Main

Endometrial cancer accounts for 5% of all cancers in women (Ferlay et al, 2010). In 2008, 287 000 endometrial cancer cases were reported worldwide, making it the sixth most common cancer in women. Endometrial cancer is more common in developed vs undeveloped regions and is the fourth most common cancer in women in Europe and the United States (Ferlay et al, 2010; Jemal et al, 2010).

Early-stage endometrial cancer (International Federation of Gynaecology and Obstetrics stage I or II) can be effectively treated with surgery, with or without adjuvant radiotherapy or chemotherapy, and is associated with a 5-year survival rate of 80% to 90% (Creasman et al, 2006). Treatment of recurrent and/or metastatic endometrial cancer is limited to cytotoxic chemotherapy and, for patients with hormone receptor-positive disease, hormonal therapy (Colombo et al, 2011; National Comprehensive Cancer Network Inc., 2012). First-line chemotherapy for advanced disease typically includes a platinum salt, paclitaxel, and/or anthracyclines and is associated with a median progression-free survival (PFS) of <12 months and a median overall survival (OS) of <20 months (Humber et al, 2007; Sovak et al, 2007; Pectasides et al, 2008). Although there is no standard of care in second- and third-line settings, patients may receive anthracyclines if not used in the first-line setting, cyclophosphamide, 5-fluorouracil, topotecan, or progestational agents. However, many patients are not eligible for chemotherapy owing to its associated toxicity profile.

Therapies targeted to signal transduction pathways dysregulated in endometrial cancer may provide improved efficacy and safety for recurrent or metastatic endometrial cancer. The phosphatidylinositol 3-kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) pathway is frequently overactivated in endometrial cancer, making it an attractive treatment target. Activating mutation or amplification of PIK3CA is observed in 27% to 52% of endometrial cancers, with mutations more common in endometrioid vs non-endometrioid tumours (Oda et al, 2005; Hayes et al, 2006; Catasus et al, 2009, 2010; Konopka et al, 2011; Rudd et al, 2011). Mutations in PTEN, a negative regulator of the PI3K/Akt/mTOR pathway, are observed in 11% to 79% of endometrial cancers depending on histological type, with mutations more common in endometrioid vs non-endometrioid tumours (Hayes et al, 2006; Catasus et al, 2009, 2010; Konopka et al, 2011; Rudd et al, 2011).

Everolimus is an oral mTOR inhibitor currently approved in various countries for treatment-refractory advanced renal cell carcinoma; progressive, unresectable, advanced pancreatic neuroendocrine tumours; and renal angiomyolipoma and subependymal giant-cell astrocytomas associated with tuberous sclerosis complex. Preclinical data suggest mTOR inhibition may provide benefit in advanced endometrial cancer (Podsypanina et al, 2001; Milam et al, 2007; Block et al, 2010). For example, in a mouse Pten heterozygote model, everolimus significantly reduced endometrial hyperplasia and the proliferation index, and significantly increased apoptosis compared with control (Milam et al, 2007).

The objective of the current study (ClinicalTrials.gov identifier NCT00870337) was to assess the efficacy and safety of everolimus monotherapy in women with advanced or metastatic endometrial cancer refractory to one or two previous chemotherapy regimens.

Patients and Methods

Patients

Patients eligible for study inclusion were aged 18 years and had histologically confirmed, metastatic or locally advanced endometrial adenocarcinoma not eligible for surgery that progressed after one or two lines of chemotherapy; at least one line of chemotherapy must have contained a platinum derivative. Additional inclusion criteria included a 3-month interval since the last chemotherapy line; 1 metastatic lesion located outside previously irradiated zones and measurable according to the Response Evaluation Criteria In Solid Tumours, version 1.0 (RECIST 1.0) (Therasse et al, 2000); Eastern Cooperative Oncology Group performance status 2; negative pregnancy test within 7 days of first dose of study treatment and use of adequate contraception throughout the study (women of child-bearing potential only); and adequate bone marrow and hepatic and renal function. Exclusion criteria included uncontrolled cerebral metastases; other serious or uncontrolled medical conditions; history of other cancer, with the exception of adequately treated cervical carcinoma in situ or basal or squamous cell carcinoma; and previous mTOR inhibitor treatment. After enrolment of the first 22 patients, a protocol amendment was introduced that excluded patients who received treatment with other chemotherapeutic drugs, hormonal therapy, radiation, or other antitumor agent within 21 days of the first dose of study drug.

