Advertisement

Targeted Oncology

, Volume 13, Issue 5, pp 641–648 | Cite as

Everolimus Exposure and Early Metabolic Response as Predictors of Treatment Outcomes in Breast Cancer Patients Treated with Everolimus and Exemestane

  • Annelieke E. C. A. B. Willemsen
  • Lioe-Fee de Geus-Oei
  • Maaike de Boer
  • Jolien Tol
  • Yvonne Kamm
  • Paul C. de Jong
  • Marianne A. Jonker
  • Allert H. Vos
  • Willem Grootjans
  • Johannes W. B. de Groot
  • Sasja F. Mulder
  • Erik H. J. G. Aarntzen
  • Winald R. Gerritsen
  • Carla M. L. van Herpen
  • Nielka P. van Erp
Open Access
Original Research Article

Abstract

Background

Treating breast cancer patients with everolimus and exemestane can be challenging due to toxicity and suboptimal treatment responses.

Objective

We investigated whether everolimus exposure and early metabolic response are predictors for toxicity and effectiveness in these patients.

Patients and Methods

We performed pharmacokinetic assessments 14 and 35 days after starting treatment. [18F]fluorodeoxyglucose-positron emission tomography (18F-FDG-PET) was performed at baseline, and 14 and 35 days after the start of the therapy. We recorded toxicity, defined as dose interventions within 3 months, and progression-free survival (PFS).

Results

Among 44 evaluable patients, the geometric mean (GM) Ctrough was higher in patients with toxicity compared to patients without (17.4 versus 12.3 μg/L (p = 0.02)). The optimal cut-off value to predict toxicity was Ctrough > 19.2 μg/L. GM Ctrough of patients with and without progressive disease (PD) within 3 months was not significantly different (12.0 versus 15.2 μg/L (p = 0.118)). In 28 evaluable patients, PD within 3 months could best be predicted using the percentage decrease in peak standardized uptake value normalized by lean body mass of the lesion with highest FDG uptake (SULpeak high) at day 14. Patients with <11% versus >11% decrease in SULpeak high at day 14 had a median PFS of 90 days versus 411 days, respectively (p = 0.0013) and more frequently had PD within 3 months: 70 vs 11%, respectively.

Conclusions

Our results show that everolimus toxicity is related to everolimus Ctrough. No relation was observed between everolimus exposure and treatment effectiveness. An early FDG-PET can identify patients at high risk of nonresponse. These results warrant further validation. Clinicaltrials.gov identifier: NCT01948960.

Key Points

In breast cancer patients treated with everolimus and exemestane, everolimus Ctrough is an important predictor of toxicity, as patients with Ctrough > 19.2 μg/L have a higher risk of clinically relevant toxicity.

A lack of decrease in metabolic activity measured by FDG-PET/CT after 14 days of treatment can identify a subgroup of patients at high risk of nonresponse, for whom an early adjustment of therapy or extra follow-up could be considered.

1 Introduction

The introduction of the combination treatment of everolimus and exemestane markedly improved clinical outcomes in patients with hormone receptor positive metastatic breast cancer (mBC) [1]. Since the BOLERO-2 study, several trials have subsequently confirmed the beneficial effect on treatment outcomes in these patients [2, 3, 4]. Unfortunately, not all patients benefit from this treatment. In the BOLERO-2 study, 10% of patients had progressive disease (PD) as best overall response and the clinical benefit rate (stable disease, partial response, and complete response) for ≥24 weeks was only 51% [5].

Importantly, patients treated with this regimen can experience severe everolimus-induced side effects, as 27% of patients experience grade 3 or 4 toxicity [6]. Moreover, low grade toxicities have a significant impact on the quality of life of these patients as well and can necessitate dose adjustments. Therefore, treating breast cancer patients with everolimus can be challenging. Early prediction of toxicity and nonresponse would greatly facilitate clinical decision-making. Everolimus exposure, patient characteristics, and early metabolic response could be predictors for toxicity and response and may be used as pharmacokinetic and pharmacodynamic biomarkers. Currently, these biomarkers are not used in routine clinical practice.

Everolimus for mBC patients is prescribed at a fixed dose, but shows substantial interindividual variation in drug exposure of 36–45% [7, 8, 9, 10]. Evidence is mounting that the everolimus trough concentration (Ctrough) is an important predictive factor for toxicity and effectiveness [11, 12, 13, 14]. The largest study so far, with 938 patients with solid tumors, showed that a twofold increase in everolimus Ctrough increased the likelihood of tumor size reduction by approximately 40% [12]. In patients with mBC, metastatic renal cell carcinoma (mRCC), or neuro-endocrine tumors, an everolimus Ctrough > 11.9 μg/L was associated with a threefold increase in progression free survival (PFS) and a Ctrough < 26.3 μg/L was associated with a fourfold decreased risk of toxicity [11].

Patient characteristics can be relevant, as subgroups such as obese or elderly patients can show different pharmacokinetics compared to other patients. Obesity can lead to an increased volume of distribution, increased hepatic blood flow, higher CYP3A4 activity, and a change in plasma binding [15, 16]. Everolimus pharmacokinetics in elderly patients can differ as a result of decreased absorption due to gastric atrophy, lower first pass effect due to a decrease in hepatic blood flow, and reduced CYP3A4 activity [17, 18, 19, 20]. The net result of all these influences on drug disposition in obese or elderly patients is yet unknown.

