妊娠相关乳腺癌定义已过时并应弃用
2021年6月1日，英国《柳叶刀》肿瘤学分册在线发表比利时鲁汶大学、荷兰癌症研究所、美国科罗拉多大学、德克萨斯大学MD安德森癌症中心、哈佛大学医学院达纳法伯癌症研究所、俄勒冈医科大学、意大利热那亚大学圣马蒂诺综合医院、米兰大学欧洲癌症研究院、德国乳腺癌协作组、法兰克福癌症中心的述评：妊娠相关乳腺癌定义已过时并应弃用。
目前，妊娠相关乳腺癌定义并不统一，有些定义为完全在妊娠期诊断出的乳腺癌，有些定义还包括出生后6个月至1年内诊断出的乳腺癌。虽然妊娠期与产后期密不可分，但是越来越多的证据支持将妊娠期乳腺癌与产后期乳腺癌（根据最新数据，可延长至产后5～10 年）视为独立且不同的乳腺癌，因为两种乳腺癌的生物学特征和预后有所不同。
妊娠期乳腺癌是在妊娠期被诊断出来的，大约占45岁以下女性全部乳腺癌病例的4%。乳腺癌是年轻女性妊娠期最常见的癌症类型之一，根据估计每10万次活产大约发生5.1例。妊娠期乳腺癌与非妊娠期乳腺癌相比，预后是否较差？现有文献的看法存在分歧。虽然已知妊娠期乳腺癌的基因存在差异且总生存率较低，但是妊娠期乳腺癌的大型队列研究发现，与年轻非妊娠期女性乳腺癌相比，肿瘤临床特征和预后相似。
产后期乳腺癌发生于妊娠后5～10年内，大约占45岁以下女性全部乳腺癌病例的35%～55%。产后期乳腺癌与曾妊娠或妊娠期年轻绝经前女性乳腺癌相比，生存率较低且转移风险增加两倍以上。对诊断年龄、诊断年份、分期、分级、激素受体状态等临床病理因素进行校正后，预后仍然不佳。重要的是，当根据传统妊娠相关乳腺癌定义对相同数据进行分组时，与生物学不同的妊娠期乳腺癌和产后期乳腺癌定义相反，上述结局差异消失。就逻辑而言，如果将高风险女性（例如产后期乳腺癌女性）纳入研究作为对照，那么在任何已知的分析中，可能都无法发现既往分娩的生物学影响。
图、导致产后期乳腺癌临床结局较差的细胞相互作用示意图
A、妊娠相关乳腺癌的定义导致临床结局数据互相矛盾。当妊娠相关乳腺癌病例被区分为妊娠期乳腺癌和产后期乳腺癌时，该模型中的不良预后尤其适合产后期乳腺癌。该区别可以解释已发表研究的预后差异。
B、妊娠期和哺乳期，乳腺发生增殖和分化，据称可导致妊娠对乳腺癌的长期预防作用。产后期乳腺癌的短期风险增加归因于乳腺退化，该过程随后通过细胞凋亡清除乳腺上皮细胞，将乳腺恢复至妊娠前状态（缓慢断奶）。退化与促肿瘤微环境具有许多共同的基质（间质）特征，包括免疫抑制，这可能导致浸润性较强的侵袭性肿瘤形成（快速断奶）。
自从妊娠相关乳腺癌的定义出现以来，关于妊娠期乳腺癌与产后期乳腺癌之间的细胞和分子差异（包括其对应组织微环境，图B）已有数据越来越多。妊娠期乳腺上皮发生增殖和分化，为哺乳做准备。分娩后和未哺乳或断奶期间，乳腺通过退化过程重塑至形态和功能类似妊娠前的状态。根据推测，当存在亚临床病变时，退化可增加产后期乳腺癌的转移风险。在对人类乳腺退化深入研究的支持下，动物研究表明乳腺退化微环境可发生类似伤口愈合的变化，免疫抑制细胞流入还可促进肿瘤细胞从乳腺逃逸。对于产后期乳腺癌患者，免疫浸润方式、细胞因子特征的变化都可在分娩后的原发肿瘤微环境中持续数年。如图B所示，亚临床病变、伤口愈合样基质重塑、免疫耐受细胞浸润之间的相互作用，可以解释产后期乳腺癌与妊娠期乳腺癌相比预后较差以及在一个定义下两种乳腺癌研究结果互相矛盾的原因。
生殖因素与乳腺癌之间的关系十分复杂，因此很难区分未生育、单胎、多胎、初产年龄、哺乳、退化、遗传因素、乳腺肿瘤本身的相关机制。不过，越来越明确的是，应该区分妊娠期发生的乳腺癌和产后5～10年内发生的乳腺癌。在临床实践中，该差异显而易见：妊娠期乳腺癌的治疗应根据胎龄个体化并考虑胎儿安全，而产后期乳腺癌的治疗决策不需要考虑这些问题。此外，需要深入了解产次状态作为绝经前乳腺癌结局独立预后因素的重要性。只有对这两种乳腺癌分别进行研究，才能提高对妊娠、哺乳、退化以及此后乳腺癌生物学的认识，并有助于破解肿瘤生物学差异的具体机制。
因此，作者推荐不再采用妊娠相关乳腺癌定义，让研究者能够分别关注妊娠期乳腺癌或产后期乳腺癌，这可能最终带来治疗模式的优化，尤其对于预后不佳的产后期乳腺癌。
Lancet Oncol. 2021 Jun 1;22(6):753-754.
