病理学完全缓解与长期结局——2016年我们知道什么？

　　2016年5月11日，英国《柳叶刀肿瘤学分册》在线发表了意大利、波兰、韩国、中国台湾、西班牙、巴西、加拿大、俄罗斯、泰国、英国（罗氏、阿斯利康）的多中心非盲Ⅱ期随机研究（NeoSphere）5年分析：帕妥珠单抗和曲妥珠单抗新辅助治疗局部晚期、炎性或早期HER2阳性乳腺癌患者。

　　该研究发现，虽然随访5年无进展生存和无病生存的置信区间太大且重叠，但是支持主要终点（病理学完全缓解）并表明曲妥珠单抗和多西他赛联合帕妥珠单抗新辅助治疗是有益的。此外，总的病理学完全缓解可能是早期HER2阳性乳腺癌长期预后的早期指标。

　　对此，奥地利维也纳医科大学综合癌症中心的乳腺内科和外科专家发表了同期述评：病理学完全缓解与长期结局——2016年我们知道什么？

Lancet Oncol. 2016 May 11. [Epub ahead of print]

Pathological complete remission and long-term outcome—what do we know in 2016?

Michael Gnant, Guenther G Steger, Rupert Bartsch.

Breast Health Centre, Comprehensive Cancer Centre, Medical University of Vienna, Vienna, Austria.

The tale of HER2-positive breast cancer is a story of success. From the identification of HER2 protein overexpression as strong predictor of poor outcome, to the introduction of the HER2-directed antibody trastuzumab in the metastatic setting, and to the adjuvant trastuzumab trials, major improvements in the outcome of patients with HER2-positive breast cancer have been achieved.

In parallel to the overall evolution of anti-HER2 therapy, neoadjuvant chemotherapy became a widely adopted treatment strategy in early-stage breast cancer. In addition to the obvious surgical benefit of preoperative tumour response, pathological complete remission has been established as a putative surrogate parameter for long-term outcomes. In the NSABP trial B-27, sequential administration of doxorubicin and cyclophosphamide followed by docetaxel doubled the proportion of patients achieving pathological complete remission, compared with doxorubicin and cyclophosphamide alone. Moreover, patients achieving pathological complete remission had better survival results than those who did not, which gives reason to hope that further increase in pathological complete remission could yield even better long-term outcomes.

A subsequent study showed that trastuzumab could be successfully added to neoadjuvant chemotherapy in HER2-positive disease, and is now considered a standard of care. In metastatic breast cancer, the addition of pertuzumab, another HER2-antibody blocking HER2 hetero-dimerization, to trastuzumab plus docetaxel further improved progression-free survival and overall survival compared with chemotherapy plus trastuzumab alone. The NeoSphere trial investigated the potential role of pertuzumab in the neoadjuvant setting, and the primary results of this randomised four-arm phase 2 study have already been published. A significantly higher proportion of patients treated with four cycles of docetaxel plus trastuzumab and pertuzumab achieved pathological complete remission (46%) compared with patients treated with either docetaxel plus trastuzumab (29%) or docetaxel plus pertuzumab (24%), or the combination of both antibodies without chemotherapy (17%). These results eventually led to the approval of dual HER2 inhibition using trastuzumab and pertuzumab in the neoadjuvant setting.

Despite this success, the question remained whether these pathological complete remission improvements would eventually result in improved long-term outcome. While a meta-analysis indicated that, on the level of individual patients, pathological complete remission is a reliable predictor of favourable outcome in high-risk breast cancer subtypes, this does not necessarily mean that achievement of pathological complete remission will automatically yield superior disease-free survival results in comparative trials. Indeed, this putative correlation between pathological complete remission and disease-free survival is a matter of ongoing debate, and the 5-year analysis of NeoSphere published in The Lancet Oncology adds an important piece of information. However, it must be noted that these results arise from an exploratory analysis that is not powered for formal statistical hypothesis testing.

The achievement of total pathological complete remission (pathological complete remission in breast and axilla) predicted improved 5-year progression-free survival, defined as the interval from breast cancer diagnosis to a progression event compared with patients who did not achieve total pathological complete response (85% [95% CI 76-91] vs 76% [71-81], HR 0.54 [95% CI 0.29-1.00]). This is reassuring for both patients and clinicians, and strongly suggests that individual total pathological complete remission is a reliable predictor of outcome.

When analysing the long-term outcomes in the trial group, the 5-year disease-free survival of patients treated with docetaxel plus trastuzumab and pertuzumab was 84% (95% CI 72-91) compared with patients on docetaxel plus trastuzumab (81% [95% CI 72-88]; HR 0.60 [95% CI 0.28-1.27]). Does this marginal and statistically not significant improvements in disease-free survival suggest that, on the trial level, no correlation exists between pathological complete remission and disease-free survival?

In this context, mathematics must be considered. With a 20% increase in pathological complete remission, an improvement in event-free survival of approximately 20% (HR 0.8) might be expected in HER2-positive breast cancer. Furthermore, Hatzis and colleagues showed that even with pathological complete remission hypothetically increasing from 30% to 60%, a trial would require 3550 patients per arm in order to detect a significant overall survival difference, if 10-year recurrence-free survival was 74% in the control arm.

To summarise, what have we learned from this analysis of NeoSphere? The data on long-term outcome again emphasise that patients with total pathological complete remission have a favourable outcome, and we believe that this alone justifies continuing the quest for increasing pathological complete remission. We remain less likely to achieve the important goal of being able to fully predict the adjuvant application of an innovative treatment from neoadjuvant phase 2 studies. Also, our goal of de-escalating therapies for selected patients remains unresolved and urgently needs further translational work. Nonetheless, the results of the pivotal NeoSphere trial are in line with data predicted in mathematical models, and support the existence of an association between achievement of pathological complete remission and improved long-term outcome in high-risk breast cancer subtypes.

DOI: 10.1016/S1470-2045(16)30038-9