Functional imaging to assess metabolic response could serve as an early biomarker for the evaluation of response to treatment. The inhibition of activity of oncogenic proteins by everolimus and exemestane can result in rapid metabolic changes. These metabolic changes can be detected using [18F]fluorodeoxyglucose-positron emission tomography (FDG-PET). Early detection of nonresponse to treatment can facilitate early decision making, including dose adjustments or a change of therapy. Preclinical data suggest that FDG-PET is a suitable imaging modality for response assessment to everolimus treatment [21, 22, 23, 24, 25]. A clinical study of FDG-PET changes in mRCC patients treated with everolimus showed a modest correlation between metabolic changes and change in tumor size [26]. In mBC patients on endocrine therapy or chemotherapy, metabolic response on FDG-PET was a good predictor of PFS [27, 28]. For mBC patients treated with everolimus and exemestane, to the best of our knowledge, no clinical study assessing the predictive value of FDG-PET for treatment response has yet been published.

The current study aimed to investigate whether everolimus exposure is related to the development of toxicity and treatment effectiveness in patients with mBC. We hypothesized that a therapeutic window for everolimus can be established, in which the balance between toxicity and effectiveness is optimal. Additionally, we explored whether early FDG-PET evaluation can be used as an early read-out for effectiveness. In this way, an ineffective therapy could be stopped earlier, preventing unnecessary toxicity and costs.

2 Materials and Methods

2.1 Patients and Study Design

The present study was an explorative, prospective, open-label, comparative study carried out in six hospitals in the Netherlands. We included women planning to start treatment with exemestane and everolimus. Patients using medication recognized as being strong inhibitors or inducers of CYP3A isoenzymes were excluded and no such medication was allowed during the study.

Patients could participate either in the pharmacokinetic part of the study or in the combined pharmacokinetic and imaging part. For the pharmacokinetic part, patients were enrolled in three subgroups: control (<70 years and body mass index (BMI) <30 kg/m2), obese (<70 years and BMI ≥30 kg/m2), and elderly (≥70 years and BMI <30 kg/m2). Planned enrollment was 25 patients per subgroup, but due to slow recruitment the study closed earlier. For the combined pharmacokinetic and imaging part, 40 patients were planned for inclusion. In the imaging part of the study, patients who were elderly and obese could also be included. The study protocol was approved by the Radboudumc Ethics Committee (reference number 2013–284) and registered as clinicaltrials.gov identifier: NCT01948960. All patients gave written informed consent. The study was performed in accordance with the Declaration of Helsinki.

2.2 Clinical Evaluation

We recorded baseline characteristics, including previous treatments. Clinical evaluation took place before the start of therapy, and after 14, 35, 60, and 90 days of treatment. Response assessment was performed with CT of chest and abdomen at baseline, and subsequently every three months. On the physician’s discretion, additional investigations, such as bone scintigraphy, could be performed. Follow-up took place until progressive disease (PD) (even if everolimus or exemestane were discontinued earlier) and the reason for treatment discontinuation was recorded. PFS was defined as the time from start of treatment until PD or death. PD was assessed using Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 [29]. If RECIST assessment was not possible, PD was assessed by other imaging techniques, CA 15.3 marker, or was clinically determined. If patients had PD within 14 days after stopping everolimus treatment, patients were scored as PD on treatment. Patients were censored for PFS and time to dose change event at time of everolimus discontinuation.

We recorded adverse events using Common Terminology Criteria for Adverse Events 4.0. We considered adverse events leading to a dose change clinically most relevant. Therefore, we defined toxicity as adverse events requiring a dose change within 3 months after start of treatment. All dose changes were recorded until discontinuation of treatment. In our analysis, we included all dose reductions or dose discontinuations that were a result of toxicity at least probably related to everolimus. Temporary drug interruptions without dose changes were not included.

2.3 Pharmacokinetic Analysis and Intervention

We performed everolimus pharmacokinetic assessment after 14 days of treatment. In the imaging part, a second pharmacokinetic assessment was performed at day 35. In the case of dose interruptions, treatment had to be restarted at least 7 days prior to pharmacokinetic assessment to ensure steady state pharmacokinetics. Blood samples were collected before (Ctrough) and 1, 2, 3, 5, and 8 h after administration of everolimus. The trough sample was extrapolated as a 24 h sample. The area under the concentration time curve (AUC0-24h) was calculated using a noncompartmental trapezoidal approach (Phoenix WinNonlin v6.3®). We measured everolimus concentrations in whole blood by a validated liquid chromatography tandem mass spectrometry technique.

Patients in the imaging part who had a low everolimus AUC0-24h (<550 μg*h/L) at day 14 were randomized 2:1 to have a dose escalation or not [7, 8]. Dose escalation was calculated to achieve an AUC >550 μg*h/L. Patients in whom the dose was escalated were excluded for the analysis of the relationship between Ctrough and toxicity and effectiveness.

