Background
Phyllodes tumors (PTs) of the breast are typically large, rapidly growing tumors that account for up to 1% of all breast neoplasms [
1]. The World Health Organization (WHO) classifies phyllodes tumors into three histologic subtypes: benign, borderline, and malignant, based on stromal cellularity, stromal cell mitotic activity, stromal nuclear atypia, stromal overgrowth and type of borders (infiltrating or pushing) [
2‐
4].
Although many phyllodes tumors tend to behave in a benign manner, the clinical outcomes of phyllodes tumors are hard to predict because of relatively high recurrence rates and occasional distant metastases. The current approach to preventing local relapse and metastasis is surgical resection with wide margins. However, even with wide surgical resection, the local recurrence rate remains as high as 8 to 36% [
5]. Furthermore, recurrent phyllodes tumors can progress toward more malignant phenotypes [
6], in which metastases have been estimated to occur in up to 25% of patients [
7].
Despite these data, no well-established adjuvant therapy applies to high grade phyllodes tumors, which is partly due to the controversial roles of adjuvant radiotherapy and chemotherapyl [
8]. The use of radiotherapy lacks sufficient prospective study data regarding borderline and malignant PT [
9], while the adoption of chemotherapy is yet to solve the rarity of disease presentation [
10]. Furthermore, no randomized clinical trials of chemotherapy and/or radiation therapy have been published.
To help address this problem, we performed a literature review and meta-analysis to evaluate the roles of radiotherapy and chemotherapy in PT treatment.
Methods
Search strategy
We performed a comprehensive search of databases, including PubMed, Web of Science, Embase and the Cochrane Library, from 1985 to Feb 1, 2019. The following MeSH terms and their combinations were searched: (breast tumor/ sarcoma/ neoplasm) and (phyllodes or phyllode) and (radiotherapy/ chemotherapy).
Inclusion and exclusion criteria
Articles concerning radiotherapy and chemotherapy as adjuvant therapy in women with breast phyllodes tumors were included, regardless of prospective or retrospective ones. However, case reports were excluded. So were studies without the outcomes of interest, such as local recurrence, survival and distant metastasis.
Data were extracted independently by two of the authors (N.Y. and C.X). Consensus was reached between the two authors if there was a discrepancy among the collected data. For each included study, the following information was collected: author, year of publication, country, months of follow-up, number of total cases and patients treated with adjuvant therapy, surgery type and percentage, type and details of adjuvant therapy, age, tumor size, histological type, margins, number of local recurrences, number of metastases, number of patients surviving disease-free and number of overall survivors. A quality assessment of the included studies was also performed based on the tool for case series studies provided by U.S. Department of Health & Human Services (
https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools). Per study, there were many aspects of study design that required termwise evaluation, including objective of the study, cases definition, subjects, interventions, outcome measures, follow-up, statistical methods and results. Quality reviewers could select “yes,” “no,” or “cannot determine/not reported/not applicable” in response to each item on the tool. For each item where “no” was selected, reviewers were instructed to consider the potential risk of bias that could be introduced by that flaw in the study design or implementation. Cannot determine and not reported were also noted as representing potential flaws. The procedure was done independently by two reviewerss (X.C and K.C.). A review grade (good, fair or poor) was allocated to each study. A “good” study has the least risk of bias, and results are considered to be valid. A “fair” study is susceptible to some bias deemed not sufficient to invalidate its results. A “poor” rating indicates significant risk of bias.
As the data included in our study were all either < 0.2 or > 0.8, these data do not follow a normal distribution, and therefore continuous outcome data could not be carried out. The Free-Tukey double arcsine transformation transformed the data to follow a normal distribution, which then could be used for meta-analysis. Once the meta-analysis is done, the output data were back-transformed and therefore through this process we can compare the data obtained via meta-analysis with original data [
11]. Separate analysis was performed based on the transformed proportions using random models. All data were back-transformed to determine the rate and 95% confidence interval (CI). Statistical heterogeneity among studies was assessed using Q and I
2 statistics. For the Q statistics, data were heterogeneous if
p < 0.1. The Q test was used to test for effect size heterogeneity. I
2 values of < 25, 25–75% and > 75% corresponded to low, moderate, and high degrees of heterogeneity, respectively. Sensitivity analysis and meta-regression analyses were conducted to determine the origin of heterogeneity. Meta-regression analysis was achieved by using liner regression models. We also conducted subgroup analyses based on study size, surgery type, age, tumor size, histological type and margins. All statistical analyses were performed by using Stata 13.0 (Stata, College Station, Texas, USA).