All patients provided written informed consent. The study was approved by a central ethics committee and conducted in accordance with international standards of good clinical practice and all local laws and regulations.

Procedures

In this French, multicenter, open-label, phase II trial, all patients received everolimus 10 mg once daily as two 5-mg pills taken with water on an empty stomach or after a light, low-fat meal. Treatment was administered continuously until disease progression or unacceptable toxicity. In the case of adverse events (AEs) or toxicity thought to be related to everolimus, everolimus dosing could be delayed or reduced according to an algorithm outlined in the study protocol. The first and second dose reductions were 5 mg per day and 5 mg every other day, respectively. If the AE or toxicity did not resolve within 21 days of treatment interruption or recurred after everolimus reintroduction, treatment was discontinued. Grade 3 hyperlipidemia was managed per local clinical practice; everolimus was withheld for grade 4 hyperlipidemia. Hyperglycaemia was managed by everolimus dose adjustment and/or the addition of metformin as investigator choice.

Tumour measurements included X-ray and/or computed tomography (CT) scan of the chest and abdominal and pelvic CT scan or magnetic resonance imaging. Measurements were performed at screening, every 3 months during treatment and follow-up, and at the time of treatment discontinuation. Safety assessment included monitoring and recording of all AEs, regular laboratory evaluations of haematology and clinical chemistry, regular measurement of vital signs, performance of physical examinations, and recording of all concomitant medications. Adverse events and laboratory abnormalities were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 3.0 (National Cancer Institute, 2006).

Statistical analysis

The primary study end point was the rate of non-progressive disease at 3 months, defined as the proportion of patients with a complete or partial response (CR or PR, respectively) or stable disease (SD) as assessed by local review according to RECIST 1.0. The choice of non-progressive disease at 3 months as the primary end point is consistent with other phase II studies of rare cancers, including sarcoma (Schoffski et al, 2011). Secondary end points included the rate of non-progressive disease at 6 months and best overall response according to RECIST 1.0; duration of response, defined as the time from the date of the first confirmed response to the date of disease progression or death due to cancer; PFS, defined as the time from enrolment to the date of disease progression or death due to any cause; OS, defined as the time from enrolment to the date of death due to any cause; and safety and toxicity. The potential predictive value of select biomarkers is reported separately.

This study was designed using a Simon two-stage mini–max design (Simon, 1989). To show a >15% success rate (i.e., >15% of patients without progressive disease at 3 months) with 90% power and α=5%, 44 evaluable patients were required. In stage one of the study, if at least four of the first 19 evaluable patients demonstrated CR, PR, or SD, recruitment was continued until 44 evaluable patients were enrolled. In stage two, if at least 11 patients demonstrated CR, PR, or SD, everolimus was considered to have shown sufficient efficacy to warrant further study.

Kaplan–Meier methodology was used to estimate PFS and OS. Progression-free survival and OS equality in subgroups was assessed using the log-rank test. The protocol-specified population evaluable for non-progressive disease included all enrolled patients who had no protocol deviations and received everolimus for 1 month according to the study protocol. The protocol-specified safety population and the population evaluable for clinical benefit included all patients who received 1 dose of study drug.