2.4 FDG-PET Acquisition and Analysis

In the imaging part of the study, patients underwent FDG-PET at baseline, and after 14 and 35 days of treatment. Activity was analyzed using standardized uptake values normalized by lean body weight (SUL), using the Janmahasatian formula [30]. We performed analysis by PERCIST criteria [31]. We measured SULpeak (mean SUL in a volume of interest (VOI) of 1 cm3) and SULmax (voxel with highest activity). For five target lesions we determined the sum of the SULmax (sumSULmax) and the sum of the SULpeak (sumSULpeak), and the SULpeak of the lesion with the highest uptake (SULpeak high). For total lesion glycolysis (TLG), a tumor VOI was determined with a threshold of minimally two times mean activity of selected reference tissue. TLG then was defined as the mean uptake times the volume of this VOI, and the sum for up to five lesions was determined (sumTLG). Additionally, a qualitative judgment was performed independently by two nuclear medicine physicians.

Further details can be found in the Supplementary Material.

2.5 Statistical Analyses

Before statistical tests were performed, Ctrough data were log (base 10) transformed, in order to obtain more symmetrically distributed data. A two sample t-test based on the transformed data was used to test for differences in the mean Ctrough between patients with and without toxicity and response. To test for differences in the mean of the transformed Ctrough between the three subgroups (control, obese, elderly), a one-way ANOVA was used, followed by post-hoc analysis (equality of variances was tested with Bartlett’s test). No Bonferroni adjustment was performed for this post-hoc analysis, as this is an explorative study.

For the analyses described below, the untransformed Ctrough data were used. The association between Ctrough and age and BMI were analyzed, first visually and next with a multivariable linear regression model with Ctrough as dependent variable and age and BMI as independent variables. For the association of Ctrough with toxicity and effectiveness, receiver operating characteristic (ROC) curve analysis was performed and the AUC was determined. Optimal cut-off values were determined using the Youden index, the value associated with the highest sum of sensitivity + specificity – 1. The correlation between Ctrough and metabolic change on FDG-PET was estimated with Spearman’s correlation.

PFS and the time to dose change were estimated by the Kaplan-Meier estimator. A log rank test was performed to test whether the survival curves were significantly different. Two separate multivariable Cox proportional hazards regression models were fitted for PFS and the time to dose change, using backward selection based on a likelihood ratio test with p > .10 for removal. Age and BMI were used as covariates for the analysis of time to a dose change event. The ROC analysis was used to assess which variable of everolimus exposure and which variable of metabolic change had the strongest association with PD within 3 months. These variables were then used as covariates in the multivariable model for the analysis of time to PFS. Model assumptions, including proportional hazards and linearity, were checked and if needed the model was adjusted. Furthermore, the number of previous treatments was added in this model as a covariate, as this generally is a significant predictor of PFS.

Analyses were performed with SPSS v22 and graphs were created using GraphPad Prism v5.03.

3 Results

3.1 Study Subjects

Between October 2013 and October 2017, we enrolled 61 subjects. Four patients dropped out due to screen failure and from two patients we could not obtain pharmacokinetic data. The remaining 55 patients were subdivided into 34 control patients, ten obese patients, nine elderly patients, and two patients that were elderly and obese. Baseline characteristics of these patients are presented in Table 1. Twenty patients had an AUC0-24h < 550 μg*h/L, and 14 patients were randomized to have dose escalation. In three patients dose escalation was not executed, as everolimus treatment had already been interrupted or discontinued. Therefore, 44 patients were suitable for analysis of the relationship between everolimus Ctrough and toxicity; 43 patients were evaluable for the relationship between everolimus Ctrough and effectiveness, as one patient had used exemestane pretreatment, which was not allowed in the study. In four out of these 43 patients, PD was established based on imaging or clinical aspects, as RECIST assessment was not possible. In the imaging part of the study, 39 patients were included. Of these, 29 patients had lesions with sufficient FDG avidity at baseline and were eligible to proceed for metabolic response monitoring using FDG-PET. From one of these 29 patients no FDG-PET was performed at day 14.
Table 1

Baseline patient characteristics

Characteristic

All patients (n = 55)

Control patients (n = 34)

Obese patients (n = 10)

Elderly patients (n = 9)

Age (years), mean (SD)

61.6 (9.7)

58.1 (8.6)

59.5 (6.5)

74.2 (2.9)

BMI (kg/m2), mean (SD)

26.2 (4.4)

24.4 (3.0)

32.8 (3.0)

24.6 (1.9)

Number of previous treatment lines, mean (min-max)

 In adjuvant setting

1.0 (0–4)

0.8 (0–4)

1.9 (0–3)

0.8 (0–3)

 In metastatic setting

2.1 (0–6)

1.9 (0–6)

1.8 (0–3)

3.1 (2–6)

 Total

3.2 (1–6)

2.8 (1–6)

3.7 (3–6)

3.9 (2–6)

Histology (%)

 Ductal carcinoma

71

71

45

89

 Lobular carcinoma

26

23

55

11

3.2 General Pharmacokinetics of Everolimus

In the 55 patients, the geometric mean (GM) Ctrough was 12.6 μg/L (range 4.6–37.5 μg/L) and AUC0-24h was 565 μg*h/L (range 261–1162 μg*h/L). Ctrough was a good predictor of AUC0-24h, with coefficient of determination (r2) = 0.84 (p < 0.001). Nineteen patients had everolimus concentrations measured at 2 and 5 weeks, without a concurrent dose change. In these patients, the intra-individual coefficient of variation for everolimus at day 14 and day 35 was low: 16.8% for Ctrough and 14.8% for AUC0-24h.

