临床研究模板

题目

• --
  

背景

• --
  

方法

• 研究设计
• 研究对象
• 干预措施
• 研究终点
  

结果

• 生存
• 安全性
  

结论

• --
  

评论

• --
  

存在问题

• --
  

相似研究

• 对比图表
  

参考文献

• --
  

微信扫二维码转发分享

循证医学
黎明前的黑暗，曙光将至——ARCS-M研究带来的启示

冲出重围，双免为剑直指恶性胸膜间皮瘤——评CheckMate 743研究

免疫治疗开启新辅助新篇，精准策略探索前路漫漫——评NSCLC新辅助免疫治疗的机遇与挑战

需要深度精简！尽量简洁，表格和图片为主。


1.   文献来源

Pan Y，Zhang JT，Gao X，et al. Dynamic circulating tumor DNA during chemoradiotherapy predicts clinical outcomes for locally advanced non - small cell lung cancer patients[J]. Cancer Cell，2023，41（10）：1763-1773.e4.

2.   证据水平

2b。

3.   背景

PACIFIC研究与GEMSTONE-301研究奠定了同步/序贯放化疗联合巩固免疫治疗在局部晚期非小细胞肺癌（non-small cell lung cancer，NSCLC）的标准治疗地位。然而这种治疗模式带来了新的临床问题：患者对放化疗呈现不同的治疗反应，有19%的患者经根治性放化疗已达到治愈，继续给予免疫治疗不仅没有生存获益，反而增加了患者的不良反应和经济负担。然而现有的计算机断层扫描（computed tomography，CT）、正电子发射断层/计算机断层扫描（positron emission tomography/computed tomography，PET-CT）影像学检查手段并不能找到放化疗治愈的患者。

近年来由于检测技术的进步，基于循环肿瘤DNA（circulating tumor DNA，ctDNA）的分子残留病灶（molecular residual disease，MRD）检测开始探索运用于实体肿瘤。最新的研究显示：早期NSCLC术后MRD持续阴性患者是潜在治愈人群，而术后MRD阳性患者预后差。但是，对于接受根治性放疗的患者，ctDNA-MRD的相关研究较少，并且研究关注放化疗后ctDNA-MRD对生存的影响，对于放疗中ctDNA的变化及其影响尚未有报道。

4.   目的

ctDNA-MRD在不可切除局部晚期NSCLC放化疗期间的动态变化特征及其对预后的影响。

5.   研究设计

•研究条件：中国单中心研究。

•研究时间：2019年7月至2021年12月。

•研究方法：单中心、前瞻性、观察性研究。

•研究对象：不可切除局部晚期NSCLC，接受常规分割的根治性放疗（60~66 Gy/30~33 F）。

•干预措施：在基线、放疗前、放疗中、放疗结束时间节点，以及巩固治疗/随访期间每3~6个月抽血检测ctDNA-MRD（图 1）。研究一共入组139例患者，收集761个血液样本。

图  1  研究流程图

•评价指标：ctDNA动态变化，landmark时间点以及纵向ctDNA-MRD与无进展生存（progression-free survival，PFS）之间的关系。

6.   主要结果
6.1   放疗早期MRD清零患者的预后好，接受巩固免疫治疗没有生存获益

诱导治疗后仅7%的患者ctDNA-MRD由阳性转为阴性，而在放疗中、放疗结束分别有30.9%、23.9%的患者转阴；ctDNA的浓度在放疗中出现明显下降，表明在根治性放化疗的整个过程中，放疗发挥了更为重要的作用。放疗中ctDNA-MRD清零并维持到放疗结束，即放疗早期MRD清零的患者，其中位PFS为22.9个月，明显优于晚期清零患者（10.3个月）；早期清零患者接受巩固免疫治疗并未有生存获益（20.3 vs. 22.9个月，P=0.917）。

6.2   放化疗结束后MRD持续阴性为潜在治愈人群

研究共有90例患者出现肿瘤复发，其中97.8%在随访中ctDNA-MRD出现阳性；20.1%的患者纵向MRD检测持续阴性，其1、2年的无进展生存率高达94.7%、88.4%，明显优于MRD阳性患者；MRD持续阴性患者有5例由于感染或免疫相关性肺炎等非肿瘤原因死亡。与手术患者不同，所有脑转移患者ctDNA-MRD均检测为阳性，仅在2例局部复发患者中ctDNA-MRD检测为阴性。与影像学相比，ctDNA-MRD中位提前4个月发现肿瘤复发。

6.3   基线血液ctDNA是局部晚期肺癌MRD检测的必要补充

仅41.0%的局部晚期NSCLC患者基线有足够的肿瘤组织进行基因检测；检测的突变中仅来源于血液ctDNA的占51.9%，而仅来源于肿瘤组织的占1.9%；对比三种ctDNA-MRD检测技术：未知肿瘤信息、已知肿瘤组织信息和已知基线血液ctDNA信息，结果显示已知基线血液ctDNA的MRD检测技术可以获得更多的肿瘤突变信息（P=0.048）。

7.   结论

放化疗期间MRD早期清零患者预后更好；持续MRD阴性患者为潜在治愈人群；基线血液ctDNA是局部晚期肺癌MRD检测的必要补充。

8.   评论

随着基因测序技术的不断进步，基于血液标本的MRD检测在NSCLC疗效预测、预后评估和个体化治疗决策中发挥了越来越大的作用。首先，基于外周血ctDNA的MRD动态监测，可以准确地预测早期肺癌术后复发[1-3]。著名的TRACERx研究，最早探讨了MRD监测在NSCLC术后复发中的预测作用[4]：在14例复发患者中，有13例在出现明显临床症状之前可检测到ctDNA “复阳”，且ctDNA “复阳”与影像学确认之间的时间间隔中位数为70天。此外，MRD检测也有助于预测晚期NSCLC患者的治疗反应[5-7]。例如，对于接受奥希替尼和贝伐珠单抗治疗的晚期EGFR突变NSCLC患者而言，治疗后6周ctDNA中EGFR突变的丰度可以预测早期进展风险[6]。在评估卡瑞利珠单抗联合化疗用于晚期肺鳞癌一线治疗的CameL-sq研究中，ctDNA清除患者的PFS和总生存期（overall survival，OS）显著延长；治疗过程中ctDNA的动态变化能精准筛选预后不良人群[8]。