Results

Forty-four patients were enrolled at 18 French centres between April 2008 and October 2009. Between stages one and two, enrolment was stopped for 6 months. At the time of analysis, 43 patients had discontinued treatment owing to progressive disease (65%) or AEs (35%). Median age was 65 years (range, 52–77 years), and 64% of patients had endometrioid tumours (Table 1). Per protocol, all patients previously received one (66%) or two (34%) lines of chemotherapy. The treatment-free interval was 6 months in 64% of patients. In addition to chemotherapy, a majority of patients were previously treated with surgery (89%) and radiotherapy (80%); 14% of patients received previous hormonal treatment.

Table 1 Baseline demographics and disease characteristics of the total population (N=44)

Efficacy

At 3 months, 16 patients (36%) in the total population had non-progressive disease, including two (5%) with PR and 14 (32%) with SD (Table 2). The trial, therefore, met the prespecified criteria for efficacy of everolimus in patients with chemotherapy-refractory advanced endometrial cancer. At 6 months, the rate of non-progressive disease remained 36%, with an additional two patients experiencing PR (Table 2). At both 3 and 6 months, the rates of non-progressive disease in patients with endometrioid (n=28), serous (n=11), and other (n=5) histology were 39%, 27%, and 40%, respectively (P=not significant) (Table 2). The best overall response was PR in four patients (9%), SD in 12 patients (27%), and progressive disease in 25 patients (57%); three patients (7%) were not evaluable because they received <3 months of treatment or had their treatment discontinued owing to toxicity. Partial response was observed in three patients (11%) with endometrioid histology, one patient (9%) with serous histology, and no patients with other histology. The median (range) duration of response was 3.1 months (2.5−19.8+ months). The median (range) duration of SD was slightly longer at 4.3 months (2.1–14.9 months).

Table 2 Disease response rates at 3 and 6 months in the total population (N=44)

Median PFS in the overall population was 2·8 months (95% confidence interval (CI), 0.6–5.1) (Figure 1A). No difference in median PFS was noted between patients previously treated with one vs two chemotherapy regimens (3.0 months (95% CI, 0.0–6.1) vs 2.8 months (1.3–4.3); P=0.784). Median OS was 8.1 months (95% CI, 5.1–11.1) in the overall population (Figure 1B). No difference in median OS was observed between patients previously treated with one vs two chemotherapy regimens (9.3 months (95% CI, 5.0–13.6) vs 7.7 months (95% CI, 6.6–8.9); P=0.735). No effect of tumour type, histological grade, presence of abdominal, pelvic, lung, lymph node, liver, or bone/soft tissue metastases, or duration of treatment-free interval on PFS or OS was observed (data not shown).

Figure 1
figure 1

Efficacy results in the total population ( N =44). (A) progression-free survival; (B) OS

Safety

Median duration of everolimus exposure was 2.5 months (range, 12 days to >25.7 months). Dose reductions and interruptions occurred at 18 (13%) and 31 (23%) of 137 total study visits, respectively. The most common cause of dose reduction or interruption was mucositis, which accounted for 39% of reductions and 23% of interruptions. Other AEs that led to dose reduction and interruption were asthenia, fever, vomiting, hyperlipidemia, rash, thrombocytopenia, nausea, thromboembolism, and diarrhoea.

