本次大会的主题为“砥砺深耕·笃行致远”，热情邀请全国相关领域专家、学者、科研骨干、企业代表与会交流。
致癌性试验的目的是在动物中识别潜在致癌作用，从而评价人体中的相关风险。实验室研究、动物毒理学试验和人体数据中出现的任何担忧因素，均可能导致需要进行致癌性试验。对于预期在患者一生中大部分时间连续使用的药物，通常会要求采用啮齿类动物进行致癌性试验。这些试验的设计和结果解释出现在目前已有的测试潜在遗传毒性的大多数技术和评估系统暴露量的最新技术之前。这些研究也出现在我们目前对非遗传毒性物质致癌性的认识之前。
现在遗传毒性试验、毒代动力学和机制研究的结果已被常规用于临床前安全性评价。这些新增数据不仅有助于判断是否需要进行致癌性试验，而且对解释研究结果与人体安全性的相关性也是十分重要的。由于致癌性试验耗时耗力，只有当人体暴露情况确实需要动物终生给药研究信息来评价其潜在致癌性时，才进行致癌性试验。
(本推文致癌研究内容章节由于较多，将分2个章节介绍。)
二乙基己烯雌酚（DES）于 1938 年首次合成，是一种合成形式的雌激素，从 20 世纪 40 年代开始被用于孕妇（尤其是美国、英国等欧洲国家和澳大利亚的孕妇），以预防流产。然而，直到 1971 年，《新英格兰医学杂志》上发表了一篇具有里程碑意义的文章，将母亲在怀孕期间服用 DES 的妇女在子宫内接触 DES 与一种罕见的阴道肿瘤联系起来。这一发现促使美国和欧洲的监管机构分别于1971年（美国）和1978年（欧洲）建议不再向孕妇开具DES处方。如今，DES 已被国际癌症研究机构（IARC）列为 1 类致癌物（即对人类具有致癌性），世界各地已基本停止使用该药物。
DES 的故事对药品开发商和监管机构都是一个重要的警示故事。尽管自DES问世以来，药物开发已经发生了广泛的变化，首先是监管机构将安全性测试的重任转移给了申办者，最近又出现了全球统一的测试要求，但致癌性测试仍然远远不只是 固定程序。近年来，随着一系列新型药物产品类别的推出，这一事实得到了进一步强调，这也带来了独特的测试挑战。例如，先进的细胞和基因疗法尽管前景广阔，但也引起了人们对病毒整合的长期影响和干细胞群意外扩增的担忧。随着这些技术和其他技术逐渐成熟，成为可行并可广泛使用的疗法，监管专业人员需要做好准备，在与监管机构沟通时，证明开发过程中使用的各种测试策略是合理的。
本章旨在概括介绍适用于致癌性研究的全球监管要求的现状。在阅读本章时，应结合前一章的遗传毒性评估（见第 7 章），因为这些评估通常先于致癌性测试计划，并为其提供某些支持信息，在对致癌可能性进行总体评估时也可加以利用。讨论将首先简要介绍与监管决策有关的致癌机制背景。随后将介绍与致癌性测试的进行和解释有关的全球监管要求和豁免的现状，这些要求和豁免主要来自国际人用药品技术要求协调理事会 (ICH)。最后，将介绍与模型选择和研究设计有关的具体考虑因素，然后讨论应如何在向全球监管机构提交的营销和上市前呈件中提交致癌性数据。
1
癌症和致癌性
世界卫生组织（WHO）将癌症定义为由 "异常细胞快速增殖引起的一组疾病"，"异常细胞的生长超出其正常边界，然后侵入身体的邻近部位，扩散到其他器官"，这一过程被称为 "转移"。最初的生长或肿胀被称为肿瘤或新生物。重要的是，虽然肿瘤可被定义为恶性或良性或（即癌或不癌），但监管科学历来将导致良性病变在统计上显著增加的药物也纳入 "致癌物 "的定义中。这一决定是切实可行的，因为良性病变会阻塞附近的组织，导致疼痛和其他病症，从而增加不可接受的风险。以下内容旨在概述从监管角度理解致癌性测试要求所必需的几个重要概念。如果读者希望对这一主题有更全面的了解，我们建议他们从现有的众多优秀文献22-24中选择一本进行阅读。
1
基因毒性和非基因毒性致癌物质
除了将肿瘤定义为良性或恶性外，致癌物质还可根据其作用机制分为基因毒性、非基因毒性或两者兼有。直接与脱氧核糖核酸（DNA）发生作用从而对其造成损害的药物被认为具有基因毒性，而通过间接机制致癌的药物则被认为具有非基因毒性。
• 基因毒性致癌物是指对 DNA 有致突变作用并有可能成为 "完全致癌物"（即能够诱发、促进和发展癌症，详见下一节）的致癌物。虽然这些药物在某些情况下可能存在 "安全 "剂量阈值（如秋水仙碱等无性繁殖化合物），但默认的监管假设是，这些药物具有低剂量线性风险与反应曲线（见图 8-1），因此即使少量也能造成直接的 DNA 损伤，最终导致癌症。
• 与此相反，非遗传毒性致癌物是表观遗传的，通过改造微环境使细胞更容易受到DNA损伤而起作用。这通常是通过增加各种浓度依赖性的细胞信号传导过程来实现的(例如，基因转录，或影响参与生长和增殖的各种受体)。因此，这些化合物的剂量-反应曲线假定有一个“安全”剂量阈值，低于该阈值，这些过程不会启动(图8-1)。
2
致癌过程的各个阶段