3.3 Main Outcomes

3.3.1 Relationship of Ctrough and Toxicity

The reason for everolimus discontinuation was PD in 70% of patients, versus 30% due to toxicity. A dose change due to toxicity within 3 months occurred in 45% of patients. The median time to the first dose change was 52 days (9–159 days). Of 24 patients with a dose change event, this occurred in the first 14 days of treatment in only two patients. No dose changes occurred after day 159 of treatment. The most frequent reason for a dose change or interruption was stomatitis and the most frequent reason for discontinuation was pulmonary toxicity. Patients with a dose change <3 months had a significantly higher GM Ctrough compared to patients without a dose change <3 months: 17.4 μg/L and 12.3 μg/L, respectively (p = 0.02). ROC analysis to predict dose change using Ctrough showed an AUC of 0.71 (p = 0.02). Ctrough was a slightly better predictor of toxicity than AUC0-24h, with an AUC of 0.66. Optimal cut-off value was Ctrough > 19.2 μg/L, with sensitivity 55% and specificity 92%. Thirteen out of 44 patients (30%) had a Ctrough above this threshold. Eleven out of these 13 patients (85%) had toxicity leading to a dose change <3 months, compared to nine out of 31 (29%) patients who did not develop toxicity, resulting in a relative risk of 2.9. Median time to first dose adjustment was significantly shorter for patients with Ctrough > 19.2 μg/L compared to patients with Ctrough < 19.2 μg/L, 42 days versus not reached, p = 0.0008 (Fig. 1), hazard ratio (HR) 5.8 (95% CI 2.1–16.1). When a cut-off of 26.3 μg/L was used, as proposed previously [11], median time to first dose adjustment for patients with Ctrough > 26.3 μg/L was 47 days, compared to not reached for patients with Ctrough < 26.3 μg/L, p = 0.0028.
Fig. 1

Inverse Kaplan-Meier estimates of time to dose change event, based on differences in exposure

3.3.2 Relationship of Ctrough and Effectiveness

GM everolimus Ctrough of patients with and without PD <3 months was 12.0 μg/L and 15.2 μg/L, respectively (p = 0.118). ROC analysis to predict PD <3 months using Ctrough showed an AUC = 0.64, p = 0.15. Ctrough was a slightly better predictor of effectiveness than AUC0-24h with an AUC of 0.60. The optimal cut-off value for Ctrough was <12.6 μg/L, with sensitivity 69% and specificity 60%. Twenty-one out of 43 patients (48%) had an everolimus Ctrough below this threshold. Nine of these 21 patients (43%) had PD <3 months, compared to four out of 23 (17%) patients with everolimus Ctrough > 12.6 μg/L, resulting in a relative risk of 2.5. No relevant difference was seen in the frequency of censoring in the Kaplan-Meier survival analysis, as four out of 21 patients were censored in the group patients with Ctrough < 12.6 μg/L, and six out of 22 in the group patients with Ctrough > 12.6 μg/L. Median PFS time was not significantly different for patients with Ctrough > 12.6 μg/L and < 12.6 μg/L: 318 versus 153 days respectively (p = 0.19), HR 0.85 (95% CI 0.41–1.77) (Fig. 2).
Fig. 2

Kaplan-Meier estimates of progression-free survival, based on differences in exposure

The relationship between Ctrough and PFS was different for the period of the first 365 days of treatment and after this period, as shown in a Cox regression analysis with time-dependent Ctrough, dichotomized for above or below 12.6 μg/L.

The outcomes of the effect of weight (obese) and age (elderly) on everolimus pharmacokinetics are presented in the Electronic supplementary material.

3.4 Exploratory Outcomes: Response assessment Using FDG-PET

Of all performed PET scans, 70 met the acquisition criteria as set by PERCIST. Lesions had sufficient FDG uptake in 20 of 22 patients (91%) with ductal carcinoma, compared to eight of 12 patients (67%) with lobular carcinoma.

ROC analysis showed that of sumSULmax, sumSULpeak, SULpeak high, and sumTLG at day 14 and day 35, the percentage decrease in SULpeak high at day 14 (SULpeak high d14) was the best predictor of PD <3 months. A cut-off of 11% decrease in SULpeak high d14 was optimal (Fig. 3), with an AUC of 0.78, sensitivity 0.78, and specificity 0.84. Patients with PD <3 months had a significantly lower mean decrease in SULpeak high d14 compared to patients without PD <3 months: 7.1% versus 23.0% (p = 0.02). Median PFS was 90 days for patients with <11% decrease in SULpeak high d14, compared to 388 days for patients with >11% decrease, p = 0.0013 (Fig. 4). Seven out of ten patients (70%) with <11% decrease in SULpeak high d14 had PD <3 months, compared to two out of 18 patients (11%) with a decrease >11%, resulting in a relative risk of 6.4.
Fig. 3

Decrease in SULpeak high d14 for patients with and without progressive disease within 3 months. Each bar represents one patient

Fig. 4

Kaplan-Meier estimates of progression free survival, based on differences in metabolic /response at day 14

When using the European Organisation for Research and Treatment of Cancer (EORTC) response criteria of > versus <15% decrease in uptake in SULpeak high d14, median PFS was 440 versus 110 days (p = 0.0039). Using the PERCIST response criteria of > versus <30% decrease and ≥ 0.8 SUL units in SULpeak high d14, median PFS was 440 versus 137 days (p = 0.115).