Although all patients enrolled in the study met the protocol-specified criteria for inclusion in the safety population, one female patient was excluded from analysis because she died of unknown causes prior to the 1-month visit, thus precluding the collection of safety data. Of the 43 patients included in the safety population, all experienced at least one AE, a majority of which were of grade 1 or 2 severity (Table 3). The most common non-haematologic AEs were fatigue (93%), nausea (51%), rash (49%), vomiting (49%), and mucositis (49%). The most common grade 3/4 AEs were fatigue (42%), anorexia (26%), and infection (16%). Pneumonitis was observed in five patients (12%; grade 3, 5%). Three patients who experienced SD developed pulmonary toxicities necessitating permanent everolimus discontinuation after 1.6 months, 2.5 months, and 7.0 months of treatment. In one patient, pulmonary toxicity resolved without supportive therapy. In the other two patients, pulmonary toxicity resolved after treatment with ceftriaxone (n=1) or ceftriaxone plus metronidazole followed by ceftriaxone plus piperacillin (n=1). Thromboembolism occurred in seven patients (16%). Thromboembolic events included one pulmonary embolism (grade 4) and six venous thromboembolisms (one grade 4, three grade 3, and two grade 2). The most common haematological AEs were anaemia (100%), lymphopenia (81%), and leucopenia (49%) (Table 3). Thrombocytopenia (all grades) was observed in 21% of patients, with one patient each experiencing grade 3 and 4 events. Biochemical abnormalities included hypercholesterolaemia (81%), hypertriglyceridemia (69%), and hyperglycaemia (61%) (Table 3). Additional therapy included prophylactic (39%) and curative (32%) mouthwash, antibiotics (36%), and corticosteroids (20%). Hypertension at baseline reduced the risk of haemorrhage compared with normal blood pressure at baseline (P=0.045). No other relationships between comorbid conditions and risks of toxicity were observed.

Table 3 Adverse events reported in 10% of patients in the safety population (N=43), regardless of relationship to study drug

A total of 56 serious AEs were experienced by 34 patients (77%). Of these serious AEs, 25 (45%) were considered to be related to study drug. The most common serious AEs were reduced general condition (nine events), thrombosis (six events), infection and interstitial lung disease (four events each), and hyperglycaemia and renal insufficiency (three events each). Eleven patients (25%) died during the study. Of these deaths, 10 were related to disease progression. One patient experienced sudden death considered unlikely to be related to study drug.

Discussion

In this open-label, phase II study, everolimus enabled 36% of patients with advanced endometrial cancer refractory to one or two previous chemotherapy regimens to remain progression free at 3 months. Thus, the trial met the prespecified criteria for efficacy of everolimus in this patient population. The non-progressive disease rate remained 36% at 6 months. The rate of non-progressive was higher in patients with endometrioid vs serous histology (39% vs 27%), although this difference was statistically significant. Median PFS and OS in the total population were 2.8 months and 8.1 months, respectively, with similar results observed for patients previously treated with one and two lines of systemic chemotherapy.

Our results are generally similar to those of other phase II studies of mTOR inhibitors in pretreated recurrent/metastatic endometrial cancer (Table 4) (Colombo et al, 2007; Slomovitz et al, 2010; Fleming et al, 2011; Mackay et al, 2011; Oza et al, 2011a, 2011b; ARIAD Pharmaceuticals Inc., 2012). In another trial of everolimus for patients with progressive or recurrent endometrioid endometrial cancer previously treated with one or two chemotherapy regimens, the non-progressive disease rate at 8 weeks was 43%, and the median duration of non-progressive disease was 4.5 months (Slomovitz et al, 2010). At 20 weeks, the confirmed clinical benefit rate, defined as CR or PR or SD of 8 weeks in duration, was 21%. Interim results of trials of oral ridaforolimus for advanced endometrial cancer suggest ridaforolimus provides a response rate similar to that of everolimus but with a higher rate and a slightly longer duration of disease stabilisation (Table 4) (Colombo et al, 2007; Mackay et al, 2011; Oza et al, 2011b; ARIAD Pharmaceuticals Inc., 2012).