虽然致癌可能源于对细胞维持和生长的正常细胞过程的许多不同干扰因素，但将整个致癌过程视为一个多阶段过程是有帮助的，特定的致癌物质会在一个或所有阶段发挥作用（即 "完全致癌物"）。这些阶段被称为起始阶段、促进阶段和进展阶段（initiation, promotion, and progression）。
3
启动
启动涉及细胞 DNA 序列中不可逆转、可遗传的变化（包括突变和缺失）。由于起始剂具有诱变作用，因此也被认为具有基因毒性。大多数在起始阶段引起致癌变化的药物都是促基因毒性物质，因为它们需要经过一定程度的内部活化或生物转化（见第 6 章）后才能与DNA发生作用。因此，早期遗传毒性评估（见第 7 章）通常包括 S9 微粒体部分，因为这种混合物含有重要的肝酶，可在细胞（即体外）系统中形成药物的直接作用毒性代谢物。重要的是，起始事件本身并不能保证一定会导致癌症。在许多情况下，这些变化只是使细胞失去活力或无法增殖。但是，如果起始事件的影响是可持续的，那么癌变的下一步就是促进。
4
促进
促进是启动细胞进行克隆扩增以产生癌前病变的一般过程。这个阶段的致癌性通常是非遗传毒性的，通过无数的表观遗传机制起作用，包括激素调节、刺激和炎症、免疫抑制和细胞信号通路抑制。
鉴于其影响的间接性，如前所述，假定促进致癌过程阶段的致癌物具有“安全”阈值。此外，与开始和进步相反，促进通常被认为是可逆的。本章开头提到的致癌物二乙基己烯雌酚（DES）主要通过诱导雌激素受体介导的细胞增殖增加而起肿瘤启动子的作用。29
5
进展
癌变的最后阶段称为进展期。在这一阶段，连续的基因毒性变化导致单个肿瘤细胞内积累更多的突变。这些过程最终会选择出最具侵袭性（如增殖）的克隆，进而导致肿瘤不可逆转的发展。与起始制剂类似，针对肿瘤进展的制剂通常具有基因毒性，因此被认为没有“安全”的使用阈值。
2
监管概况
自美国国家癌症研究所 (NCI) 于 20 世纪 70 年代中期提出 两年啮齿类动物致癌性生物测定的概念以来，该方法已成为估算新药致癌可能性的黄金标准（the 2-year rodent bioassay for carcinogenicity has become the gold standard for estimating the carcinogenic potential for new drug products）。尽管在 20 世纪 90 年代制定 ICH 致癌研究指南时，ICH 的创始成员地区（即美国、欧盟国家和日本）已经制定了长期研究的要求，但这些指南在全球范围内的采用使得测定的许多特定方面得以统一。
这些方面包括高剂量选择（以前欧盟和日本允许使用每公斤毫克大于 100 倍的高剂量倍数来替代最大耐受剂量 (MTD)，但美国不允许），以及证明长期研究合理性的最短临床暴露时间（以前美国为 3 个月，欧盟和日本为 6 个月）。ICH S1 指导原则 通过统一这些预期和其他预期，力求减少全球药物开发过程中的冗余。
尽管如此，“两年检测”现在和过去都不是一项简单的工作。它的设计有几个方面有别于非临床试验中的许多传统试验。特别是，进行两项为期两年的传统啮齿类动物致癌性研究（通常在大鼠和小鼠模型中进行）需要较长的持续时间和较多的动物数量（即每组至少有 50 只雌雄动物），这使其具有独特的资源密集性，因此对申办者来说成本较高。
构思 ICH S1 指南时，我们认识到必须让监管期望与业界对癌症本身的认识同步发展。该附录描述了综合证据权重 (WoE) 评估，以帮助申办者确定为期 两年的啮齿类动物致癌性研究是否可能为整体致癌性测试方法带来附加值。如果不能增加价值，申办者现在可以寻求监管咨询，以支持完全豁免两年期研究的决定。
如表 8-1 所示，ICH 成员国几乎普遍实施了致癌性测试的相关指南。许多 ICH 观察员国和非成员国的监管机构也广泛推荐使用该指南。
1
致癌性测试要求
ICH S1B(R1)指南概述的一般要求是进行一项长期研究（通常是在大鼠体内进行为期两年的研究），以及第二项长期研究（通常是在小鼠体内进行）或一项短期或中期的体内（即动物）研究。最常用的较短时间研究包括啮齿动物癌症诱发和促进模型，以及使用转基因（如 p53+/-、Tg.AC、Tg.rasH2 和 Xpa 缺失）和新生啮齿动物进行研究。
2
考虑因素
从历史上看，暴露持续时间是决定是否需要进行研究的主要因素，对于打算服用 六个月以上或长期间歇性服用的产品，需要进行研究。
具体的 "关注原因 "是另一个重要的考虑因素。根据 ICH 指导原则 S1A，引起关注的原因可能包括：
1.先前证明与人类相关的产品类别具有致癌性、
2.结构-活性关系表明有致癌风险、
3.重复给药毒理试验中出现肿瘤前病变的证据，以及
4.母体化合物或代谢物在组织中长期滞留，导致局部组织反应或其他病理生理反应。
值得注意的是，虽然延长给药时间超过两个月被视为需要进行致癌性研究的理由，但根据 ICH 指南，这本身并不被视为 "值得关注的原因 "。
3
补充机理研究