Discussion
PTs are divided into three types: benign, borderline and malignant, according to their histological characteristics [
3]. Surgical treatment, including breast-conserving surgery and mastectomy, are the mainstays of curative treatment of PTs. However, these patients still encounter a fairly high overall recurrence rate of 19.1% [
28] and a 25% malignant recurrence rate, according to a review involving 5530 patients [
28]. Furthermore, malignant and borderline PT had metastasis rates ranging from 22 to 75% and mortality rates ranging from 23 to 32% [
29‐
32]. These data emphasize the importance of local control and prevention of metastasis.
Currently, the use of adjuvant therapy for PTs remains controversial because of inadequate data from large prospective studies. The absence of these data may be due to the low incidence of PTs and the limited utilization of adjuvant therapy.
Recently, adjuvant radiotherapy has been more frequently utilized. According to a study in the National Cancer Database from the American College of Surgeons’ Commission on Cancer involving 3120 patients, adjuvant radiotherapy for PTs in 2008–2009 was used in 19.5% of cases, more than doubled compared to the rate of 9.5% in 1998–1999. [
18] Data from this large retrospective study suggested that radiotherapy could extend the time to local recurrence and decrease the local recurrence rate, with no significant influence on survival. Another prospective study also discussed the effectiveness of radiotherapy for local disease control [
12]. However, scant data has discussed the relationship between radiotherapy and metastasis. The pooled LR rate for patients who underwent both radiotherapy and surgery was (8%, 95% CI: 1–22%). These data were lower than both our calculated LR rate of 19% (95% CI: 7–18%) and the pooled LR rates of borderline (13%) and malignant (18%) reported by a meta-analysis [
33]. Furthermore, the results of our subgroup analysis suggested that irradiation may be more effective in younger patients (< 45 years), patients with larger tumors, patients with malignant tumors and patients with wider excision. The meta-regression analysis also confirmed the importance of margin status in local control. These data emphasize the importance of ensuring adequate surgical margins. However, surgery type showed less impact on disease control based on the subgroup analysis. We suggest that for those PTs with high malignancy, radiotherapy should be used as adjuvant therapy without consideration of the surgery type. Our calculated metastasis rate of 4% (95% CI: 0–13%) in patients treated by radiotherapy, compared with a metastasis rate of 8% (95% CI: 3–15%) in patients receiving only surgical treatment, also suggested that radiotherapy may be effective in the prevention of metastasis. Although our calculated survival data (disease-free survival rate: 93%, overall survival rate: 96%) for radiotherapy are higher than those for the surgery group (disease-free survival rate rate: 70%, overall survival rate: 76%), they are similar to the data from the Surveillance, Epidemiology, and End Results database (SEER). The estimated 5-, 10-, and 15-year rates of cancer-specific survival for all women were 91, 89, and 89%, respectively [
34]. This suggests that radiotherapy may have little effect on prolonging survival. Further subgroup analysis also showed improved survival among patients with cleaner margins.
Our review of the available data shows a negligible role for chemotherapy in the treatment of PTs. Most clinicians avoid chemotherapy as a first-line treatment due to lack of evidence. To date, there has been only one prospective study involving 28 patients, which has showed that chemotherapy has little effect on survival [
10]. The sample sizes of other retrospective studies were also too small to prove the efficacy of chemotherapy for PTs treatment. Moreover, PTs with higher histological grades have higher metastatic potential. Few studies discuss the treatment for patients with metastatic PTs. However, the results of these few studies seem promising as nearly half of the patients exhibited partial responses to treatment [
20,
35].
Our study was the first meta-analysis designed to evaluate the efficacy of radiotherapy in PTs treatment. However, there were some limitations to our study. First, we calculated only recurrence and survival rate data without considering elapsed time. Second, the studies in our meta-analysis were all observational studies, the quality of which may be sub-optimal. In conclusion, the limited incidence of PT remains a challenge for the study of PT.
Conclusions
In summary, this study indicates that radiotherapy is effective in PTs disease control without prolonging survival. However, the data examined were mostly retrospective and permitted comparative analysis between published treatments, which is a common limitation throughout the literature. Therefore, further studies, particularly prospective studies, are needed to prove the efficacy of adjuvant therapy.
Acknowledgements
We thank all the women who participated in this study for their willingness to contribute their data.