Deregulation of mammalian target of rapamycin (mTOR) signalling is common in human hepatocellular carcinoma (HCC).
Deregulation of mammalian target of rapamycin (mTOR) signalling is common in human hepatocellular carcinoma (HCC).
To determine the maximum tolerated dose (MTD) of the oral mTOR inhibitor everolimus in advanced HCC patients.
Patients with locally advanced or metastatic HCC (Child-Pugh class A or B) were enrolled in an open-label phase 1 study and randomly assigned to daily (2.5–10 mg) or weekly (20–70 mg) everolimus in a standard 3 + 3 dose-escalation design. MTD was based on the rate of dose-limiting toxicities (DLTs). Secondary endpoints included safety, pharmacokinetics and tumour response. In a post hoc analysis, serum hepatitis B virus (HBV) DNA levels were quantified.
Thirty-nine patients were enrolled. DLTs occurred in five of 21 patients in the daily and two of 19 patients in the weekly cohort. Daily and weekly MTDs were 7.5 mg and 70 mg respectively. Grade 3/4 adverse events with a ≥10% incidence were thrombocytopenia, hypophosphataemia and alanine transaminase (ALT) elevation. In four hepatitis B surface antigen (HBsAg)-seropositive patients, grade 3/4 ALT elevations were accompanied by significant (>1 log) increases in serum HBV levels. The incidence of hepatitis flare (defined as ALT increase >100 IU/mL from baseline) in HBsAg-seropositive patients with and without detectable serum HBV DNA before treatment was 46.2% and 7.1% respectively (P < 0.01, Fisher exact test). Disease control rates in the daily and weekly cohorts were 71.4% and 44.4% respectively.
The recommended everolimus dosing schedule for future hepatocellular carcinoma studies is 7.5 mg daily. Prophylactic anti-viral therapy should be mandatory for HBsAg-seropositive patients (ClinicalTrials.gov NCT00390195).
In 2008, it is estimated that approximately 748 000 people were diagnosed with liver cancer and that approximately 695 000 individuals died of the disease globally. Approximately 85–90% of primary liver cancers manifest as hepatocellular carcinoma (HCC). Most patients with HCC present with co-existing cirrhosis caused primarily by chronic hepatitis B virus (HBV) and/or hepatitis C virus (HCV) infection, chronic alcohol abuse or metabolic syndrome.[2, 3]
Therapies with curative potential, such as liver resection, liver transplantation and local ablation, are limited to patients with early-stage HCC.[4-6] However, the majority of patients present with advanced or metastatic disease. In these patients, systemic cytotoxic chemotherapy provides only a modest benefit.[3, 4] HCC is a complex, heterogeneous, highly vascularised tumour with evidence of aberrant activation of multiple signaling cascades.[2, 3] Therapies targeting these molecular pathways provide a new approach to address the unmet need of effective therapy for advanced HCC. Sorafenib, a multikinase inhibitor that targets tyrosine kinase receptors, including the vascular endothelial growth factor receptor (VEGFR), is currently the only targeted agent recommended for the treatment of patients with advanced HCC.[5, 6] Although sorafenib provides significant clinical benefit compared with no treatment, it is associated with a limited survival gain [overall survival (OS) 10.7 months vs. 7.9 months with placebo in the phase 3 SHARP trial and 6.5 months vs. 4.2 months with placebo in the phase 3 Asia-Pacific trial], low tumour response rate (2–3% partial response rate) and a high incidence of diarrhoea and hand-foot skin reaction.[8, 9]
Mammalian target of rapamycin (mTOR) is a key protein kinase that regulates cell growth, proliferation, metabolism and angiogenesis. Upregulation of mTOR signalling has been observed in 40–45% of patients with HCC,[11, 12] and in HepG2 cells, elevated levels of phosphorylated mTOR were correlated with increased proliferation. Preclinical studies showed that inhibition of mTOR with rapamycin not only reduced HepG2 proliferation, but also inhibited tumour growth and metastasis and improved survival in mouse and rat models of HCC via both antitumour and anti-angiogenic effects.[13-15] Everolimus, an oral mTOR inhibitor, also has been shown to decrease HCC cell viability in vitro and to inhibit the growth of both cell line-derived and patient tissue-derived HCC tumours in murine xenograft models.[16, 17] Clinically, mTOR inhibitor-based immunosuppression has been shown to reduce the risks of HCC recurrence and de novo tumour development, avoid rejection and improve OS in patients with HCC who undergo liver transplantation.[18, 19]
Everolimus has been extensively evaluated in patients with advanced solid tumours and is currently approved in various countries for the treatment of postmenopausal women with hormone receptor-positive, HER2-negative breast cancer in combination with exemestane after failure of treatment with letrozole or anastrozole; adults with metastatic renal cell carcinoma (mRCC) that progressed on previous VEGF-targeted therapy; adults with progressive pancreatic neuroendocrine tumours (pNET) that are unresectable, locally advanced or metastatic; adults with renal angiomyolipoma associated with tuberous sclerosis complex (TSC) not requiring immediate surgery; and paediatric and adult patients with unresectable subependymal giant cell astrocytomas associated with TSC that require intervention. The recommended everolimus dose in patients with breast cancer, mRCC and pNET is 10 mg daily. However, given that the everolimus half-life is significantly prolonged in patients with impaired liver function compared with healthy volunteers, everolimus 10 mg/day may not be the optimal dose in patients with HCC. The primary objective of this randomised, phase 1, dose-escalation study was to define the maximum tolerated dose (MTD) and optimal schedule of everolimus in patients with advanced HCC.