虽然不是一项具体要求，但可能有必要进行额外的机理研究，以了解从致癌性测试中获得的任何结果。此类研究可侧重于各种终点，包括
• 细胞变化（如细胞凋亡或增殖）
• 生化变化（如生长因子和特定激素的血浆水平）
• 当标准组合检测结果为阴性且无法确定表观遗传机制时，进行额外的遗传毒性评估；以及
• 其他修改方案，以进一步确定效应的特征（如可逆性、用药中断等）。
这些额外机理研究的主要目的是更好地描述药物的作用机理及其与潜在人体风险的关系。例如，虽然已知激动过氧化物酶体增殖激活受体 α 的药物可诱发某些啮齿类动物癌症，但这一机制已被证明与评估人类癌症风险无关。
4
开发计划中的时间安排
致癌试验应在遗传毒性试验之后进行，如 ICH S2(R1) 所述。同时还应考虑到预期的患者人群和给药方案，以及早期药代动力学和药效学分析结果（见第 6 章）和重复给药毒理试验结果（包括必要的单独开展的剂量探索试验；见第 8 章）。
这些信息不仅对确定是否存在肿瘤前病变或其他重要变化（如免疫抑制或细胞增殖），而且对了解任何研究结果的机理基础，包括是否与人类癌症发展相关，都极具价值。假设没有 "令人担忧的原因"，对致癌性研究的一般期望是在申请上市批准之前完成，但不一定在启动大规模（即III期）临床试验之前完成。
5
监管要求预期的潜在例外情况
一般监管要求有多种例外情况。这些例外可能是基于目前对特定产品或产品类别的科学理解；或基于对偏离正常时间表、数据预期或两者可能产生的任何风险的审慎评估，同时考虑到任何潜在的益处。
6
“明确的”基因毒性药物
人们可能会认为，要从遗传毒性研究中获得明确的阳性结果（即基于一整套测试而非单一测试的结果），必然需要进行长期的致癌性研究，但事实并非如此。
S1B(R1) 指南指出，"无论是在没有遗传毒性风险的情况下，还是在有明确遗传毒性风险的情况下，为期 两年的大鼠研究都不太可能有价值。"换句话说，如果已经有足够的信息可以得出遗传毒性的结论，那么就可以推定该药物是一种跨物种致癌物，不太可能需要进行耗费大量资源开展两年的致癌试验“a “2-year rat study is less likely to be of value either in cases whether there is no genotoxicity risk, or in cases with unequivocal genotoxicity risk.” In other words, where already sufficient information exists to draw a conclusion regarding genotoxicity, the drug is presumed to be a transspecies carcinogen and a resource-intensive 2-year study is not likely to be warranted”。
7
抗癌产品和预期寿命有限的适应症
根据 ICH S918 和 ICH S1A，用于治疗晚期癌症的药物通常不需要进行致癌性研究；预期寿命短于两至三 年的其他严重适应症也可能不需要进行致癌性研究。
对于抗癌药物来说，这种例外情况似乎很奇怪；但这是基于这样一种认识，即许多癌症治疗方法也是 "癌症诱导剂"，下面举例说明。
• 放射治疗（如使用高能 X 射线、伽马射线或其他带电粒子）通过诱导 DNA 链断裂而起作用，这更有可能导致快速分裂的细胞死亡。不过，放疗也会增加将来罹患其他癌症的风险，这一点已得到公认。
• 烷化剂（如环磷酰胺、顺铂和丁硫胺）通过烷基的化学附着破坏癌细胞的 DNA，阻止细胞分裂，导致受影响细胞死亡。不过，这种治疗也会影响正常细胞（尽管影响程度较小），并可能在日后导致继发性癌症。
• 如果这种疗法能立即治疗现有的癌症，而且几乎没有或根本没有其他疗法可供选择，那么这些疗法和其他疗法增加的未来癌症风险通常被认为是可以接受的。但是，这种例外情况并不包括对无肿瘤患者使用的新辅助药物进行致癌性测试，也不包括对非癌症患者使用的新辅助药物进行致癌性测试。
对于仍需进行研究的严重疾病，致癌性研究通常可在药物获得上市批准后完成。同样，由于需要加快获得挽救生命或延长生命的疗法，延迟进行此类研究会略微增加风险。
8
组合产品
如果产品的单个成分已经分别进行了致癌和遗传毒性风险评估，且不存在新的令人担忧的原因，则一般不建议在临床试验或批准上市前对复合产品进行致癌和遗传毒性研究。
9
替代疗法
如果替代疗法（如胰岛素）的化学结构和生物活性与 "天然 "（‘natural’）对应物相当，而且在正常生理水平下使用，特别是类似产品已被确定为安全时，通常不需要进行额外的致癌试验。当新药偏离这些参数时，应评估是否需要进行额外研究。
10
配方和生产变更
微小的生产和配方变更需要逐案评估，以确定是否需要进行新的致癌性测试。不过，根据 ICH S1A 指南，如果现有测试支持当前制剂的安全性，且药代动力学、药效学或毒性没有发生重大变化，无需进行额外的衔接研究，则可能不需要进行额外的研究（additional bridging studies）。