Using the qualitative assessment according to EORTC criteria, PERCIST criteria, and the consensus reading of two nuclear medicine physicians, the accuracy to predict RECIST classification after three months of treatment was poor, with 0.24, 0.34, and 0.39, respectively. The negative predictive value of a partial metabolic response to predict the absence of progressive disease at three months according to EORTC criteria, PERCIST criteria, and the physician consensus reading was 0.78, 0.88, and 0.70, respectively.

No significant correlation was found between everolimus Ctrough and the decrease in SULpeak high d14 (Spearman’s rho = −0.31, p = 0.108). Dose escalation did not result in a stronger decrease in metabolic activity between day 14 and 35 compared to no dose escalation (n = 8 vs. n = 17).

3.5 Multivariable Analysis to Predict PFS

The total number of previous lines of treatments in the adjuvant and metastatic settings was a significant predictor of PFS. Median PFS time was 341 days for patients with 1 to 3 previous treatments (n = 38) and 90 days for patients with 4 to 6 previous treatments (n = 19) (p = 0.003), HR 0.32 (95% CI 0.15–0.86). Multivariable Cox regression analysis to estimate time to PFS event was performed using dichotomized Ctrough at day 14, SULpeak high d14, and the number of previous treatments as covariates. Based on the plots of the Kaplan-Meier curves (Fig. 2), the HR for the dichotomized variable Ctrough in the Cox proportional hazards model was allowed to be different before and after one year after diagnosis. From these covariates, SULpeak high d14 and the number of previous treatments were significantly associated with PFS (Table 2).
Table 2

Cox proportional hazards model for time to PFS event

 

HR

95% CI

P value

Ctrough at day 14 > or < 12.6 μg/La

0.999

0.904–1.103

0.978

SULpeak high d14

0.960

0.932–0.990

0.087

Number of previous treatments

1.352

0.957–1.910

0.009

a, Ctrough < 12.6 μg/L served as the reference group

SUL, standardized uptake values normalized by lean body weight

4 Discussion

The present study shows that patients with high everolimus trough concentrations (>19.2 μg/L) have a higher chance of clinically relevant toxicity leading to dose changes. It is clinically highly relevant to know that one third of patients are exposed to high everolimus Ctrough, with a nearly threefold increased risk to develop toxicity, as this toxicity can potentially be prevented by dose adjustments.

The upper limit for toxicity that we found is comparable with the results of a recent, independently performed study [11]. As such, we were able to validate the importance of Ctrough as a predictor of toxicity. If early dose reduction in the group of patients with Ctrough > 19.2 μg/L can reduce the dose change risk to the risk of patients with Ctrough < 19.2 μg/L, severe toxicity could potentially be prevented in ~50% of these patients.

Secondly, we show that, as early as 14 days after start of treatment, FDG-PET can identify a subgroup of patients with a high risk of nonresponse to everolimus and exemestane. To our knowledge, this is the first prospective evaluation of the value of FDG-PET to predict outcomes of treatment with everolimus and exemestane in mBC patients. A decrease in SULpeak high d14 of <11% was found to be the best predictor for early PD, but the cut-off as proposed by EORTC (≥15% decrease) was a significant predictor as well. The cut-off used in the PERCIST classification (≥30% decrease) in this study was not a significant predictor. Perhaps the criterion of PERCIST is too stringent for response monitoring of agents that inhibit the activity of oncogenic proteins. If a patient has a poor decline in metabolic activity on FDG-PET at day 14, the treating physician can consider to adjust therapy or to follow this patient extra cautiously. This can aid to give patients optimal outcomes from their anticancer therapy and to avoid side effects and costs of an ineffective treatment. Additionally, lobular carcinomas more frequently had lesions with low FDG activity, consistent with earlier experiences [32], but the majority of lobular carcinomas were suitable for FDG-PET response monitoring.

In our data, Ctrough was no significant predictor of effectiveness. The relatively small sample size of our study increases the chance that we were not able to demonstrate a difference between these groups. A previous study did show a significant relationship between Ctrough and effectiveness [11]. Another study showed no correlation between dose intensity and response rate or PFS, but no everolimus concentrations were measured in that study [33]. The weak relationship between everolimus Ctrough and outcomes can be explained by the presence of exemestane-sensitive tumors that would also have responded without adding everolimus to the treatment. This also can explain the crossing of the PFS survival curves that was seen in our data (Fig. 2). Unfortunately, up to now it cannot be predicted which tumors are exemestane-sensitive and do not need the combination with everolimus treatment. The fact that no relationship was seen between everolimus Ctrough and metabolic response supports the notion that other factors influence the effectiveness of everolimus and exemestane besides everolimus Ctrough.

An additional finding of this study, described in the supplementary file, was that no significant effect of elderly age and obesity was established on everolimus Ctrough. However, due to the small sample size there was limited power to detect a possible difference, and no firm conclusions can be drawn. Elderly patients more frequently had dose changes due to toxicity compared to control patients. This is in contrast with a previous study in which everolimus dose modifications were not seen more frequently in patients of ≥70 years (n = 118) compared to patients <70 years (n = 364) [34].