吴一龙教授团队围绕着MRD这一主题做了一系列精彩的研究，不仅推动了科学研究的进步，也为临床实践提供了宝贵的参考。2022年发表在《Cancer Discov》的研究纳入了261例接受根治性手术的Ⅰ~Ⅲ期NSCLC患者，结果表明，持续MRD阴性患者可能代表潜在治愈的人群。亚组分析显示，MRD阴性患者可能无法从辅助治疗中受益 [1]。在最近的这项发表在《Cancer Cell》的研究中，吴一龙教授团队探索了ctDNA - MRD在不可切除局部晚期NSCLC放化疗期间的动态变化特征，发现放化疗早期MRD清零患者的预后更好，而且早期MRD清零患者接受巩固免疫治疗并未表现出生存获益。放化疗结束后MRD持续阴性患者是潜在治愈人群[9]。上述研究揭示了MRD在NSCLC患者治疗管理中的重要价值，提示ctDNA-MRD可用于精确地区分术后哪些患者需要辅助治疗以及放化疗后哪些患者需要继续免疫巩固治疗。在此基础上，基于MRD状态的前瞻性、干预性的临床研究有助于进一步明确MRD在临床实践中用于辅助治疗决策的价值。

此外，发表在《Cancer Cell》的这项研究中观察到放化疗早期（放疗剂量达到40 Gy）有30.9%的患者ctDNA清零。这一比例与既往新辅助放化疗（放疗剂量40~46 Gy）用于Ⅲ期NSCLC的临床研究中观察到的病理完全缓解率（16%~27%）基本相符[10-14]。值得注意的是，诱导化疗后仅7%的患者ctDNA-MRD由阳性转为阴性，而在放疗中有30.9% 的患者转阴，这提示相较于化疗，放疗在Ⅲ期NSCLC的治疗中可以发挥更好的根治作用。

与此同时，这项研究还发现基线血液ctDNA是指导局部晚期NSCLC MRD检测过程中靶基因富集、分子探针设计等环节重要的参考依据。不可手术切除的局部晚期NSCLC患者可能经常面临没有足够的肿瘤组织用于基因检测的困境；同时，由于肿瘤异质性，基于穿刺活检标本的检测结果，可能无法反映全貌。这项研究对比了三种ctDNA-MRD检测技术（未知肿瘤信息、已知肿瘤组织信息和已知基线血液ctDNA信息），发现已知基线血液ctDNA的MRD检测技术可获得更多的肿瘤突变信息。而且，基线外周血ctDNA测序也可为后续MRD检测提供必要的技术支持[2, 15]。

传统观点认为由于血脑屏障的存在，外周血ctDNA检测并不能很好地监测脑转移[16-17]。然而，发表在《Cancer Cell》的这项研究发现对于放化疗后出现单纯脑转移的患者，外周血ctDNA也可早期预警脑转移的发生。研究者认为这一现象背后的原因，是由于原发灶进展和放射性纤维化在影像学上很难区分，这些脑转移患者可能还存在原发灶的进展。关于ctDNA检测在脑转移患者中的临床价值仍值得进一步探索：一方面，受到血脑屏障的影响，患者血浆中ctDNA含量往往非常少[18]，而脑脊液中ctDNA的含量远高于血浆，脑脊液ctDNA检测vs. 外周血ctDNA检测在脑部转移监测中的应用值得深入探索；另一方面，放疗后原发灶（放疗野内病灶）的复发和放射性改变在临床上确实难以鉴别。能否用ctDNA鉴别原发灶复发和放射性改变值得进一步的前瞻性研究。

综上，发表在《Cancer Cell》的这项研究深化了我们对于NSCLC的MRD监测的理解，提示ctDNA-MRD可用于精确地区分放化疗后免疫巩固治疗的获益人群。未来，基于MRD状态的前瞻性、干预性的临床研究有助于推动NSCLC的个体化治疗的进一步发展。

1.   文献来源

Planchard D，Jänne PA，Cheng Y，et al. Osimertinib with or without chemotherapy in EGFR - mutated advanced NSCLC[J]. N Engl J Med，2023，389（21）：1935-1948.

2.   证据水平

1b。

3.   背景

奥希替尼是表皮生长因子受体（epithermal growth factor receptor，EGFR）突变晚期非小细胞肺癌（non-small cell lung cancer，NSCLC）患者的首选一线治疗方案。该疗法的推荐是指南基于Ⅲ期FLAURA临床研究的结果显示与一代EGFR酪氨酸激酶抑制剂（tyrosine kinase inhibitors，TKIs）相比，一线奥希替尼治疗具有更长的无进展生存期（progression-free survival，PFS）和总生存期（overall survival，OS）。既往研究发现，与单独使用靶向药物相比，一代EGFR-TKI吉非替尼联合卡铂和培美曲塞的疗效更好。

•奥希替尼联合化疗与奥希替尼单药靶向治疗在EGFR敏感突变未经治疗晚期NSCLC患者中的疗效对比情况尚不明确。

4.   目的

旨在评估奥希替尼联合化疗在EGFR敏感突变阳性未经治疗的晚期NSCLC的疗效是否优于奥希替尼单药方案。

5.   研究设计

•研究条件：全球21个国家共计176个中心参与的FLAURA2研究（NCT04035486）。

•研究方法：多中心、Ⅲ期、开放标签、随机对照临床试验。

•研究时间：2020年06月至2021年12月。

•研究对象：年龄≥18岁（日本地区≥20岁）；初治型携带EGFR敏感突变（exon19del/exon21 L858R）的局部晚期或晚期非鳞NSCLC患者；WHO PS 0~1，未接受过系统/EGFR-TKI治疗；若存在中枢神经系统（central nervous system，CNS）转移灶，则要求神经系统症状稳定。