11
儿科研究
ICH M3(R2)12阐述了在将儿科人群纳入临床试验之前进行额外研究的必要性。就致癌性研究而言，预计在开始长期研究之前将确定是否需要进行此类研究，但除非根据一般毒性、遗传毒性或机理研究的结果存在令人担忧的原因，否则没有必要进行此类研究。
12
生物技术产品
ICH S6(R1) 13 中讨论了生物技术产品临床前安全性评估指南。该文件指出，除非因接触时间长短、患病人群或生物作用机理而有必要，否则通常不适合对生物技术产品进行致癌性测试。后者非常重要，因为生物技术产品的毒性通常是由引人注目的药理作用而非 "脱靶 "毒性引起的（results from exaggerated pharmacology rather than ‘off-target’ toxicity）。例如，有意抑制免疫功能或促进细胞增殖的产品，如果反应过度，可能会致癌。如果潜在的疾病状态使病人对这些影响更加敏感，情况就更是如此。
如果担心有致癌性，则应进行WoE评估。评估的输入信息应包括已公布的有关一般药物类别的信息、相关实验室报告（例如，使用转基因动物或靶标缺陷动物）、靶标的生物学特性、作用机制，以及申办者可参考的任何可用细胞或动物模型数据。同样，如果评估结果支持该结论，也不需要进行额外的研究，在这种情况下，指南建议将重点转移到产品标签和风险管理方法上。当评估结果不明确时，可能需要在特征明确的细胞和动物模型中进行额外研究，以了解并在可能的情况下管理任何基于机制的问题。
假定该产品在啮齿类动物中无免疫原性且具有生物活性，在某些有限的情况下，申办者仍可考虑在相关动物物种或品系中进行单项研究。但是，不建议对同类产品进行研究，因为研究结果可能与人体使用无关。在这种情况下，阳性数据会给所有相关人员（包括监管审查人员）带来问题，而且进行该研究可能会占用申办者的宝贵资源，而更多的机理研究可以更好地促进整体风险评估。
13
有限的全身暴露
如果局部（如皮肤或眼部）使用的产品导致全身暴露程度较低，且施用部位不存在致癌活性问题（如光致癌性），则可能不需要进行广泛的测试。
14
光致癌性
根据ICH S10.19所述，在门诊研究之前应进行光毒性潜在评估，考虑的要点包括药物是否:
1.吸收自然光(290- 700nm)范围内的光
2.吸收紫外线-可见光后产生反应物质
3.充分分布于暴露于光下的组织(如皮肤、眼睛)
在发现潜在风险的情况下，ICH S10 还指出，现有啮齿动物模型的光致癌性测试不足以评估药物对人体的风险。相反，ICH M3(R2)建议在试验过程中采取保护措施（如在标签上和患者同意过程中发出适当警告、使用皮质类固醇和其他干预措施等）来控制任何潜在的光毒性风险。(本章较长，未完待续)
英文版原文
一、Chapter 8 Carcinogenicity Studies
First synthesized in 1938, Diethylstilbestrol (DES) is a synthetic form of estrogen that was prescribed to pregnant women (especially in the US, European countries such as the UK, and Australia) beginning in the 1940s to prevent miscarriage. It was not until 1971, however, that a landmark article published in the New England Journal of Medicine linked in utero DES exposure to a rare type of vaginal tumor in women whose mothers had received the drug during their pregnancy. This finding led US and European regulators to recommended that DES no longer be prescribed to pregnant women in 1971 (US), and 1978 (Europe), respectively. Today, DES is classed as a Group 1 carcinogen (i.e., carcinogenic to humans) by the International Agency for Research on Cancer (IARC), and its use has largely been phased out around the world.