Furthermore, we have shown that Ctrough is a better predictor for toxicity and effectiveness compared to a full AUC with eight measurements. The importance of Ctrough compared to AUC was also demonstrated in a recent preclinical study showing that continuous low dosing above a certain threshold can be as effective as intermittent dosing with a higher AUC [35]. It has substantial practical benefits if the collection of only one blood sample is sufficient and even superior.

The strong points of our study are the prospective nature of our study with the collection of extensive pharmacokinetic data. Our pharmacokinetic results are in agreement with the findings of several other studies.

Our study also has limitations. Firstly, the sample size of our study is relatively small. Hence, this is an exploratory, hypothesis generating, study. Furthermore, the value of FDG-PET as an early read-out is limited by the fact that very few patients are classified as PD after three months. As such, it was not possible to identify individuals with 100% chance on early progression.

To advance the care for patients with mBC, we need to further individualize treatment in order to improve effectiveness and reduce toxicity. The most important finding to emerge from this study is that monitoring everolimus Ctrough has the potential to identify patients at high risk of developing toxicity. For the future, a patient should have everolimus Ctrough determined seven days after start of treatment, so that toxicity potentially can be prevented through early dose adjustments. A prospective evaluation to study whether therapeutic drug monitoring to reach everolimus drug concentrations in a target window leads to better patient outcomes is currently ongoing (M17 TDM study). Furthermore, after start of treatment a poor decrease in metabolic activity at day 14 on FDG-PET seems to be an important signal of possible ineffective treatment, but these results require further validation.

Notes

Acknowledgements

First of all, we thank all the patients that were willing to participate in this study. Further we would like to thank the project staff of all participating hospitals in acquiring the data.

Compliance with Ethical Standards

Funding

This study was funded by a grant from Novartis (grant number 2878/13). This article was published open access under a Springer Compact agreement with Dutch universities and Academy institutes.

Conflict of Interest

Maaike de Boer received an institutional grant from Novartis for research. Carla M.L. van Herpen and Nielka P. van Erp received an institutional grant from Novartis to perform this study. Annelieke E.C.A.B. Willemsen, Lioe-Fee de Geus-Oei, Jolien Tol, Yvonne Kamm, Paul C. de Jong, Marianne A. Jonker, Allert H. Vos, Willem Grootjans, Johannes W.B. de Groot, Sasja F. Mulder, Erik H.J.G. Aarntzen, and Winald R. Gerritsen declare that they have no conflicts of interest that might be relevant to the contents of this manuscript.

Supplementary material

11523_2018_596_MOESM1_ESM.pdf (166 kb)
ESM 1 (PDF 165 kb)