This was an open-label, parallel-group, noncomparative, randomised phase 1 study of daily vs. weekly schedules of everolimus (Afinitor; Novartis Pharma AG, Basel, Switzerland) for patients with advanced HCC conducted at National Cheng Kung University Hospital and Tri-Service General Hospital in Taiwan (ClinicalTrials.gov identifier NCT00390195). A standard 3 × 3 dose-escalation design was used. Permuted block randomisation with a block size of 6 was used to assign eligible patients to either daily or weekly everolimus. Randomisation was performed at the Bio-statistical Center of Taiwan Cooperative Oncology Group (TCOG) under the direction of one of the authors (CF Hsiao). The planned everolimus doses were 2.5, 5.0, 7.5 and 10 mg for the daily schedule and 20, 30, 50 and 70 mg for the weekly schedule. Dose-limiting toxicities (DLTs) were defined as any nonhaematological toxicity of grade ≥3 excluding nausea, vomiting and alopecia; grade 4 neutropenia lasting >7 days; neutropenic fever; or grade 4 thrombocytopenia. Enrolment was initiated at the lowest dose level. In the absence of a DLT, the subsequent cohort of patients was randomised to the next dose level. If one out of three patients developed a DLT, three additional patients were treated at the same dose level. If none of the three additional patients experienced a DLT, three patients were enrolled at the next dose level. The MTD was defined as the dose level immediately below the level at which two or more patients of a dose level of up to six patients experienced a DLT. If one or fewer patients experienced a DLT at the last dose level of either dosing schedule, the last dose level was considered to be the MTD.
This study was approved by the institutional review boards of the participating institutions and the Department of Health, Executive Yuan, Taiwan.
Eligible patients were 20–75 years of age with locally advanced or metastatic HCC that was measurable and progressed after or was too advanced for definitive local therapy (i.e. surgical resection, radiofrequency ablation, percutaneous ethanol or acetic acid injection) or transcatheter arterial chemoembolisation. The HCC diagnosis had to be established by cytology or histopathology or, in patients with cirrhosis of the liver or chronic HBV or HCV infection, characteristic radiographic findings with α-fetoprotein (AFP) ≥400 ng/mL. Additional eligibility criteria included an Eastern Cooperative Oncology Group (ECOG) performance score ≤2; Child-Pugh score ≤9; serum levels of total bilirubin ≤2.0 mg/dL, alanine transaminase (ALT) ≤2.5 times the upper limit of normal (ULN) and creatinine ≤2.0 × ULN; white blood cell count ≥3000/μL; and platelet count ≥50 000/μL. Prior chemotherapy or targeted therapy were allowed but had to be completed ≥4 weeks before study entry. Other exclusion criteria included the presence of uncontrolled inter-current illness, a history of other malignancies within 3 years of study entry, previous therapy with a rapamycin analogue or use of inhibitors or inducers of P-glycoprotein, CYP3A4 or CYP3A5 within 2 weeks of study initiation. All patients provided written informed consent.
Everolimus was given at the specified dose and schedule in continuous 4-week cycles until disease progression, unacceptable toxicity, patient refusal, withdrawal of informed consent or death. For patients who experienced a DLT or significant toxicity, dose reduction was allowed for subsequent treatment cycles. Physical examination and laboratory tests, including serum levels of albumin/globulin, aspartate transaminase, ALT, alkaline phosphatase, gamma-glutamyltranspeptidase and total/direct bilirubin; fasting blood sugar, blood urea nitrogen, creatinine and electrolytes; complete blood count with differential count, prothrombin time/activated partial thrombin time; and urinanalysis were performed once weekly during the first treatment cycle, once every 2 weeks during the second and third treatment cycles and once per treatment cycle thereafter. The serum level of AFP was assessed at baseline and every 4 weeks during treatment. Prophylactic anti-HBV nucleoside/nucleotide therapy and periodic monitoring of serum levels of HBV DNA were not specifically recommended in the study protocol because the risk of HBV reactivation was not fully appreciated before study initiation.