The story of DES is an important cautionary tale for both drug developers and regulators alike. Although drug development has evolved extensively since DES was introduced, first by shifting the burden of safety testing from regulators to sponsors, and more recently with the advent and global harmonization of testing requirements, carcinogenicity testing is still considerably more than just a ‘box-checking’ exercise. This fact has been further highlighted in recent years with the introduction of a host of novel drug product classes, which have brought unique testing challenges. For example, despite their great promise, advanced cell and gene therapies have raised concerns regarding the longterm impacts of viral integration and the unintended expansion of stem cell populations. As these and other technologies mature into viable and widely available therapies, regulatory professionals will need to be ready to justify the various testing strategies that have been used for their development in their communications with regulators.
This chapter is intended to provide a broad overview of the current state of global regulatory requirements as they apply to carcinogenicity studies. It should be read in concert with the preceding chapter on assessing genotoxicity (see Chapter 7) since these assessments generally precede and inform the carcinogenicity testing plan and are leveraged in making overall assessments regarding carcinogenic potential. The discussion will start with a brief background on mechanisms of carcinogenicity as they relate to regulatory decision making. This will be followed by an introduction to the current state of global regulatory requirements and exemptions pertaining to the conduct and interpretation of carcinogenicity testing, largely from the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH).9–20 Finally, specific considerations related to model choice and study design are presented, followed by a discussion of how carcinogenicity data should be presented in marketing and premarket submissions to global regulatory bodies.
01
Cancer and Carcinogenicity
The World Health Organization (WHO) defines cancer as a group of diseases caused by the rapid proliferation “of abnormal cells that grow beyond their usual boundaries, and which can then invade adjoining parts of the body to spread to other organs” in a process known as “metastasis”. The initial growth or swelling is referred to as a tumor or neoplasm. Importantly, while neoplasms may be defined as being either malignant or benign or (i.e., cancerous or not cancerous), regulatory science has historically included drugs that cause a statistically significant increase in benign lesions within the definition of a “carcinogen” as well. This decision is practical since such growths can cause an unacceptable increase in risk through their ability to obstruct nearby tissues, resulting in pain and other medical conditions. The following is meant to provide an overview of several important concepts that are necessary for understanding carcinogenicity testing requirements from a regulatory perspective. Readers interested in a more holistic treatment of the topic are encouraged to seek out one of the many excellent texts that are available on the subject.
（1）Genotoxic and Nongenotoxic Carcinogens
In addition to defining neoplasms as benign or malignant, carcinogens can be classified as being genotoxic, nongenotoxic, or both dependent upon their mechanism(s) of action. Drugs that interact directly with DNA to cause damage to it are considered to be genotoxic, while drugs that cause cancer through indirect mechanisms are considered to be nongenotoxic.
• Genotoxic carcinogens are those which are mutagenic to DNA and have the potential to be ‘complete carcinogens’ (i.e., capable of initiation, promotion, and progression of cancer as described in the next section). Although a ‘safe’ dose threshold may exist for these drugs in some cases (e.g., aneugenic compounds, like colchicine),22,25 the default regulatory assumption is that they have a low-dose linear risk vs response curve (see Figure 8-1), such that even small amounts are capable of causing direct DNA damage that can culminate in cancer.
• In contrast, nongenotoxic carcinogens are epigenetic and act by engineering a microenvironment that makes cells more susceptible to DNA damage. This is generally accomplished by augmenting various concentration-dependent, cellular signaling processes (e.g., gene transcription, or affecting various receptors involved in growth and proliferation). As a result, the dose-response curve for these compounds is assumed to have a ‘safe’ dose threshold below which these processes are not initiated (Figure 8-1).
（2）Stages of Carcinogenesis
While carcinogenicity can arise from many different perturbations of normal cellular processes involved in the maintenance and growth of cells, it is helpful to think of carcinogenesis overall as a multi-phase process, with specific carcinogens contributing at one, or all stages (i.e., a ‘complete carcinogen’). These stages are known as initiation, promotion, and progression.