•干预措施：557例符合入组条件的患者按1∶1比例随机分配至靶向联合化疗组[奥希替尼80 mg，qd，培美曲塞500 mg/m2，顺铂75 mg/m2或卡铂（曲线下面积，area under the curve，AUC）= 5 mg/mL/min]，3周为一个周期]或者单药靶向治疗组（奥希替尼80 mg，qd）。最终有276例患者接受奥希替尼联合化疗，275例患者接受奥希替尼单药靶向治疗。根据实体瘤疗效评价（Response Evaluation Criteria in Solid Tumors，RECIST）1.1标准评估疗效直到疾病进展或出现不可耐受毒性。研究流程见图 1。

图  1  FLAURA2研究的流程和给药方案

•评价指标：主要终点为由研究者根据RECIST 1.1评估的PFS，敏感性分析为BICR根据RECIST 1.1进行评估的PFS。次要终点包括OS，客观缓解率（objective response rate，ORR），缓解持续时间（Duration of Response，DoR），疾病控制率（disease control rate，DCR），药物的安全性、耐受性和生存质量等。

6.   主要结果

共入组557例患者进行1∶1随机，结果表明奥希替尼联合化疗一线治疗方案的疗效优于奥希替尼单药治疗方案，安全性可控。奥希替尼联合化疗组中位PFS明显优于单药奥希替尼治疗组[25.5 vs. 16.7个月，风险比（hazard ratio，HR）=0.62，P＜0.001，95%可信区间（confidence interval，CI）0.49~0.79]，OS数据尚未成熟。ORR在两组分别为83%和76%，详见表 1。

以下的表格是差不多抄自原文，这是不合适的，建议制作与传统资料的表格，既有历史数据或者其他研究的对比。
表  1  FLAURA2研究的主要疗效结果

评价指标	奥希替尼联合化疗组（n=279）	奥希替尼单药组（n=278）	HR（95%CI）	P值
中位PFS/月	25.5	16.7	0.62（0.49~0.79）	＜0.001
伴有CNS转移	24.9	13.8	0.47（0.33~0.66）	-
不伴有CNS转移	27.6	21.0	0.75（0.55~1.03）	-
ORR/%	83（78~87）	76（70~80）		-
中位药物反应期/月	24.0（20.9~27.8）	15.3（12.7~19.4）		-
24个月生存率/%	79（73~83）	73（67~78）		-

在主要不良反应方面，奥希替尼联合化疗组3级不良事件发生率较高，但与既往报告一致，无新的不良事件发生，详见表 2。

表  2  FLAURA2研究的主要不良反应 n（%）

分组	奥希替尼联合化疗组			奥希替尼单药组
	所有程度	3/4级		所有程度	3/4级
所有不良反应事件	276（100）	176（64）		268（97）	75（27）
贫血	128（46）	55（20）		22（8）	1（＜1）
腹泻	120（43）	8（3）		112（41）	1（＜1）
恶心	119（43）	4（1）		28（10）	0
食欲减退	85（31）	8（3）		26（9）	2（1）
便秘	81（29）	1（＜1）		28（10）	0

7.   结论

对于携带EGFR敏感突变的局部晚期或转移性NSCLC患者，一线使用奥希替尼联合培美曲塞和铂类化疗的疗效优于单药奥希替尼方案。同时，联合治疗组整体药物耐受性良好。

8.   评论

EGFR - TKI极大地改善了EGFR突变晚期NSCLC患者的预后，同时提高了患者的生活质量，已成为该类患者的一线治疗首选。FLAURA研究显示奥希替尼较第一代EGFR-TKI具有更好的PFS和OS获益[1-2]。以奥希替尼为代表的第三代EGFR- TKI是目前EGFR突变晚期NSCLC的一线标准治疗，已在临床广泛使用。然而，由于原发性和继发性耐药，部分患者不可避免的会出现疾病进展[3-4]。大量研究证实联合疗法可以克服或延缓耐药，还可以通过协同作用来增强抗肿瘤效果，是治疗EGFR突变NSCLC的一种有前景的策略[5-6]。近年来EGFR突变晚期NSCLC领域开展了多项联合治疗方案的临床研究。其中，NEJ009研究开启了EGFR-TKI联合化疗的治疗模式探索，结果显示与单独使用吉非替尼相比，吉非替尼联合培美曲塞和卡铂方案显著改善了患者的PFS[7-8]；JO25567[9]、NEJ026[10]、CTONG1509[11]研究进行了EGFR-TKI联合抗血管生成药物的探索，结果均显示贝伐珠单抗联合厄洛替尼的PFS显著优于厄洛替尼单药。基于RELAY研究[12]，2020年5月29日美国食品药品监督管理局（Food and Drug Administration，FDA）批准VEGFR2拮抗剂雷莫芦单抗与EGFR-TKI厄洛替尼联合用于EGFR 19del及ex21 L858R突变转移性NSCLC的一线治疗。另外，在EGFR领域一些新型抗体药物偶联（antibody-drug conjugate，ADC）药物正在涌现，绝大多数处于后线研究，针对一线联合治疗方案也在探索中，例如奥希替尼联合HER3-DXd一线治疗的Ⅱ期试验等。

FLAURA2研究是首个报道三代EGFR-TKI奥希替尼联合化疗在EGFR敏感突变初治晚期NSCLC的疗效是否优于奥希替尼单药的国际多中心随机Ⅲ期临床研究，研究结果显示奥希替尼联合化疗组中位PFS明显优于奥希替尼单药治疗组（25.5 vs. 16.7个月，HR=0.62）[13]。在当前EGFR-TKI单药一线治疗PFS陷于瓶颈之际，FLAURA2研究的数据无疑给我们带来了鼓舞和希望。本评论将从FLAURA2的研究设计和临床实践的挑战等方面进行多维度解析。

