To our knowledge, this is the first meta-analysis to evaluate the impact of breast surgery on OS and QoL outcomes in patients with dnMBC. The results suggested that breast surgery had no beneficial outcome compared with no surgery in terms of OS or QoL in patients with dnMBC. In addition, we found that breast surgery had a significant advantage in LPFS, but no benefit in DPFS. Overall analysis showed that breast surgery was not significantly associated with improvement in OS, DPFS, or QoL in dnMBC patients. Subgroup analysis indicated that findings were generally conformed, regardless of timing of surgery, site and number of metastases and tumor molecular subtype.
Our study supports several moderate quality findings.55 The meta-analysis of prospective randomized trials yielded results opposite those of retrospective studies. We found that locoregional control was improved by breast surgery in all trials, but without improvement in OS and QoL.
Prior meta-analyses have reported on the association between breast surgery and disease outcome in patients with dnMBC.30,31,34−40However, our analysis differentiates itself from prior studies by focusing on RCTs using OS, PFS and QoL as study endpoints, to explore whether breast surgery is associated with disease outcome in the dnMBC population. Our results showed that breast surgery for therapeutic rather than palliative intent did not improve OS or QoL in these patients. We found that no significant difference between the two groups in terms of OS (HR = 0.87), and QoL (SMD=-0.26), although LPFS (HR = 0.36) in the breast surgery group was longer. Furthermore, 3-year OS was reported by five RCTs, and the pooled analysis failed to demonstrate any significant difference in OS. In addition, a pooled analysis of 3 studies that reported 2-year OS showed the same results.24–26 Interestingly, we found that breast surgery was significantly associated with improved LPFS in dnMBC patients. Obviously, the result of LPFS is tantamount to locoregional control. A major reason for the benefit of breast surgery on LPFS may be the reduction of overall tumor burden after primary tumor resection.56However, the effect of breast surgery on DPFS could be explained by an enhanced metastatic growth after tumor resection, which has been demonstrated in animal and human studies.57–59 There was substantial heterogeneity across studies in the differences in clinical data between the two groups of patients with dnMBC. For example, two studies randomization to breast surgery occurred after achieving clinical benefit from ST,25,28but in three other studies, prior breast surgery was followed by ST.24,26,29 The heterogeneity may stem from differences in the choice of treatment sequences among different populations in the surgery group.
Focusing on different subgroups, we did not verify that breast surgery can significantly benefit on OS. However, it is possible certain subgroups of patients may beneft more or less from breast surgery. For example, patients with solitary bone metastasis appear to benefit from breast surgery.29A recently published prospective study demonstrated that breast surgery improves the survival rates of dnMBC patients with bone metastases, particularly among women with solitary bone metastasis.3Another retrospective study also showed that breast surgery may improve prognosis of the patients with oligometastasis.60However, a previously study reported no improvement in OS for patients with bone metastases who underwent breast surgery.61 Another subgroup showed that breast surgery performed before ST seemed to yield a better OS. This finding is consistent with the results reported by Karagiannis and colleagues.62Unfortunately, we did not have sufficient trial-level data to perform a number of planned subgroup analyses. The TSA found that the information size was not enough to exclude an important effect with the intervention.Therefore, future studies need to focus on these specific populations, with a large enough sample size, including dnMBC patients undergoing breast surgery in specific clinical settings, such as oligometastatic disease or the sole active primary site remaining after ST.
In our study, three trials examined whether breast surgery improved QoL of dnMBC patients compared with no surgery, and nearly unanimous conclusions from all three studies indicated that breast surgery did not significantly improve QoL of patients. QoL outcomes as predictors for OS in breast cancer have been described,63and we chose QoL scores as the study endpoint. Results based on two eligible studies showed that breast surgery was not associated with improved QoL in patients with dnMBC. In MF07-01Q study,53as a secondary outcome of the MF07-01 study,27they compared the effects of breast surgery versus ST on QoL in dnMBC patients. There was no difference in QoL scores among patients who received breast surgery and ST only at least 36 months after randomization. Posytive trial,54concluded that breast surgery did not improve or alter QoL. Equally, E2108 study,28 showed that breast surgery had no overall impact on QoL in patients with dnMBC. Clearly, these results are consistent with our findings, which are supported by several studies.64,65
Compared with previous meta-analyses, the present study analyzed RCTs and included QoL data. The pooled QoL outcomes data is the first to be reported, and may have implications for assessing the effects of breast surgery in dnMBC population. Moreover, TSA was used in our meta-analysis.52 Nevertheless, this study has several limitations. First, relatively small sample size and number of studies are limitations of this study. Second, our review confirms high heterogeneity among randomized trials evaluating breast surgery. We attempted to determine the source of heterogeneity, but failed to perform meta-regression according to Cochrane's criteria due to the small number of included studies. Third, a significant metric, 5 years or 10 years of follow-up, was reported in only 1 study and hence could not be meta-analyzed.29 Fourth, pooled analyses of QoL data were limited as only 2 studies reported these results with the low quality certainty. We planned to report data on longer-term quality of life, but none of the included trials have reported these data as yet. Fifth, we did not report on adverse events because our main objectives were OS, PFS and QoL. Finally, our analysis was performed with trial-level data rather than individual patient data, which would limit the power of the analysis. Better understanding of the estimated effects could be gained from a patient-level meta-analysis in more selected populations. Despite these limitations, we believe that our study allows for the most comprehensive and up-to-date analysis by providing simultaneous assessment of QoL in dnMBC patients undergoing breast surgery.