References

  1. 1.
    Baselga J, Campone M, Piccart M, Burris HA 3rd, Rugo HS, Sahmoud T, et al. Everolimus in postmenopausal hormone-receptor-positive advanced breast cancer. N Engl J Med. 2012;366(6):520–9.CrossRefGoogle Scholar
  2. 2.
    Lousberg L, Jerusalem G. Safety, efficacy, and patient acceptability of everolimus in the treatment of breast Cancer. Breast Cancer. 2016;10:239–52.Google Scholar
  3. 3.
    Li N, Hao Y, Xie J, Lin PL, Koo V, Ohashi E, et al. Effectiveness of verolimus versus endocrine monotherapy or chemotherapy among HR+/HER2- mBC patients with multiple metastatic sites. Clin Ther. 2016;38(4):905–17.CrossRefGoogle Scholar
  4. 4.
    Lin PL, Hao Y, Xie J, Li N, Ohashi E, Koo V, et al. Real-world effectiveness of everolimus-based therapy versus endocrine monotherapy and chemotherapy in patients of HR+/HER2- breast cancer with liver metastasis in the USA. Expert Opin Pharmacother. 2015;16(14):2101–11.CrossRefGoogle Scholar
  5. 5.
    Yardley DA, Noguchi S, Pritchard KI, Burris HA 3rd, Baselga J, Gnant M, et al. Everolimus plus exemestane in postmenopausal patients with HR(+) breast cancer: BOLERO-2 final progression-free survival analysis. Adv Ther. 2013;30(10):870–84.CrossRefGoogle Scholar
  6. 6.
    Jerusalem G, Mariani G, Ciruelos EM, Martin M, Tjan-Heijnen VC, Neven P, et al. Safety of everolimus plus exemestane in patients with hormone-receptor-positive, HER2-negative locally advanced or metastatic breast cancer progressing on prior non-steroidal aromatase inhibitors: primary results of a phase IIIb, open-label, single-arm, expanded-access multicenter trial (BALLET). Ann Oncol. 2016;27(9):1719–25.CrossRefGoogle Scholar
  7. 7.
    US Food and Drug Administration. Everolimus (Afinitor). Clinical pharmacology and biopharmaceutics review. 2008. Available at http://www.accessdata.fda.gov/drugsatfda_docs/nda/2009/022334s000TOC.cfm. Accessed 25th January 2013.
  8. 8.
    O'Donnell A, Faivre S, Burris HA 3rd, Rea D, Papadimitrakopoulou V, Shand N, et al. Phase I pharmacokinetic and pharmacodynamic study of the oral mammalian target of rapamycin inhibitor everolimus in patients with advanced solid tumors. J Clin Oncol. 2008;26(10):1588–95.CrossRefGoogle Scholar
  9. 9.
    Awada A, Cardoso F, Fontaine C, Dirix L, De Greve J, Sotiriou C, et al. The oral mTOR inhibitor RAD001 (everolimus) in combination with letrozole in patients with advanced breast cancer: results of a phase I study with pharmacokinetics. Eur J Cancer. 2008;44(1):84–91.CrossRefGoogle Scholar
  10. 10.
    van Erp NP, van Herpen CM, de Wit D, Willemsen A, Burger DM, Huitema AD, et al. A semi-physiological population model to quantify the effect of hematocrit on Everolimus pharmacokinetics and pharmacodynamics in Cancer patients. Clin Pharmacokinet. 2016;55(11):1447–56.CrossRefGoogle Scholar
  11. 11.
    Deppenweiler M, Falkowski S, Saint-Marcoux F, Monchaud C, Picard N, Laroche ML, et al. Towards therapeutic drug monitoring of everolimus in cancer? Results of an exploratory study of exposure-effect relationship. Pharmacol Res. 2017;121:138–44.CrossRefGoogle Scholar
  12. 12.
    Ravaud A, Urva SR, Grosch K, Cheung WK, Anak O, Sellami DB. Relationship between everolimus exposure and safety and efficacy: meta-analysis of clinical trials in oncology. Eur J Cancer. 2014;50(3):486–95.CrossRefGoogle Scholar
  13. 13.
    Thiery-Vuillemin A, Mouillet G, Nguyen Tan Hon T, Montcuquet P, Maurina T, Almotlak H, et al. Impact of everolimus blood concentration on its anti-cancer activity in patients with metastatic renal cell carcinoma. Cancer Chemother Pharmacol. 2014;73(5):999–1007.CrossRefGoogle Scholar
  14. 14.
    Verheijen RB, Yu H, Schellens JHM, Beijnen JH, Steeghs N, Huitema ADR. Practical recommendations for therapeutic drug monitoring of kinase inhibitors in oncology. Clin Pharmacol Ther. 2017;102(5):765–76.CrossRefGoogle Scholar
  15. 15.
    Hunt CM, Watkins PB, Saenger P, Stave GM, Barlascini N, Watlington CO, et al. Heterogeneity of CYP3A isoforms metabolizing erythromycin and cortisol. Clin Pharmacol Ther. 1992;51(1):18–23.CrossRefGoogle Scholar
  16. 16.
    Blouin RA, Warren GW. Pharmacokinetic considerations in obesity. J Pharm Sci. 1999;88(1):1–7.CrossRefGoogle Scholar
  17. 17.
    Grassi M, Petraccia L, Mennuni G, Fontana M, Scarno A, Sabetta S, et al. Changes, functional disorders, and diseases in the gastrointestinal tract of elderly. Nutr Hosp. 2011;26(4):659–68.PubMedGoogle Scholar
  18. 18.
    Schmucker DL. Liver function and phase I drug metabolism in the elderly: a paradox. Drugs Aging. 2001;18(11):837–51.CrossRefGoogle Scholar
  19. 19.
    Le Couteur DG, McLean AJ. The aging liver. Drug clearance and an oxygen diffusion barrier hypothesis. Clin Pharmacokinet. 1998;34(5):359–73.Google Scholar
  20. 20.
    Hilmer SN, Shenfield GM, Le Couteur DG. Clinical implications of changes in hepatic drug metabolism in older people. Ther Clin Risk Manag. 2005;1(2):151–156.CrossRefGoogle Scholar
  21. 21.
    Honer M, Ebenhan T, Allegrini PR, Ametamey SM, Becquet M, Cannet C, et al. Anti-Angiogenic/vascular effects of the mTOR inhibitor Everolimus are not detectable by FDG/FLT-PET. Transl Oncol. 2010;3(4):264–75.CrossRefGoogle Scholar
  22. 22.
    Nogova L, Boellaard R, Kobe C, Hoetjes N, Zander T, Gross SH, et al. Downregulation of 18F-FDG uptake in PET as an early pharmacodynamic effect in treatment of non-small cell lung cancer with the mTOR inhibitor everolimus. J Nucl Med. 2009;50(11):1815–9.CrossRefGoogle Scholar
  23. 23.
    Cejka D, Kuntner C, Preusser M, Fritzer-Szekeres M, Fueger BJ, Strommer S, et al. FDG uptake is a surrogate marker for defining the optimal biological dose of the mTOR inhibitor everolimus in vivo. Br J Cancer. 2009;100(11):1739–45.CrossRefGoogle Scholar
  24. 24.
    Wei LH, Su H, Hildebrandt IJ, Phelps ME, Czernin J, Weber WA. Changes in tumor metabolism as readout for mammalian target of rapamycin kinase inhibition by rapamycin in glioblastoma. Clin Cancer Res. 2008;14(11):3416–26.CrossRefGoogle Scholar
  25. 25.
    Thomas GV, Tran C, Mellinghoff IK, Welsbie DS, Chan E, Fueger B, et al. Hypoxia-inducible factor determines sensitivity to inhibitors of mTOR in kidney cancer. Nat Med. 2006;12(1):122–7.CrossRefGoogle Scholar
  26. 26.
    Chen JL, Appelbaum DE, Kocherginsky M, Cowey CL, Rathmell WK, McDermott DF, et al. FDG-PET as a predictive biomarker for therapy with everolimus in metastatic renal cell cancer. Cancer Med. 2013;2(4):545–52.CrossRefGoogle Scholar
  27. 27.
    Mortazavi-Jehanno N, Giraudet AL, Champion L, Lerebours F, Le Stanc E, Edeline V, et al. Assessment of response to endocrine therapy using FDG PET/CT in metastatic breast cancer: a pilot study. Eur J Nucl Med Mol Imaging. 2012;39(3):450–60.CrossRefGoogle Scholar
  28. 28.
    Dose Schwarz J, Bader M, Jenicke L, Hemminger G, Janicke F, Avril N. Early prediction of response to chemotherapy in metastatic breast cancer using sequential 18F-FDG PET. J Nucl Med. 2005;46(7):1144–50.PubMedGoogle Scholar
  29. 29.
    Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer. 2009;45(2):228–47.CrossRefGoogle Scholar
  30. 30.
    Janmahasatian S, Duffull SB, Ash S, Ward LC, Byrne NM, Green B. Quantification of lean bodyweight. Clin Pharmacokinet. 2005;44(10):1051–65.CrossRefGoogle Scholar
  31. 31.
    Wahl RL, Jacene H, Kasamon Y, Lodge MA. From RECIST to PERCIST: evolving considerations for PET response criteria in solid tumors. J Nucl Med. 2009;50(Suppl 1):122S–50S.CrossRefGoogle Scholar
  32. 32.
    Fujii T, Yajima R, Kurozumi S, Higuchi T, Obayashi S, Tokiniwa H, et al. Clinical significance of 18F-FDG-PET in invasive lobular carcinoma. Anticancer Res. 2016;36(10):5481–5.CrossRefGoogle Scholar
  33. 33.
    Ciccarese M, Fabi A, Moscetti L, Cazzaniga ME, Petrucelli L, Forcignano R, et al. Dose intensity and efficacy of the combination of everolimus and exemestane (EVE/EXE) in a real-world population of hormone receptor-positive (ER+/PgR+), HER2-negative advanced breast cancer (ABC) patients: a multicenter Italian experience. Breast Cancer Res Treat. 2017;163(3):587–94.CrossRefGoogle Scholar
  34. 34.
    Pritchard KI, Burris HA 3rd, Ito Y, Rugo HS, Dakhil S, Hortobagyi GN, et al. Safety and efficacy of everolimus with exemestane vs. exemestane alone in elderly patients with HER2-negative, hormone receptor-positive breast cancer in BOLERO-2. Clin Breast Cancer. 2013;13(6):421–32 e8.CrossRefGoogle Scholar
  35. 35.
    Laborde L, Oz F, Ristov M, Guthy D, Sterker D, McSheehy P. Continuous low plasma concentrations of everolimus provides equivalent efficacy to oral daily dosing in mouse xenograft models of human cancer. Cancer Chemother Pharmacol. 2017;80(4):869–78.CrossRefGoogle Scholar