（3）Initiation
Initiation involves irreversible, heritable changes (including mutations and deletions) in the DNA sequence of cells. As initiating agents are mutagenic, they are also considered to be genotoxic. The majority of drugs that cause carcinogenic changes at the initiation stage are pro-genotoxins because they require some amount of internal activation or biotransformation (see Chapter 6) before they can interact with DNA. For this reason, early genotoxicity assessments (see Chapter 7) often include an S9 microsomal fraction as this mixture contains important liver enzymes that allow the formation of the direct-acting toxic metabolite of the drug in cellular (i.e., in vitro) systems. Importantly, initiating events in themselves are not guaranteed to result in cancer. In many cases,these changes just make the cell non-viable or otherwise unable to proliferate. But presuming that the initiating event’s effect is sustainable, the next step in carcinogenesis is promotion.
（4）Promotion
Promotion is the general process by which the initiated cells undergo clonal expansion to produce a preneoplastic lesion. Carcinogenicity at this stage is generally nongenotoxic, acting through a myriad of epigenetic mechanisms including hormone modulation, irritation and inflammation, immunosuppression, and inhibition of cellular communication pathways.27,28 Given the indirect nature of their effects, carcinogens that encourage the promotion phase of the carcinogenic process are assumed to have a ‘safe’ threshold as described earlier. Furthermore, in contrast to initiation and progression, promotion is generally thought to be reversible.22,26 The carcinogen DES, which was mentioned at the beginning of this chapter, primarily acts as a tumor promoter by inducing estrogen receptor mediated increases in cellular proliferation.29
（5）Progression
The final stage of carcinogenesis is known as progression. During this stage, successive genotoxic changes result in accumulation of additional mutations within individual tumor cells. These processes ultimately select for clones with the most aggressive traits (e.g., proliferation), and this in turn culminates in the irreversible development of a neoplasm. Similar to initiating agents, agents that target progression are typically genotoxic, and thus, are assumed to not have a ‘safe’ threshold for their use. 22
02
Overview of Regulatory Landscape
Since it was initially conceived by the National Cancer Institute (NCI) in the mid-1970s, the 2-year rodent bioassay for carcinogenicity has become the gold standard for estimating the carcinogenic potential for new drug products.30 Although requirements for long-term studies were already in place in founding member regions of the ICH (i.e., the US, EU countries, and Japan) by the time the ICH guidelines for carcinogenicity studies were developed in the 1990s, global adoption of these guidance enabled alignment on many specific aspects of the assay. These aspects included high dose selection (which previously allowed use of high dosage multiples of greater than 100 times on a milligram per kilogram basis as an alternative to the maximum tolerated dose(MTD) in the EU and Japan, but not the US), and the minimum duration of clinical exposure to justify long-term studies (previously 3 months in the US versus 6 months in the EU and Japan). Through harmonization of these and other expectations, the ICH S1 guidances9 sought to reduce redundancies in the global drug development process.
This alignment notwithstanding, the 2-year assay is not now, nor has it ever been, a simple undertaking. There are several aspects of its design that set it apart from many of the traditional tests in the nonclinical testing battery. In particular, the long duration and high number of animals (i.e., at least 50 animals of each sex per group) required for performing two, traditional, 2-year rodent carcinogenicity studies (typically in rat and mouse models) make them uniquely resource intensive, and therefore costly for sponsors.
The ICH S1 guidance was conceived with the understanding that regulatory expectations must be allowed to evolve alongside the industry’s understanding of cancer itself. Perhaps the most dramatic update in recent years has been the finalization of the ICH S1B(R1) addendum.10 This addendum describes an integrated weight of evidence (WoE) assessment to aid sponsors in determining whether a 2-year rodent carcinogenicity study is likely to add value to the overall carcinogenicity testing approach. When it does not add value, the sponsor can now seek regulatory consultation to support the decision to exclude 2-year studies entirely.
As outlined in Table 8-1, relevant guidance for carcinogenicity testing has been implemented almost universally by ICH member countries. It is also widely recommended by regulators in numerous ICH observer and non-member countries.
（1）Requirements for Carcinogenicity Testing
The general expectations outlined in the ICH S1B(R1) guidance10 are for one long-term study (typically a 2-year study in rats), and either a second long-term study (typically in mice) or a short or medium length in vivo (i.e., animal) study. The most commonly used shorter length studies include rodent models of initiation and promotion of cancer, as well as studies using transgenic (e.g., p53+/-, the Tg.AC, Tg.rasH2, and Xpa-deficient) and neonatal rodents.