The definition of pregnancy-associated breast cancer is outdated and should no longer be used.
Frédéric Amant, Hanne Lefrère, Virginia F Borges, Elyce Cardonick, Matteo Lambertini, Sibylle Loibl, Fedro Peccatori, Ann Partridge, Pepper Schedin.
KU Leuven, Leuven, Belgium; Netherlands Cancer Institute, Amsterdam, Netherlands; University of Colorado Anschutz Medical Campus, Aurora, CO, USA; MD Anderson Cancer Center at Cooper, Camden, NJ, USA; Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA; University of Oregon Health & Science University, Portland, OR, USA; University of Genova, Genoa, Italy; IRCCS Ospedale Policlinico San Martino, Genoa, Italy; European Institute of Oncology IRCCS, Milan, Italy; German Breast Group, Neu-Isenburg, Germany; Centre for Haematology and Oncology Bethanien, Frankfurt, Germany.
The definition of pregnancy-associated breast cancer is inconsistently given as either breast cancer diagnosed exclusively during pregnancy, or combined with cases diagnosed within 6 months to 1 year after the birth. Although pregnancy and the postpartum period are intertwined, evolving evidence supports considering breast cancer that occurs during pregnancy as a separate and distinct entity from breast cancer that occurs during the postpartum period (which, according to newer data, can extend to 5-10 years after the birth) because each type has unique biological attributes and prognosis.
Breast cancer during pregnancy is diagnosed during pregnancy and represents about 4% of cases of breast cancer in women younger than 45 years. With an estimated incidence of 5.1 cases per 100000 livebirths in young women, breast cancer is one of the most common cancer types occurring during pregnancy. The existing literature gives a mixed view of whether a breast cancer diagnosis during pregnancy confers a worse prognosis than breast cancer not diagnosed during pregnancy. Although genetic differences and reduced overall survival in breast cancer during pregnancy have been described, large cohort studies of breast cancer during pregnancy find that clinical tumour characteristics and prognosis are indistinguishable from breast cancers occurring in young, non-pregnant women.
Postpartum breast cancer occurs within 5-10 years after pregnancy and is estimated to represent 35-55% of all cases of breast cancer in women younger than 45 years. Postpartum breast cancer is associated with worse survival rates and with a more than two times increased risk of metastasis than breast cancer diagnosed in young, premenopausal women during pregnancy or who have ever been pregnant. Poor prognosis persists after adjustment for several clinicopathological factors, including age at diagnosis, year of diagnosis, stage, grade, and hormone receptor status. Importantly, when the same data were grouped in the traditional pregnancy-associated breast cancer definition, as opposed to the biologically distinct proposed breast cancer during pregnancy and postpartum breast cancer definitions, this difference in outcomes was lost. Logically, if women at increased risk, such as women with postpartum breast cancer, are included in a study as controls, the biological effect of previous childbirth is likely to go unidentified in any given analysis.
Since the conception of the definition of pregnancy-associated breast cancer, more data on the cellular and molecular differences between breast cancer during pregnancy and postpartum breast cancer (including their respective tissue microenvironments [figure, B]) have become available. During pregnancy, the mammary gland epithelium undergoes proliferation and differentiation in preparation for lactation. After childbirth and in the absence of lactation, or during weaning, the gland remodels to a state that is morphologically and functionally similar to the prepregnancy state, via a process called involution. It is hypothesised that involution, in the presence of subclinical disease, increases the metastatic potential of postpartum breast cancer. Animal studies, supported by detailed studies of human breast involution, show wound healing-like alterations in the microenvironment of the involuting breast and an influx of immunosuppressive cells that promote the escape of tumour cells from the mammary gland. In patients with postpartum breast cancer, patterns of altered immune infiltration, cytokine profiles, or both can persist in the primary tumour microenvironment for several years after delivery. As depicted in figure B, the interaction between subclinical disease, wound healing-like stromal remodelling, and infiltration of tolerant immune cells could explain the worse prognosis of postpartum breast cancer compared with breast cancer during pregnancy, as well as the conflicting results observed in studies that consider both entities under a single definition.
The association between reproductive factors and breast cancer is complex, making it hard to differentiate between mechanisms related to nulliparity, uniparity, multiparity, age at first birth, lactation, involution, hereditary factors, and the breast tumour itself. Nevertheless, it is becoming clear that a distinction should be made between breast cancers that occur during pregnancy and those that occur within 5-10 years postpartum. In clinical practice, this difference is obvious: the treatment of breast cancer during pregnancy is individualised according to gestational age and taking foetal safety into consideration, whereas treatment decisions for postpartum breast cancer do not need to account for these concerns. Moreover, there is a need for a better understanding of the importance of parity status as an independent prognostic factor for worse outcomes in premenopausal breast cancer. Only a separate investigation of both entities will improve our understanding of the biology of breast cancer during pregnancy, lactation, involution, and thereafter, and help to decipher the pathways underlying differences in tumour biology. Therefore, we recommend that the term pregnancy-associated breast cancer is no longer used, allowing investigators to focus specifically on breast cancer during pregnancy or during the postpartum period, which could ultimately lead to optimised therapeutic modalities, particularly for postpartum breast cancer with a poor prognosis.
DOI: 10.1016/S1470-2045(21)00183-2