Copyright information

© The Author(s) 2018

Open Access This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors and Affiliations

  • Annelieke E. C. A. B. Willemsen
    • 1
  • Lioe-Fee de Geus-Oei
    • 2
    • 3
  • Maaike de Boer
    • 4
  • Jolien Tol
    • 5
  • Yvonne Kamm
    • 6
  • Paul C. de Jong
    • 7
  • Marianne A. Jonker
    • 8
  • Allert H. Vos
    • 9
  • Willem Grootjans
    • 2
  • Johannes W. B. de Groot
    • 10
  • Sasja F. Mulder
    • 1
  • Erik H. J. G. Aarntzen
    • 11
  • Winald R. Gerritsen
    • 1
  • Carla M. L. van Herpen
    • 1
  • Nielka P. van Erp
    • 12
  1. 1.Department of Medical Oncology, Radboud Institute for Health SciencesRadboud university medical centerNijmegenthe Netherlands
  2. 2.Department of RadiologyLeiden University Medical Center (LUMC)Leidenthe Netherlands
  3. 3.Biomedical Photonic Imaging Group, MIRA InstituteUniversity of TwenteEnschedethe Netherlands
  4. 4.Division of Medical Oncology, Department of Internal Medicine, GROW-School of Oncology and Developmental BiologyMaastricht University Medical CenterMaastrichtthe Netherlands
  5. 5.Department of Medical OncologyJeroen Bosch Hospital‘s-Hertogenboschthe Netherlands
  6. 6.Department of Medical OncologyMaasziekenhuis PanteinBoxmeerthe Netherlands
  7. 7.Department of Medical OncologySt. Antonius ZiekenhuisUtrechtthe Netherlands
  8. 8.Department for Health EvidenceRadboud university medical centerNijmegenthe Netherlands
  9. 9.Department of Medical OncologyBernhoven ZiekenhuisUdenthe Netherlands
  10. 10.Isala Oncology CenterZwollethe Netherlands
  11. 11.Department of Radiology and Nuclear MedicineRadboud university medical centerNijmegenthe Netherlands
  12. 12.Department of Pharmacy, Radboud Institute for Health SciencesRadboud university medical centerNijmegenthe Netherlands

Personalised recommendations