Abstract P5-20-11: Comparative effectiveness of neoadjuvant therapy for HER2-Positive breast cancer: Addition of new clinical evidence to network meta-analysis and data update after 5 years

Background: It is becoming more popular to perform neoadjuvant chemotherapy including anti-HER2 agents to operable HER2-positive breast cancer patients. Increasing HER2-targeted treatment options urge us to define the best neoadjuvant therapy. In 2014, we reported the systematical assessment of the efficacy and safety of neoadjuvant therapy for HER2-positive breast cancer, using network meta-analysis based on Bayesian model (Nagayama et al., JNCI 2014). Network meta-analysis synthesizes information from a network of trials, which helps interpret the randomized evidence and can rank treatments from different trials. After five years from our first literature search, we decided to update our analysis due to accumulation of new clinical evidence.

Methods: We assessed odds ratio for pathological complete response (pCR), completion, and safety in seven treatment arms utilizing pooling effect sizes. The treatment arms included the combinations of chemotherapy (CT), trastzumab (tzmb), lapatinib (lpnb) and pertzumab (pzmb). All statistical tests were two-sided, and we followed Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA) guidelines.

Results: A database search identified 993 articles with 13 studies meeting the eligibility criteria, adding three studies (a trial of CT + tzmb vs CT + lpnb, and two trials of CT + tzmb vs CT + lpnb vs CT + tzmb + lpnb) to previous analysis. In direct comparison, CT + tzmb significantly achieved more pCR than CT + lpnb (OR=0.68, 95% CI = 0.52 to 0.89, p=.005) despite no statistical difference was found previously. In indirect comparison, treatment arms of dual anti-HER2 agents with CT achieved more pCR than other arms, reducing their credibility intervals against all other arms. This trend was stronger in CT + tzmb + lpnb arm (CT + tzmb + lpnb vs CT + tzmb, OR = 1.62, 95% CrI = 1.19 to 2.22, p = .003), which we added sufficient clinical evidence. Moreover, it exposed the need for additional clinical data for pzmb relative arms. Values of surface under the cumulative ranking (SUCRA) suggested that CT + tzmb + pzmb had the highest probability of being the best treatment arm for pCR (SUCRA = 0.95), followed by CT + tzmb + lpnb (SUCRA = 0.87), and CT + tzmb (SUCRA = 0.62), widening the gap and differentiating the top two dual blockade arms which were close in our previous report. All outcomes from our present analysis were consistent with our previous report and strengthened data solidity by reducing confidence or credibility intervals.

Conclusion: Consistent results in not only in pCR but also in completion rates and adverse events indicate that we are looking at the results which are close to the truth. Additional trials of lpnb relative regimens are not probable to change the results, but pzmb relative trials are required to improve evidence solidity. New clinical data established stronger evidence in network meta-analysis that combining two anti-HER2 agents with CT is most effective in the neoadjuvant setting for HER2-positive breast cancer.

Citation Format: Nakashoji A, Hayashida T, Yokoe T, Maeda H, Watanuki R, Kikuchi M, Seki T, Takahashi M, Abe T, Kitagawa Y. Comparative effectiveness of neoadjuvant therapy for HER2-Positive breast cancer: Addition of new clinical evidence to network meta-analysis and data update after 5 years [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P5-20-11.