Chest X-rays and computed tomography of the abdomen were performed before treatment and repeated every 8 weeks thereafter or whenever clinically indicated. Additional imaging was performed 4 weeks after the first radiological evidence of tumour response, as determined by an independent radiologist blinded to treatment assignment, according to the Response Evaluation Criteria in Solid Tumors (RECIST).
Adverse events (AEs) were collected throughout the study and graded according to the National Cancer Institute Common Toxicity Criteria, Version 3.0. Because the baseline ALT level could be as high as 2.5 × ULN and ALT fluctuation is common in patients with HCC, the grading of ALT toxicity was modified based on the absolute ALT increase above baseline: absolute increase of ≤100 IU/mL = grade 1, 101–200 IU/mL = grade 2, 201–800 IU/mL = grade 3 and ≥801 IU/mL = grade 4. To determine whether the baseline serum HBV DNA level was associated with the occurrence of grade 3/4 ALT elevations in hepatitis B surface antigen (HBsAg)-seropositive patients, a reverse transcriptase-polymerase chain reaction (RT-PCR) assay (COBAS TaqMan; Roche Diagnostics, Laval, Quebec, Canada) was used for post hoc quantification of the HBV DNA level in stored serum samples. The lower limit of detection was 48 IU/mL.
Blood samples for pharmacokinetic analysis in the daily schedule were collected predose on days 1, 27, 28 and 29 and 1, 1.5, 2, 2.5, 3, 4, 6, 8, 12 and 24 h postdose on days 1 and 29. For the weekly schedule, blood samples were collected predose at weeks 1, 3, 4 and 5 and 1, 2, 3, 4, 6, 8, 12, 24, 72 and 168 h postdose at weeks 1 and 5. A predose 5-mL blood sample was drawn into an anticoagulant-free collection tube to yield 2 mL of serum for protein binding measurements. At the remaining time points, 2 mL of EDTA-containing blood were obtained. Samples were transferred to polypropylene tubes and frozen at −70°C until analysis. Levels of everolimus were determined using a fluorescent polarisation immunoassay (INNOFLUOR CERTICAN assay system; Seradyn, Inc., Indianapolis, IN, USA) based on the competitive binding principle. The pharmacokinetic parameters of area under the time–concentration curve from time 0 to time t (AUC0−t) and from time 0 to infinity (AUC0−∞), the maximum (Cmax), minimum (Cmin) and average (Cave, AUC/dosing interval) plasma concentration at steady state, elimination half-life (t1/2) and time to maximum plasma concentration (Tmax) were determined using noncompartmental approaches.
Genomic DNA was prepared from peripheral mononuclear cells obtained from 5-mL EDTA-containing blood samples. PCR products of genomic DNA were subjected to direct sequencing to study polymorphisms of P-glycoprotein (MDR1) and CYP3A4 and CYP3A5 alleles. The pharmacogenetic results were correlated with the pharmacokinetic parameters.
The primary study endpoint was the MTD of the daily and weekly schedules of everolimus. The secondary endpoints were DLTs, objective tumour response per RECIST, everolimus pharmacokinetics and correlation with pharmacogenetics and safety. Post hoc analyses of progression-free survival (PFS) and OS were performed to compare the efficacy of the daily and weekly dosing schedules. All patients were evaluated for efficacy and safety (intent to treat principle). The Chi-squared test and Fisher exact test were used to compare descriptive variables, and the Wilcoxon Mann–Whitney approach was used to compare continuous variables. Any P-value <0.05 (two-sided test) was considered statistically significant.
This was an investigator-initiated study. Novartis Pharmaceuticals provided study drug and funding support, and grant support was provided by the Department of Health, Executive Yuan, Taiwan (DOH99-TD-C-111-004). All the randomisation and data analyses were processed in the TCOG under the supervision of one of the authors (C.-F. Hsiao). The corresponding authors have full access to all the data in the study and have final responsibility for the decision of submission for publication.