（2）Factors for Consideration
Historically, exposure duration has been the primary determinant of whether studies are required, with studies being needed for products that are intended to be taken for longer than 6 months, or intermittently for an extended period of time.
Specific ‘cause for concern’ is another important consideration. According to ICH guideline S1A,9 causes for concern may include:
1.Previous demonstration of carcinogenic potential in the product class that is considered relevant to humans,
2.Structure-activity relationship suggesting carcinogenic risk,
3.Evidence of preneoplastic lesions in repeated dose toxicity studies, and
4.Long-term tissue retention of parent compound or metabolite(s) resulting in local tissue reactions or other pathophysiological responses.
Notably, while extended dosing beyond 6 months is considered to be justification for requiring carcinogenicity studies, it is not considered to be a ‘cause for concern’ in itself per ICH guidance.9
（3）Additional Mechanistic Studies
While not a specific requirement, additional mechanistic studies may be necessary to understand any results obtained from carcinogenicity testing. Such studies may focus on various endpoints, including:
• Cellular changes (e.g., apoptosis or proliferation),
• Biochemical changes (e.g., growth factors and plasma levelsof specific hormones),
• Additional genotoxicity assessments when the standard battery is negative and an epigenetic mechanism cannot be discerned, and
• Other modified protocols to further characterize an effect (e.g., reversibility, dosing interruptions, etc.).
The primary goal of these additional mechanistic studies is to better characterize a drug’s mechanism of action and its relationship with potential human risk. For example, while drugs that agonize the peroxisome proliferator-activated receptor α are known to induce cancer in some rodents, this mechanism has been shown to be irrelevant for assessing human cancer risk.31
（4）Timing in a Development Program
Carcinogenicity testing should follow genotoxicity testing as discussed in ICH S2(R1).11 It should also take the intended patient population and dosing regimen into consideration, as well as the results of early pharmacokinetic and pharmacodynamic analyses (see Chapter 6), and repeat-dose toxicity testing results (including independent dose-range finding studies as needed; see Chapter 8).
This information can be extremely valuable for not only determining if preneoplastic lesions or other important changes (e.g., immunosuppression or cellular proliferation) are present, but also for understanding the mechanistic basis of any findings, including whether they are relevant for cancer development in humans. Assuming there is no ‘cause for concern,’ the general expectation for carcinogenicity studies is that they be completed before applying for marketing approval, but not necessarily before initiating large-scale (i.e., Phase 3) clinical trials.9,12
（5）Potential Exceptions to Regulatory Requirement Expectations
There are multiple exceptions to the general regulatory requirements. These may be based on current scientific understanding of specific products or product classes; or on careful assessment of any risks that may occur with deviation from the normal timeline,data expectations, or both, in light of any potential benefits.
（6）‘Unequivocally’ Genotoxic Drugs
While it might be expected that obtaining unequivocally positive results (i.e., based on the findings of an entire battery of tests rather than a single test) from genotoxicity studies would inevitably require long-term carcinogenicity studies, this is not the case.
The S1B(R1) guidance,10 states that a “2-year rat study is less likely to be of value either in cases whether there is no genotoxicity risk, or in cases with unequivocal genotoxicity risk.” In other words, where already sufficient information exists to draw a conclusion regarding genotoxicity, the drug is presumed to be a transspecies carcinogen and a resource-intensive 2-year study is not likely to be warranted.
（7）Anti-Cancer Products and Indications with Limited Life Expectancy
According to ICH S918 and ICH S1A,9 carcinogenicity studies are not typically required for drugs being developed to treat advanced cancers; they may also not be required for other serious indications where the life expectancy is shorter than 2 to 3 years.
This exception may seem strange in the case of anti-cancer drugs; however, it is based on the understanding that many cancer treatments are also ‘cancer inducers’, as explained in the following examples.
• Radiation therapy (e.g., using high-energy X-rays, gamma rays, or other charged particles) works by inducing strand breaks in DNA, which is more likely to cause death in rapidly dividing cells. However, it is well-established that radiation therapy also increases the risk of developing other cancers in the future.32
• Alkylating agents (e.g., cyclophosphamide, cisplatin and busulfan) damage DNA in cancer cells by the chemical attachment of an alkyl group, preventing cell division and resulting in the death of the affected cells. Such treatment can affect normal cells (albeit to a lesser extent), however, and may result in secondary cancers at a later date.
• The increase in future cancer risk for these and other therapies is often considered acceptable if the treatment provides an immediate benefit for the existing cancer, and where there are few or no other therapeutic options available. However, this exception does not extend to carcinogenicity testing of new adjuvant drugs for use in tumor-free patients, or usage for noncancer conditions.