第一，研究设计的合理性。首先，FLAURA2研究对照组的选择是当前最佳对照疗法（奥希替尼单药）[1-2]；其次，基于第一代EGFR-TKI联合化疗对比EGFR-TKI单药治疗的疗效获益数据，支持奥希替尼联合化疗作为试验组的研究假设[7-8]；再者，FLAURA2研究在随机化前还设置了安全导入期以确认奥希替尼联合化疗的安全性，最大程度保证受试者的安全。此外，FLAURA2研究终点选择了研究者和盲态独立中心（blinded independent central review，BICR）双重评估，兼顾了临床实践及数据的一致性考量，该研究设计合理，符合伦理性和科学性。

第二，研究数据的稳健性。首先，FLAURA2研究在全球范围内进行，基线特征与流行病学一致，如EGFR 19del与L858R突变的比例在全球队列中研究组和对照组分别为61% vs. 38%和60% vs. 38%，这一比例与流行病学发生率一致[13]。其次主要研究终点（BICR评估）与敏感性分析（研究者评估）所评估的PFS均显示显著改善，获益趋势一致（HR都是0.62[13]）。研究者评估的PFS从单药治疗组的16.7个月延长至联合治疗组的25.5个月（HR=0.62），BICR评估的PFS从单药组的19.9个月延长至联合组的29.4个月（HR=0.62），研究者和BICR的疗效评估显示出很好的一致性。再者，中国数据与全球数据在疗效和安全性上获益趋势一致。

第三，未出现新的安全性信号。FLAURA2研究中，奥希替尼联合化疗的不良反应与单药已知的安全谱一致，未发现联合后出现新的安全风险，研究随机期与安全导入期的安全性结论一致。但仍然观察到，奥希替尼联合化疗组的≥3级的与治疗相关不良事件（treatment-related adverse event，TRAE）发生率更高（联合组53% vs. 单药组11%），其中联合组≥3级TRAE发生率在前3个月含铂化疗联合培美曲塞的是49%，奥希替尼联合培美曲塞的24%，联合治疗的AE发生率在含铂联合阶段最高，随后随着时间的推移而显著降低[13]。

总之，FLAURA2研究证实奥希替尼联合化疗在奥希替尼单药的基础上将PFS延长了9~10个月，研究数据也比较稳健和可靠，显然具有重要的临床意义。但是在临床实践中，我们很清晰地认识到面临的最大问题：是否能被患者或临床医生所接受？正如吴一龙教授在多个场合提及的中国胸部肿瘤研究协作组（Chinese Thoracic Oncology Group，CTONG）在中国医患人群中进行的关于EGFR-TKI联合化疗接受度的一项调查报告：调查对象包括363名肺癌患者及其家属和584名医师，在已知TKI联合化疗可延缓耐药达9个月的前提下，仍有84.5%的医生、73%的患者、77%的患者家属选择单药治疗[14]。这就说明无论是患者、家属抑或临床医生都对化疗安全性和不便利性有所顾虑。

在当今精准治疗时代，临床医生和患者可能更希望在保证疗效的同时降低不良反应带来的损害以及经济负担。那么，如何识别潜在获益人群，是FLAURA2研究在临床实践中所面临的机遇和挑战。基于现有的数据，以下几个方面可能会给我们临床带来启发和思考：

第一，合并脑转移人群。脑转移是EGFR敏感突变患者常见的转移部位，约有30%~50%的患者在5年内会发生CNS转移[15-17]。脑转移也是肺癌患者的不良预后因素，显著降低患者的生活质量[18-19]。然而，由于当前很多药物的血脑屏障通透性有限，并且部分患者因担心颅脑放疗和外科治疗可能对神经造成损害而拒绝接受颅内局部治疗，导致脑转移依然是肿瘤细胞迁徙中的“避难所”。因此控制颅内肿瘤、延缓脑转移进展是临床医生选择治疗方案的重要考量因素。奥希替尼基于良好的血脑屏障通透性，以及FLAURA研究中脑转移的显著获益的数据（与一代EGFR-TKI相比，可降低52%的脑转移进展风险）而成为控制脑转移的首选推荐。尽管奥希替尼对脑转移有效，但疗效有待进一步提升，尤其是提高脑转移的完全缓解率达到“无脑转移灶的状态”，从而提高患者生活质量仍然值得探索。基于吉非替尼联合化疗较吉非替尼单药对基线有脑转移患者的更佳控制[20-21]，同时中国进行的GAP BRAIN Ⅲ期研究也进一步验证了一线使用吉非替尼联合化疗后改善颅内转移的疗效优于吉非替尼单药[22]，这些数据都为FLAURA2研究中奥希替尼联合化疗对脑转移患者的疗效探索提供了数据支持。

FLAURA2研究设计中包括CNS全分析集（CNS full analysis set，cFAS）和可评估分析集（CNS evaluable-for-response set，cEFR）。cFAS集中，奥希替尼联合化疗组和奥希替尼单药组分别纳入了118例和104例基线存在CNS转移灶的患者；其中40例和38例患者纳入cEFR分析集（基线存在≥1个可测量CNS转移灶且直径≥10 mm）的患者。在cFAS集中，联合组对比单药组的CNS进展或死亡的HR为0.58（95%CI 0.33~1.01），CNS客观缓解率分别为73% vs. 69%；CNS完全缓解（complete response，CR）率分别为59% vs. 43%。在cEFR集中，CNS PFS HR为0.40（95%CI 0.19~0.84），CNS客观缓解率分别为88% vs. 87%；CNS CR率分别为48% vs. 16%。研究者预估从随机化分组开始的24个月内，联合治疗组患者CNS进展的累计发生率将为9%，显著低于单药治疗组的23%。