Table 4 Studies of mTOR inhibitors in endometrial cancer

Although achieving disease remission while receiving everolimus monotherapy appears unlikely, limiting disease growth by interrupting this biological pathway is a reasonable approach for disease control. Evaluation of everolimus in combination with hormonal therapy or chemotherapy should be explored. For example, everolimus in combination with the non-steroidal aromatase inhibitor letrozole has shown promising clinical efficacy in patients with recurrent endometrial cancer (Slomovitz et al 2011). In addition, everolimus in combination with tamoxifen has been shown to restore hormonal sensitivity in advanced breast cancer and provide clinical benefit (Bachelot et al, 2012). Given the high rate of comorbid diabetes in patients with endometrial cancer, combination therapy with everolimus and metformin, which has demonstrated synergistic activity in preclinical models of cancer (Liu et al, 2012), is being explored in endometrial cancer (ClinicalTrials.gov identifier NCT01205672). Conversely, more profound inhibition of the PI3K/Akt/mTOR pathway may provide improved clinical efficacy in chemotherapy-refractory endometrial cancer. The novel oral agent BEZ235, which inhibits both PI3K and mTOR, has demonstrated promising preclinical activity in both in vitro and in vivo endometrial cancer models (Yang et al, 2011; Shoji et al, 2012). The pan–class I PI3K inhibitor BKM120, which has demonstrated promising activity in preclinical models of several cancers, is being assessed in phase II studies as second-line therapy for advanced endometrial cancer (ClinicalTrials.gov identifier NCT01289041) and as first-line therapy for advanced, metastatic, or recurrent endometrial cancer (ClinicalTrials.gov identifiers NCT01397877 and NCT01550380). Other PI3K/Akt/mTOR pathway inhibitors currently in development for endometrial cancer include the pan–class I PI3K inhibitor XL147 (SAR245408; ClinicalTrials.gov identifier NCT01013324), the dual PI3K/mTOR inhibitors PF-04691502 and PF-05212384 (ClinicalTrials.gov identifier NCT01420081) and GDC-0980 (ClinicalTrials.gov identifier NCT01455493), and the Akt inhibitor MK-2206 (ClinicalTrials.gov identifier NCT01307631). Of note, stratifying patient enrolment by KRAS mutation status may be of value in future studies exploring therapies targeted to the PI3K/Akt/mTOR pathway, as in a biomarker analysis based on the present study population, KRAS mutation was significantly associated with shorter PFS (P<0.001) and OS (P=0.034) (Trédan et al, 2012).

The safety profile of everolimus was acceptable in the context of this heavily pretreated population of patients with advanced endometrial cancer. Although all patients experienced at least one AE, most were of grade 1 or 2 severity. The most common grade 3/4 AEs of any cause were fatigue (42%), anorexia (26%), and infection (16%). The overall safety profile observed in this study was similar to those observed in other studies of everolimus in cancer (Motzer et al, 2010; Yao et al, 2011; Zhu et al, 2011), including endometrial cancer (Slomovitz et al, 2010). Three patients with SD experienced pulmonary toxicity requiring study withdrawal after 1.6, 2.5, and 7.0 months of treatment. After everolimus discontinuation, all cases of pulmonary toxicity resolved either spontaneously (n=1) or with treatment (n=2). Pulmonary-related toxicity needs to be further evaluated in patients receiving prolonged mTOR inhibitor therapy.

Specific to this highly comorbid population, we did not find any correlation between comorbidities (e.g., diabetes, obesity, hypertension) and specific toxicities, except for hypertension, which reduced the risk of haemorrhage compared with normal blood pressure at baseline (P=0.045). Of note, the percentage of patients who discontinued treatment due to AEs (35%) was higher in this study than in studies of everolimus monotherapy in patients with other cancers (8%–17%) (Motzer et al, 2010; Yao et al, 2011). However, it was consistent with the high rate of discontinuation observed in other phase II studies of mTOR inhibitors in heavily pretreated advanced endometrial cancer (Table 4) (Slomovitz et al, 2010; Mackay et al, 2011; Oza et al, 2011b).

In conclusion, everolimus demonstrated clinical efficacy according to the prespecified criteria, as well as acceptable tolerability, in patients with advanced or metastatic endometrial cancer that progressed after one or two lines of previous systemic chemotherapy, supporting the further development of therapies targeted to the PI3K/Akt/mTOR pathway in endometrial cancer.