·8%), and leucopenia (eight,7·1%). One patient died in each group. Both wereconsidered to be potentially treatment-related by theinvestigator (sudden death in the afatinib group andcardiac failure in the gemcitabine and cisplatin group).One patient in the afatinib group had grade 4 treatment-related interstitial pneumonitis. This male Koreanpatient permanently discontinued afatinib treatment andeventually recovered from pneumonitis after antibioticand steroid treatment. Pharmacokinetic data will bepublished separately.

Discussion

1，主要結論做一個重申，主要意義進行一個重申。

To our knowledge, this study is the largest prospective,randomised trial to compare EGFR-directed treatmentwith chemotherapy for first-line treatment of advancedEGFR mutation-positive lung adenocarcinoma (panel).The results show that afatinib significantly delayedprogression of advanced EGFR mutation-positive NSCLCcompared with gemcitabine and cisplatin in Asianpatients. Clear benefits occurred according to bothindependent and investigator review, and were consistentacross predefined subgroups. The effect on progression-free survival was substantiated by the improvement forthe secondary endpoints—objective response, diseasecontrol, and patient-reported outcomes—showing bettercontrol of lung cancer-related symptoms. These benefitswere present despite higher ECOG performance scoresin the afatinib group than in the gemcitabine andcisplatin group at baseline. At the time of analysis, overallsurvival did not differ significantly between treatmentgroups, which is unsurprising for a trial of a first-line treatment with substantial subsequent crossover betweentreatments (ie, patients in the gemcitabine and cisplatinlater receiving EGFR tyrosine kinase inhibitors and thosein the afatinib group later receiving chemotherapy).Indeed, neither of the previous studies14,25 in this settingshowed differences in survival, despite meeting theprimary endpoint of progression-free survival.

2，與相似最新研究進行對比：LUX-Lung 3；體現出聯繫與區別。

Our findings for afatinib in an Asian population arecomparable to results of the global trial—LUX-Lung 3—in which afatinib was compared with pemetrexed andcisplatin.17 Progression-free survival, objective responses,disease control, and progression-free survival over timewere similar in the overall population and in patientswith common mutations (appendix), further supportingthe robustness of our findings.17,24 The HR forprogression-free survival differed in our study comparedwith LUX-Lung 3 (HR 0·58, 95% CI 0·43–0·78), perhapsa result of differences in the efficacy of the chemotherapycomparators used in each study. LUX-Lung 6 includedpatients with uncommon mutations (not del19 orLeu858Arg), and although the population with these raremutations was small and genetically diverse, our studycontributes to the body of data already collected in othertrials of afatinib, which will be presented in the future.

3，展示研究事實，主觀分析治療相關副作用事件。

Most patients in LUX-Lung 6 had treatment-relatedadverse events. Treatment with afatinib was associatedwith the expected EGFR-mediated adverse events,including gastrointestinal 26 and dermatological dis-orders, 27 which were managed by supportive care andprotocol-defined dose reductions. Few patients dis-continued afatinib because of adverse events and nopatients discontinued afatinib because of diarrhoea. Thisresult suggests that the systematically establishedmanagement of adverse events used in this trial workedwell to keep patients on treatment, enabling themaximum benefit from afatinib. The most commonlyreported adverse events—diarrhoea, rash or acne, andstomatitis or mucositis—are the same as those reportedwith erlotinib (diarrhoea 25–81%, rash 73–85%)14,15,28 andgefitinib (diarrhoea 31–54%, rash 45–85%).10–13,29 Liverdysfunction has been reported in 8–70% of patientstreated with gefitinib11,12,29 and 6–37% of patients treatedwith erlotinib,14,15 although it was not an importantadverse event with afatinib. Cross-trial comparisons ofadverse events with different drugs have some limitationsbecause of differences in the patient populations andhow adverse events were reported. Although adverseevents were common, substantial improvements inoverall health status and quality of life compared withgemcitabine and cisplatin support the efficacy benefitsand suggest that—with proactive management—thesafety profile of afatinib was acceptable in this population.Although adverse events were rare in LUX-Lung 6 compared with LUX-Lung 3, we think that thisdifference was because physicians in LUX-Lung 6 hadmore experience in treatment of adverse events thatcommonly occur after EGFR inhibition—all patients (n=364) were recruited from only 36 sites in LUX-Lung 6,compared with LUX-Lung 3, in which a similar numberof patients (n=345) was recruited from 133 sites.

4，自我對試驗設計進行評價，優勢與劣勢。

Key strengths of this study include the use of centralstandardised EGFR mutation test and central imagingreview, and the detailed assessment of patient-reportedoutcomes. These procedures are fundamental toguaranteeing the reproducibility of the results and highquality of data and—when combined with the large samplesize—contribute to the robustness of this trial. Althoughthe open-label design is a potential source of bias, thestudy included several features to minimise this risk,including independent central review of patients』 scans byradiologistsandoncologistswhoweremaskedtotreatmentassignment and sensitivity analyses, which did not suggestany such bias. A further potential source of bias was thedifferent performance score in each group at baseline;however, with more patients having a score of 1 in theafatinibgroupthaninthegemcitabineandcisplatingroup,any resulting bias should favour the comparator.

5，自我評價

One limitation of LUX-Lung 6 is the use of gemcitabineand cisplatin as a control treatment. The treatment foradvanced lung adenocarcinoma in patients with EGFRmutation has developed quickly since this study wasstarted, with erlotinib and gefitinib becoming the firstchoice for EGFR mutation-positive patients. Indeed,recruitment for LUX-Lung 6 began in 2010, despite beingdesignedearlierand—atthattime—little data for gefitinib and none for erlotinib in this patient population had beenpublished. However, platinum-based treatment is still anoption, and the choice of comparator was approved by theregulatory authorities. The chemotherapy regimen anddose of gemcitabine and cisplatin was commonly used inAsia at the time that the study was designed. Recruitmentfor a head-to-head trial of afatinib compared with gefitinibfor first-line treatment of EGFR common mutation-positive lung adenocarcinoma has been completed(NCT01466660; LUX-Lung 7), with the results expected toprovide the first evidence of the comparative efficacy andsafety of these drugs for this setting.

6，核心結論強調。

The results of LUX-Lung 6 show a more profound anddurable effect of afatinib than standard first-linechemotherapy with gemcitabine and cisplatin for Asianpatients with NSCLC tumours harbouring EGFRmutations. Afatinib should therefore be considered as afirst-line treatment option for this patient population.

推薦閱讀：