FLAURA2研究公布的CNS疗效和安全性数据，显示在奥希替尼单药治疗基础上联合化疗，能更进一步延缓乃至预防脑转移的发生，这对患者的全程管理是非常有利的。值得注意的是，联合组较高的CR率显示了缓解深度的改善，尤其完全缓解的患者大多有多个脑转移病灶和预后较差的软脑膜转移灶，而延缓脑转移的发生和进展，降低CNS进展的风险可有助于患者保留神经认知，维持或改善生活质量。因此，基于神经系统症状改善的渴望和对化疗毒性恐惧的综合评估，脑转移患者可能更适合联合治疗策略。

第二，合并混合突变人群。EGFR突变患者是一组异质性较高的人群。混合突变约占22%~55%，L858R合并突变高达69%，吸烟人群EGFR合并突变比例高达91%；合并突变患者，EGFR-TKI单药治疗ORR、mFPS、mOS都显著低于EGFR单纯突变[23-25]。合并TP53突变与EGFR-TKI治疗时长负相关[26]。目前FLAURA2中尚未发布合并突变人群相关结果，期待后续数据报道。正在开展的奥希替尼的TOP研究以及阿美替尼的ACROSS研究均旨在评估EGFR-TKI联合化疗针对EGFR合并共突变的疗效探索。期待这些研究数据的公布。

第三，EGFR突变未达早期清除人群。AURA3和FLAURA研究的探索性biomarker分析显示：血液中EGFR突变循环肿瘤细胞DNA（circulating tumor DNA，ctDNA）在奥希替尼单药治疗3周后依然存在的患者，疗效显著劣于ctDNA清除的患者（AURA3 mFPS 5.7 vs. 10.9个月；FLAURA 11.3 vs. 19.8个月）；6周ctDNA清除与否也显示相似的结果。3周血浆EGFR突变ctDNA清除率是不良疗效预后因素。阿美替尼联合化疗的Ⅱ期NCT04646824研究显示：ctDNA清零患者人群ORR达90.9%，远高于未出现ctDNA清零患者ORR的33.3%，ctDNA清零可能提示患者预后较好[27]；目前由王洁教授牵头的FLAME研究旨在探索奥希替尼联合化疗对比奥希替尼单药治疗3周ctDNA EGFR突变未清除患者的疗效和安全性及德国国立Network Genomic Medicine开展的类似研究正在进行当中。期待这些数据的公布能为未来基于ctDNA等biomarker精准指导临床用药方案提供更多强有力的证据。

综上所述，FLAURA2研究设计合理、研究数据稳健、结论可靠。研究结果提示奥希替尼联合化疗（培美曲塞+铂类）对比奥希替尼单药，一线治疗EGFR突变晚期NSCLC患者在PFS方面具有统计学和临床意义的显著获益。各亚组也显示了获益的趋势，特别是脑转移亚组可能是潜在优势人群。OS数据尚不成熟（目前获益趋势不明显）。联合化疗组不良反应增加明显，但符合预期，可控可管理。但是，因化疗的特殊性，目前临床要广泛推广FLAURA2治疗模式可能还有很多挑战：诸如患者、医生对化疗的接受度，治疗模式的不便捷性等。寻找能够获益的特定高危群体至关重要。未来，我们更需要通过肿瘤特征、分子特征、患者特征以及个体意愿，甚至血液分子残留病灶（molecular residual disease，MRD）动态监测来精准实施这部分患者的个体化治疗，从而让患者高质量的活得更久。

Genetic prediction of medication use patterns in cardiometabolic disease. Nat Med. 2023 Jan;29(1):43-44. doi: 10.1038/s41591-022-02124-3. PMID: 36683070.
This is a summary of:
Kiiskinen, T. et al. Genetic predictors of lifelong medication use patterns in cardio- metabolic diseases. Nat. Med. https://doi. org/10.1038/s41591-022-02122-5 (2023).

背景：
By performing a large-scale biobank-based genome-wide association study, we identified a strong link between the underlying risk of cardiometabolic disease and patterns of lifelong medication use in hyperlipidemia, hypertension and type 2
diabetes. We discover hundreds of genetic predictors of medication use behavior and show medication-use-enhanced applications for polygenic prediction in cardiometabolic diseases.

The project
Cardiovascular disease is a major cause of death worldwide1 and is underpinned by a complex interaction between environmental and genetic factors2. Although lifestyle changes can have clinical benefit, pharmaco- logical interventions remain an in- tegral part of cardiovascular disease prevention. However, challenges exist in optimizing long-term medi- cation use, including inter-individual variability in treatment responses, drug side effects and poor patient compliance. Genetic data have been proposed as a tool to optimize drug therapies, although most pharmaco- genetic studies have focused on drug responses including adverse effects as the primary outcome3, whereas little is known about the genetic basis of medication usage patterns. We therefore set out to produce the first systematic study of genetic effects on medication usage behavior, using cardiometabolic medications as a model. In doing so, we aimed to test three hypotheses. First, that medication purchase data can be used to identify genetic factors underlying cardiovascular diseases and their risk factors; second, that genome-wide data inform us about the likelihood of medication switch- ing and stopping; and third, that genetic variation associated with medication usage enables better pre- dictions of cardiovascular diseases.

The discovery
We performed a large-scale genome-wide association study(GWAS) on three fundamental medi- cation usage behavior phenotypes: the cumulative usage of medication; medication switching; and stopping the use of medication. In addition, we developed polygenic risk scores to predict patterns of medication usage, and analyzed medication usage data to identify risk factors for cardiometa- bolic diseases.
We discovered a total of 333 inde- pendent genetic loci: 74 with leading associations for hyperlipidemia, 181 for hypertension, and 78 for type 2 diabetes medication patterns. These included genetic loci associated with long-term medication use, but also with the switching or stopping of medication. Almost half of these loci were also associated with the risk of coronary heart disease. The genetic discovery also included 40 genetic loci with no previous links to the un- derlying risk factors of low-density lipoprotein (LDL)-cholesterol levels, blood pressure levels or type 2 dia- betes. These included eight loci that modify coronary artery disease risk and were reported for the first time.
We also showed a strong genetic corre- lation between the purchase of all three medications and the risk of coronary heart disease (Fig. 1).
Finally, we derived polygenic risk scores for LDL, systolic blood pres- sure and type 2 diabetes, and showed that these scores were predictive
of medication behavior, including switching from first-line treatments to second-line stronger medications and the stopping of medications
for hyperlipidemia, hypertensive or type 2 diabetes. These find- ings may lead to opportunities for genome-guided decision-support tools for treatment choices.