For serious conditions where studies are still needed, carcinogenicity studies can often be completed after a drug has obtained marketing approval. Again, the slight increase in risk associated with the delay in performing such studies is offset by the need to expedite access to lifesaving or life extending therapies.
（8）Combination Products
Carcinogenicity and genotoxicity studies are not generally recommended for combination products prior to clinical trials or marketing approval if the individual components of the products have already been assessed for these risks separately, and where no new cause for concern exists.
（9）Replacement Therapies
Additional carcinogenetic testing is typically not needed for replacement therapies (e.g., insulin) when there is a comparable chemical structure and biological activity to the ‘natural’ counterpart, and when the therapy is given within normal physiological
levels, especially when similar products have already been determined to be safe. The need for additional study should be assessed when a new drug deviates from these parameters.
（10）Formulation and Manufacturing Changes
Minor manufacturing and formulation changes need to be assessed on a case-by-case basis to determine if new carcinogenicity testing is needed. However, as per the ICH S1A guidance,9 additional studies may not be needed where existing testing supports
the safety of the current formulation, and where no significant changes in pharmacokinetics, pharmacodynamics, or toxicity exist that would necessitate additional bridging studies.
（11）Pediatric Studies
ICH M3(R2)12 addresses the need for performing additional studies before including pediatric populations in clinical trials. In terms of carcinogenicity studies, the expectation is that their need will be determined before beginning long-term studies, but that they are not necessary unless there is cause for concern based on the results of general toxicity, genotoxicity, or mechanistic studies.
（12）Biotechnology Products
Guidance on the preclinical safety evaluation of biotechnology products is discussed in ICH S6(R1).13 The document states that carcinogenicity testing is usually not appropriate for biotechnology products unless necessitated by duration of exposure, disease population, or biological mechanism of action. The latter is important because toxicity with biotechnology products often results from exaggerated pharmacology rather than ‘off-target’ toxicity. For example, products that intentionally suppress immune function or promote cellular proliferation could cause carcinogenesis if these responses are excessive. This is particularly true in cases where the underlying disease state makes patients more sensitive to such effects.
A WoE assessment should be carried out if there are concerns about carcinogenicity. Inputs for the assessment should include published information regarding the general drug class, relevant laboratory reports (e.g., using transgenic or target-deficient
animals), the biology of the target, the mechanism of action, and any available cell or animal model data the sponsor can reference.Additional nonclinical work is generally not needed if the WoE does not support a finding of carcinogenicity. Similarly, additional studies are not warranted if the assessments support that finding, in which case guidance recommends the focus shift to product labelling and risk management approaches. When the outcome of the assessment is unclear, additional studies in well-characterized cell and animal models may be appropriate to understand and possibly manage any mechanism-based concerns.
Presuming that the product is non-immunogenic and is biologically active in a rodent species, the sponsor might still consider conducting a single study in a relevant animal species or strain in some limited circumstances. However, studies with homologous products are not recommended as the results are not likely to be relevant to human use. Positive data in this scenario would be problematic for all involved (including regulatory reviewers), and performing the study would potentially divert valuable sponsor resources from more mechanistic studies that could better contribute to the overall risk assessment.
（13）Limited Systemic Exposure
Broad testing may not be needed in cases where a topically (e.g., dermal or ocular) administered product results in poor systemic exposure and there is no concern for carcinogenic activity at the application site (e.g., photocarcinogenicity).
（14）Photocarcinogenicity
An evaluation of phototoxic potential should be performed before outpatient studies as outlined in ICH S10.19 Points for consideration include whether the drug:
1.Absorbs light within the range of natural sunlight (290-700 nm)
2.Generates a reaction species following absorption of ultraviolet-visible light
3.Distributes sufficiently to light-exposed tissues (e.g., skin, eye)
In cases where a potential risk is identified, ICH S10 also states that photocarcinogenicity testing in available rodent models is not adequate enough for assessing the risk of drugs in humans.19 Instead, ICH M3(R2) recommends that any potential phototoxicity risk be managed during trials by protective measures (e.g., appropriate warnings on the label and during patient consent, treatment with corticosteroids and other interventions, etc.)
参考文献：
22. Klaassen CD, ed. Casarett & Doull’s Toxicology: The Basic Science of Poisons. 9th ed. McGraw-Hill Education; 2019. Verified October 2022.
23. Graziano M, Jacobson-Kram D, ed. Genotoxicity and Carcinogenicity Testing of Pharmaceuticals. Springer International Publishing AG; 2015. Verified 24 October 2022.
24. Weinberg, RA. The Biology of Cancer. 2nd ed. W.W. Norton & Company Inc.;2013. Verified 24 October 2022.
声明：原英文原文版权归美国法规事务协会（RAPS）所有，本译文供参考，如有任何建议，请联系我们。
更多优质内容，欢迎关注↓↓↓
