Background
Hormone receptors (HRs) including estrogen receptor (ER) and/or progesterone receptor are expressed in approximately 70% of breast cancers. Endocrine therapy is a very important treatment for patients with HR-positive advanced breast cancer. However, resistance to endocrine therapy is an immense challenge in clinical settings [
1,
2]. One mechanism of drug resistance to endocrine therapy is the activation of the phosphatidylinositol 3-kinase (
PI3K)/
AKT/mammalian target of rapamycin (
mTOR) pathway [
3]. The inhibitors of the
PI3K/AKT/mTOR pathway could reverse endocrine drug resistance [
3,
4]. Everolimus is a selective inhibitor of
mTOR that has shown potential for overcoming the resistance to endocrine therapy [
5‐
7]. However, due to the heterogeneity of breast cancer, a subset of patients do not respond to everolimus. Therefore, it is crucial to find biomarkers that predict the efficacy of everolimus in clinical settings [
8]. Several experimental studies have indicated that cancers with
PIK3CA/
PTEN mutations are sensitive to everolimus; however, clinical trials did not draw the same conclusions [
9‐
16].
To better select patients who will benefit most from or be resistant to everolimus, we conducted a retrospective analysis on data from 120 patients with metastatic breast cancer who underwent therapy at the National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College from February 2014 to March 2017. We also performed circulating tumor DNA (ctDNA) analysis on sixteen patients to determine the association between gene mutations and response to everolimus.
Discussion
The resistance of endocrine therapy is a major clinical challenge. The mechanism of endocrine therapy resistance is very complex, and one of the most important mechanisms is the activation of the
mTOR signaling pathway [
4]. Everolimus, an
mTOR inhibitor, can enhance the efficacy of endocrine therapy and may reverse drug resistance [
21]. We performed the present study to find patients who will benefit most from or exhibit resistance to everolimus.
In the present study, everolimus combined with endocrine therapy in patients with HR-positive metastatic breast cancer resulted in a PFS of 5.1 months. The PFS identified in the present study is shorter than the PFS in the BOLERO-2 study [
5]. This disparity can be accounted for by the retrospective design of this study and the higher tumor and treatment burdens of the patients. In the present study, 45.0% patients had > 2 metastatic sites, 81.7% patients received > 2 lines of previous endocrine therapies, and 43.0% received > 2 lines of previous chemotherapy. However, in the BOLERO-2 study, only 35% of patients had > 3 metastatic sites, and 26% received chemotherapy in metastatic settings before everolimus [
5,
6]. In our study, < 3 metastatic sites, < 2 lines of previous endocrine therapy, and < 2 lines of previous chemotherapy were associated with improved survival.
Additionally, the AEs profile in this study was similar to that of previous studies [
5‐
7,
22]. Consistent with previous reports, stomatitis was the most frequent treatment-related adverse event. However, the frequency of stomatitis in our study was lower than that reported in previous studies. This discrepancy may be due to each patient receiving an oral care package that prevented stomatitis in our study. Effective intervention and prevention may help reduce the occurrence of severe stomatitis. Kangfuxinye has been shown to have a potential preventive effect against stomatitis and was reported to be effective in the prevention of mucositis induced by chemoradiotherapy in a phase III clinical study of nasopharyngeal carcinoma [
23].
Subgroup analyses in the TAMRAD study indicated that the median time to tumor progression (TTP) was 14.8 months vs 5.4 months in patients with secondary resistance and patients with primary resistance, respectively [
7]. Similarly, patients with secondary resistance to AIs had a significantly higher CBR with everolimus combined with tamoxifen (74%) than with tamoxifen alone (48%) [
7]. However, in the present study, no difference in survival was observed between patients who received endocrine drugs that were used in previous treatments and patients who did not use any endocrine drugs (median PFS 5.23 and 5.10 months, respectively,
p>0.05). We also did not find any difference in outcomes between patients who received endocrine drugs to which they acquired primary resistance in previous treatments and patients who received endocrine drugs to which they acquired nonprimary resistance (
p > 0.05). A real-world retrospective study performed in China also found results similar to ours [
22].
Thus far, no study has directly compared the efficacy among different endocrine drugs combined with everolimus. In our study, multivariate analysis indicated that treatment with everolimus combined with fulvestrant was associated with improved survival. However, these results require further confirmation in large-scale randomized clinical trials.
Mutations in the
PIK3CA/AKT/mTOR pathway are frequent in breast cancer patients [
24,
25]. Experimental studies have demonstrated that breast cancers with
PIK3CA mutations are more sensitive to everolimus, but this has not been confirmed in clinical studies [
12‐
14]. This result may be due to the heterogeneity of the mutational status between primary tumors and metastases and the small number of patients included in previous studies [
8,
26]. In the present study, the frequency of
PIK3CA mutations was 62.5%, which was higher than that reported in previous studies (20–45%) [
19,
24]. This discrepancy may have occurred because patients in the present study had advanced cancer with increased treatment burden or because
PIK3CA mutations are significantly more frequent in HR-positive cancers than in HR-negative cancers [
24,
27,
28]. Based on the present study, patients with the
PIK3CA/H1047R mutation had longer PFS than patients with wild-type or other mutant forms of
PIK3CA, with median PFS rates of 8.8 and 4.1 months, respectively (
p = 0.020). However, we did not find similar differences in survival between patients with other types of
PIK3CA mutations and patients with wild-type
PIK3CA in whom the median PFS was 4.6 and 7.0 months, respectively (
p > 0.05). In the BOLERO-2 study, mutational analysis of plasma cell-free DNA (cfDNA) indicated no relationship between PFS after everolimus treatment and
PIK3CA genotypes, which was consistent with a previous analysis of tumor tissue DNA [
14]. Preclinical data suggested that
PIK3CA and
mTOR inhibitors had a lower IC50 for H1047R than for E542K or E545K [
29]. The
PIK3CA/H1047R mutation was also reported to confer sensitivity to everolimus in early-phase clinical trials in many types of cancers [
30]. Therefore,
PIK3CA mutations including H1047R, E545K, and E542K cannot predict patient responses to everolimus. Hence, it is possible that not all of the
PIK3CA mutations confer sensitivity to everolimus and that the H1047R mutation may be a potential biomarker of sensitivity to everolimus. One possible mechanism is that the
PIK3CA/H1047R mutation is a stronger driver of tumor development than other types of
PIK3CA mutations such as E542K and E545K. Therefore, patients with
PIK3CA/H1047R mutations were more sensitive to the
mTOR inhibitor everolimus [
31]. However, these results require further confirmation in clinical trials.
PTEN gene loss was another biomarker that was reported to be a marker of sensitivity to
PI3K/AKT/mTOR inhibitors in preclinical studies [
15,
16]. However, clinical trials including TAMRAD and BOLERO-2 did not find any association between
PTEN gene status and everolimus efficacy [
13,
32]. In the present study, a
PTEN gene mutation was detected in only one patient, and this patient received tamoxifen plus everolimus therapy and had SD for 5.0 months. However, based on just one patient, we were unable to find an association between
PTEN gene mutations in ctDNA and response to everolimus.
Despite the advantages described above, several limitations of this study should be noted. First, the retrospective design and the small sample of patients who underwent ctDNA analysis do not provide sufficient power to derive statistically sound conclusions. Additionally, because of the small sample size, we did not analyze the association between everolimus and other genetic alterations such as PTEN gene loss and KRAS gene mutations, which have been reported to be associated with everolimus response.