Future directions
Our genome-wide analyses with over 300 associated loci demon- strate a highly polygenic genetic architecture behind the heritable component of lifelong drug use for the treatment of hyperlipidemia, hypertension and type 2 diabetes that is largely overlapping with the genetics of cardiovascular disease risk factors.
We combined different drugs with varying mechanisms of action used in the treatment of hypercholes- terolemia, hypertension and type 2 diabetes to maximize the statistical power of our analyses. However, it is worth noting that by grouping drugs together, the possible pharmacog- enomic aspects that differ from molecule to molecule might be missed and warrants further studies with larger datasets.
We have probably barely scratched the surface of the opportunities pro- vided by decades of nationwide pre- scription medication purchase data and the genetic (and non-genetic) factors affecting our medication usage behavior. Extending the approach to other medication use patterns outside cardiometabolic medications is an obvious next step.

Figure

Fig. 1 | Relationship between drug purchases and underling cardiometabolic risk factors. The graphs show genetic correlations between the total numbers of drugs purchased for cardiometabolic indications and the underlying cardiometabolic risk factors and coronary artery disease. All three medication purchase phenotypes show high correlations with the underlying risk factors, but are also genetically correlated with the risk of coronary artery disease (CAD). LDL, low-density lipoprotein; SBP, systolic blood pressure; T2D, type 2 diabetes. © 2023, Kiiskinen, T. et al.

Behind the paper
This is, to our knowledge, the first large-scale effort to study the genetic determinants of medication usage by combining genome- wide profiling data with longitudinal medication purchase data. A considerable proportion of the project involved refining the medication purchase trajectory data and harmonizing the phenotypes across the three biobanks. We were then ready to answer two big questions: what is the role of genetic variation in medication-related behavior, and how do we potentially use genetics to guide treatment better and encourage positive medication-related health behavior? Strikingly, we not only identified more than 300 genetic loci associated with cardiometabolic medication behavior, but we were also able to develop genetic risk scores that predicted individuals likely to switch medications or stop using them. Although our findings concern cardiometabolic medications, they highlight the potential utility of using risk- factor-related genetic information for the optimization of pharmacological treatment more broadly.

References
1.        James, S. L. et al. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 392, 1789–1858 (2018).
A review article that presents global burden of 354 disease including cardiovascular diseases.
2.        Marenberg, M. E., Risch, N., Berkman, L. F., Floderus, B. & de Faire, U. Genetic susceptibility to death from coronary heart disease in a study of twins. New Engl. J. Med. 330, 1041–1046 (1994).
This paper reports heritability estimates for coronary heart disease.
3.        SEARCH Collaborative Group. SLCO1B1 variants and statin-induced myopathy—a genomewide study. New Engl. J. Med. 359, 789–799 (2008).
A review article that presents evidence for adverse reactions to statin use for individuals carrying risk variants in the SLCO1B1 gene.

From the editor
“This study stood out to us as it presents an elegant approach to identifying genetic loci for cardiometabolic diseases, based on the analysis of longitudinal medication use data from large biobanks and electronic health records.” Editorial Team, Nature Medicine.

Genomic autopsy offers answers for pregnancy loss and perinatal death. Nat Med. 2023 Jan;29(1):41-42. doi: 10.1038/s41591-022-02143-0. PMID: 36670297.

This is a summary of:
Byrne, A. B. et al. Genomic autopsy to identify underlying causes of pregnancy loss and perinatal death. Nat. Med. https://doi. org/10.1038/s41591-022-02142-1 (2023).

背景：

The cause of pregnancy loss or perinatal death often remains unexplained, even following a standard autopsy.
Comprehensive genomic investigation of pregnancy loss or perinatal death identifies a cause in over 50% of cases, particularly where congenital abnormalities are present. Causes of stillbirths without congenital abnormalities remain difficult to identify.

The question

Of well-established pregnancies, 1% result in pregnancy loss, terminations of pregnancy for fetal abnormalities or perinatal death (defined as fetal loss (stillbirths >20 weeks or >400 g) or neonatal death (up to 28 days post birth). Congenital abnormalities occur in ~30% of cases of perinatal death. Identifying an underlying reason for pregnancy loss or perinatal death has largely relied on standard autopsy, which for various reasons is performed in <50% of cases.
Standard autopsy usually does not involve any genetic testing beyond chromosome analysis, and establishes a precise cause of death in <50% of families. Perinatal death is a traumatic event for families, and establishing a cause not only provides answers but also provides reproductive opportu- nities for the future, including prenatal diagnosis and pre-implantation genetic diagnosis1.

The solution

In this study, we assessed whether whole-exome sequencing (WES) or whole-genome sequencing (WGS) sig- nificantly improves the diagnostic yield
and clinical utility of the autopsy process for 200 families. Families were required to have received a standard fetal autopsy and were recruited as trios (both parents) or quads (families with more than one affected pregnancy). The babies were required to be euploid and to not have an obvious clinical diagnosis of a genetic disease. We performed numerous differ- ent research investigations such as RNA experiments and genetic matchmaking, and even pursued animal models for novel genetic findings to provide evidence that proved novel genetic alleles and genes were causative.
We found a definitive molecular diagnosis in known disease genes in 52 families (Fig. 1a). This number includes 10 families for whom a definitive genetic diagnosis was achieved from additional genetic or experimental evidence, includ- ing additional cases from international data sharing2 and creation of an animal model3. Through RNA work including reverse transcription PCR, nanopore sequencing and RNA sequencing (RNA-seq), we established allelic patho- genicity in four cases. A further 53 families had highly suspicious findings that by the standards and interpretation guide- lines for genetic testing today on gene– disease validity and variant pathogenicity cannot be given a definitive genetic diag- nosis4. These findings include 27 with novel variants or phenotypes with known disease genes, and 18 potentially novel disease genes. Our cohort is perhaps sur- prisingly heterogeneous as only 7 of 96 genes were observed more than once. Our cohort, and similar cohorts, are im- portant for further genetic discoveries in pregnancy loss and perinatal death.
Our cohort was biased to babies with congenital defects, particularly those detectable by ultrasonography, and our diagnostic rate was low (8% had a pathogenic or likely pathogenic variant) in apparently normal stillborn babies at older age (>30 weeks) of gestation (Fig. 1b). WES or WGS identified a defini- tive or candidate genetic diagnosis in 105 of 200 families (52.5%).

Future directions


This research came about as a result of clinical need; so often families were left without answers after the devastating loss of a pregnancy or when congenital abnormalities were present in a fetus that led to termination of pregnancy or death in utero. How could no cause be found? After several unsuccessful applications, our research was funded by the National Health and Medical Research Council and then the Genomics Health Futures Mission of the Medical Research Futures Fund, which enabled us to establish a national referral process for perinatal death in pregnancy loss in Australia. Initially our turnaround times were slower than we wanted, and after improving turnaround time from sample receipt to reporting of results within three months, recruitment to the trial improved substantially. Most importantly, numerous couples have used our results to have healthy pregnancies following earlier losses.


Figure

Fig. 1 | Diagnostic yield of genomic autopsy. Stillbirth of babies with no documented congenital abnormality by standard autopsy have the lowest diagnostic rate. Circle, female proband; square, male proband. © 2023, Byrne, A. B. et al.

Behind the paper

This research came about as a result of clinical need; so often families were left without answers after the devastating loss of a pregnancy or when congenital abnormalities were present in a fetus that led to termination of pregnancy or death in utero. How could no cause be found?
After several unsuccessful applications, our research was funded by the National Health and Medical Research Council and then the Genomics Health Futures Mission of the Medical Research Futures Fund, which enabled us to establish a national referral process for perinatal death in pregnancy loss in Australia. Initially our turnaround times were slower than we wanted, and after improving turnaround time from sample receipt to reporting of results within three months, recruitment to the trial improved substantially. Most importantly, numerous couples have used our results to have healthy pregnancies following earlier losses.

References

1.        Flenady, V. et al. Stillbirths: recall to action in high-income countries. Lancet 387, 691–702 (2016).
This paper presents multinational stillbirth statistics.
2.        Byrne, A. B. et al. Pathogenic variants in MDFIC cause recessive central conducting lymphatic anomaly with lymphedema. Sci. Transl. Med. 14, eabm4869 (2022).
This paper reports a patient series and animal model.
3.        Tan, N. B. et al. Recurrent de novo missense variants in GNB2 can cause syndromic intellectual disability. J. Med. Genet. 59, 511–516 (2021).
Data sharing mutations in GNB2 that.

From the editor

“This study was of interest to us because it presented a framework for the application of genomic sequencing to clarify potential genetic causes of unexplained perinatal deaths, and included follow up data of how the genomic information was used for counselling.” Editorial Team, Nature Medicine.

Postoperative circulating tumor DNA could guide CRC adjuvant treatment


This is a summary of:
Kotani, D. A. et al. Molecular residual disease and efficacy of adjuvant chemotherapy in patients with colorectal cancer. Nat. Med. https://doi.org/10.1038/s41591-022-02115-4 (2023).



The large, ongoing CIRCULATE-Japan trial is investigating the role of circulating tumor DNA (ctDNA)-based molecular residual disease testing in patients with resectable colorectal cancer after radical surgery. An interim analysis of GALAXY, a prospective, observational arm of
CIRCULATE-Japan, establishes ctDNA as a prognostic and predictive biomarker.



The problem
Surgical resection is the standard-of-care for patients with stage II or III colorectal cancer (CRC). Adjuvant chemotherapy is recommended for high-risk patients with stage II CRC and all patients with stage III CRC1. The American Society of Clinical On- cology (ASCO) notes that the current defi- nition of ‘high-risk’ stage II is inadequate, as many such patients are recurrence-free, whereas some ‘average-risk’ patients do develop recurrence. The ASCO also acknowledges that no clinicopathologi- cal features currently listed as high-risk predict a benefit from adjuvant chemo- therapy2. Although adjuvant chemothera- py improves survival in patients with stage III colon adenocarcinoma, this treatment has widely variable outcomes among patients with stage III CRC, leading some researchers to question the absolute ben- efit of adjuvant chemotherapy in a subset of this population3.
GALAXY is an observational arm of the CIRCULATE-Japan trial4 that aims to clarify how monitoring the levels of circulating tumor DNA (ctDNA) after surgery affects patients’ outcomes, its implications for adjuvant chemotherapy decision-making, and the association between ctDNA dy- namics and prognosis.



The observation
A total of 1,039 patients with stage II–IV resectable CRC were monitored pro- spectively using Signatera molecular residual disease testing for ctDNA. The primary end point of the study was disease-free survival (DFS; defined as the interval between surgery and recurrence or death); a secondary end point was ctDNA clearance5. The interim findings of GALAXY support the use of personalized, tumor-specific ctDNA testing to identify patients at increased risk of recurrence who are likely to benefit from adjuvant therapy.
Among patients who were ctDNA-positive at 4 weeks after surgery, 61.4% experienced recurrence (18-month DFS 38.4%), com- pared to 9.5% of ctDNA-negative patients (18-month DFS 90.5%). In a multivariate analysis that accounted for all clinicopatho- logical risk factors, ctDNA positivity at 4 weeks after surgery was the strongest risk factor for CRC recurrence.
Patients with high-risk stage II or III CRC who were ctDNA-positive 4 weeks after sur- gery (18%; 113 out of 644 patients) derived a survival benefit from adjuvant chemother- apy, whereas ctDNA-negative patients (82%; 531 out of 644 patients) did not (Fig. 1).

Moreover, ctDNA dynamics are indicative of treatment responses. Among initially ctDNA-positive patients who received ad- juvant chemotherapy, those who achieved ctDNA clearance had superior DFS to those with persistent ctDNA positivity.


The implications
Our findings signify a paradigm shift in how we manage patients with CRC after surgery. The GALAXY interim findings show that ctDNA positivity at 4 weeks after surgery is a strong prognostic marker that identifies a group of patients at high risk of both CRC recurrence and decreased DFS.
Our study also supports ctDNA positivity as a predictive marker of benefit from adjuvant chemotherapy, as is exemplified by the overall benefit that ctDNA-positive patients derived from treatment versus the observation group. For ctDNA-negative patients, adjuvant chemotherapy conferred no survival ad- vantage over observation, which suggests that this group of patients can potentially be spared the toxicity of adjuvant chemo- therapy without compromising their long-term outcomes.
Although the GALAXY study is limited by its observational nature, as a part of CIRCULATE-JAPAN, we are also conduct- ing the ALTAIR trial, a randomized phase 3 study of treatment escalation or inten- sification in patients with CRC who are ctDNA-positive from 4 weeks onwards after surgery. We are also conducting the VEGA trial, a randomized phase 3 study of adjuvant chemotherapy de-escalation in patients with CRC who are ctDNA-negative after surgery.
Our findings could have important impli- cations for future CRC trial design. Enrich- ment of trial cohorts with patients who are likely to experience recurrence could enable substantial reductions in sample size, thereby making these studies faster and more cost-effective. Furthermore, our study showed that adjuvant chemotherapy can increase rates of ctDNA clearance and could potentially modify the outcomes of patients with postoperative ctDNA positiv- ity. This finding suggests that ctDNA status might have value as a surrogate end point or marker of treatment efficacy in clinical trials, which has the potential to consider- ably reduce trial duration and expedite the approval of new therapies in the future.


Expert opinion
“A key finding is that patients who are initially positive for ctDNA post-operatively can become ctDNA-negative after adjuvant chemotherapy and that this conversion correlates with improved outcome. Although anecdotal reports in a limited numbers of patients had suggested a benefit of adjuvant chemotherapy in initially ctDNA-positive patients, this large cohort provides an opportunity to assess the degree of benefit of adjuvant chemotherapy in the ctDNA-positive population, and the potential importance of ctDNA clearance.

Figure


Fig. 1 | ctDNA-based molecular residual disease testing is predictive of response to adjuvant chemotherapy in surgically treated patients with CRC. a,b, Kaplan–Meier estimates of DFS stratified by observation (red) and adjuvant chemotherapy (ACT; blue) in patients with pathological high-risk stage II and III CRC. a, Patients with ctDNA positivity at 4 weeks after surgery; hazard ratio (HR) adjusted by sex and performance status. b, Patients with ctDNA negativity at 4 weeks after surgery; HR adjusted by age, pathological disease stage, microsatellite instability and performance status. HR and 95% confidence
intervals were calculated by the Cox proportional hazard model. P values were calculated using the two-sided log-rank test. M, months. © Kotani, D. A. et al. CC BY 4.0.e2100181 (2022).
A review article that presents the clinical utility of ctDNA and its effect on clinical trial design.


Behind the papeR

We are on a mission to fight cancer and have a vision in which surgically treated patients will have no risk of recurrence at all.
The CIRCULATE-Japan project is designed to achieve this ultimate goal. Japanese surgeons and oncologists have shown keen interest in the trial and enrolment has been faster than expected. This is the first time that data from the GALAXY study have been published. Results from larger interventional trials (ALTAIR and VEGA) will be published in the near future.

The GALAXY study is notable for its incorporation of personalized ctDNA testing for each patient. As indicated in the accompanying research article, more than half of the genes selected for the tumor-informed ctDNA assay were unique to each patient. Detailed analysis of these genes could lead to new discoveries about recurrence.

fRom the editoR

“This study by Kotani et al. contributes to mounting evidence that the post-operative ctDNA status of patients with colorectal cancer can help in guiding adjuvant therapy decisions. These are important findings from a large prospective cohort that stood out to us because the delivery of adjuvant chemotherapy to only those patients that are likely to benefit is crucial for improving the quality of patients’ lives.” Editorial Team, Nature Medicine.


参考文献

1.        Yoshino, T. et al. Pan-Asian adapted ESMO Clinical Practice Guidelines for the diagnosis treatment and follow-up of patients with localised colon cancer. Ann. Oncol. 32, 1496–1510 (2021).
This publication reports the current clinical practice guidelines for CRC.
2.        O’Connor, E. S. et al. Adjuvant chemotherapy for stage II colon cancer with poor prognostic features. J Clin Oncol. 29, 3381–3388 (2011).
This study demonstrates a lack of overall survival benefit from adjuvant chemotherapy for patients with stage II CRC.

3.        Sobrero, A. F. et al. A new prognostic and predictive tool for shared decision making in stage III colon cancer. Eur. J. Cancer. 138, 182–188 (2020).
This study highlights variability in treatment outcomes among patients with stage III CRC.

4.        Taniguchi, H. et al. CIRCULATE-Japan: circulating tumor DNA-guided adaptive platform trials to refine adjuvant therapy for colorectal cancer. Cancer Sci. 1,
14926 (2021).

This publication discusses the clinical trial protocol associated with CIRCULATE-Japan.

5.        Kasi, P. M. et al. Impact of circulating tumor DNA-based detection of molecular residual 414 disease on the conduct and design of clinical trials for solid tumors. JCO Precis. Oncol. 6,e2100181 (2022).
A review article that presents the clinical utility of ctDNA and its effect on clinical trial design.