Between 7 December 2006 and 2 February 2009, 39 patients were enrolled in the study and randomly allocated to the daily (n = 21) and weekly (n = 18) dose schedules (Figure 1). All patients received their allocated dose and were eligible for analysis of the MTD. At the time of analysis (3 September 2009), all but two patients discontinued treatment, most commonly due to disease progression (n = 15 in the daily schedule and n = 16 in the weekly schedule). Demographic and baseline disease characteristics were generally consistent between the two schedules, except that all five patients with decompensating liver function (Child-Pugh class B) were assigned to the daily schedule (Table 1). All patients had Barcelona Clinic Liver Cancer (BCLC) stage C disease. Thirty-four patients (87.2%) had failed prior therapies, including 15 (38.5%) who failed prior chemotherapy or targeted therapy. Seropositivity for HBsAg and the anti-HCV antibody were observed in 27 (69.2%) patients and 15 (38.5%) patients respectively. Only one HBsAg-seropositive patient used anti-viral therapy at study entry.
|Characteristics||Daily cohort (n = 21)||Weekly cohort (n = 18)||All patients (N = 39)|
|Sex, n (%)|
|Male||19 (90.5)||17 (94.4)||36 (92.3)|
|Female||2 (9.5)||1 (5.6)||3 (7.7)|
|Age, years, median (range)||59 (28–75)||59 (34–73)||59 (28–75)|
|ECOG PS, n (%)|
|0||1 (4.8)||0||1 (2.6)|
|1||15 (71.4)||15 (83.3)||30 (76.9)|
|2||5 (23.8)||3 (16.7)||8 (20.5)|
|Seropositivity, n (%)|
|HBsAg||12 (57.1)||9 (50.0)||21 (53.8)|
|Anti-HCV||6 (28.6)||3 (16.7)||9 (23.1)|
|Both||1 (4.8)||5 (27.8)||6 (15.4)|
|Neither||2 (9.5)||1 (5.6)||3 (7.7)|
|Child-pugh class, n (%)|
|A||16 (76.2)||18 (100)||34 (87.2)|
|B||5 (23.8)||0||5 (12.8)|
|Serum α–fetoprotein level, n (%)|
|<200 ng/mL||5 (23.8)||7 (38.9)||12 (30.8)|
|≥200 ng/mL||16 (76.2)||11 (61.1)||27 (69.2)|
|Baseline laboratory values, median (range)|
|Total bilirubin, μmol/L||12.0 (5.1–34.2)||12.0 (6.8–30.8)||12.0 (5.1–34.2)|
|Albumin, g/dL||3.6 (2.8–4.5)||3.9 (2.8–4.5)||3.8 (2.8–4.5)|
|ALT, U/L||36 (16–134)||47 (10–114)||40 (10–134)|
|INR||1.10 (0.88–1.40)||1.12 (0.94–1.29)||1.10 (0.88–1.40)|
|Creatinine, μmol/L||70.7 (35.4–123.8)||79.6 (44.2–123.8)||70.7 (35.4–123.8)|
|Major vascular invasion, n (%)|
|Absent||11 (52.4)||12 (66.7)||23 (59.0)|
|Present||10 (47.6)||6 (33.3)||16 (41.0)|
|Modified TNM stage, n (%)|
|III||11 (52.4)||5 (27.8)||16 (41.0)|
|IV||10 (47.6)||13 (72.2)||23 (59.0)|
|Okuda stage, n (%)|
|I||12 (57.1)||15 (83.3)||27 (69.2)|
|II||7 (33.3)||3 (16.7)||10 (25.6)|
|III||2 (9.5)||0||2 (5.1)|
|CLIP score, n (%)|
|0–1||6 (28.6)||6 (33.3)||12 (30.8)|
|2–3||9 (42.9)||11 (61.1)||20 (51.3)|
|4–5||6 (28.6)||1 (5.6)||7 (17.9)|
|BCLC stage, n (%)|
|C||21 (100)||18 (100)||39 (100)|
|Prior treatment, n (%)|
|None||2 (9.5)||3 (16.7)||5 (12.8)|
|Surgery||11 (52.4)||8 (44.4)||19 (48.7)|
|RFA/PEI||3 (14.3)||6 (33.3)||9 (23.1)|
|TACE||15 (71.4)||12 (66.7)||27 (69.2)|
|Radiotherapy||10 (47.6)||3 (14.3)||13 (33.3)|
|Systemic chemotherapy||5 (23.8)||4 (22.2)||9 (23.1)|
|Molecular targeting agents||7 (33.3)||5 (27.8)||12 (30.8)|
|≥3 treatments||9 (42.9)||9 (50.0)||18 (46.2)|
In the daily everolimus schedule, DLTs were observed in one patient each enrolled at the 2.5-, 5- and 7.5-mg dose levels and two patients in the 10-mg dose level (Figure 1). The DLTs included grade 3 diarrhoea (n = 3), grade 3 bilirubin elevation (n = 1) and grade 4 thrombocytopenia (n = 1). Given the occurrence of DLTs in two patients at the 10-mg/day dose level, the MTD for daily everolimus dosing was determined to be 7.5 mg. DLTs were reported by one patient each enrolled at the weekly 30- and 70-mg dose levels (grade 3 ALT elevation and grade 3 infection, n = 1 each) (Figure 1). The MTD for weekly everolimus dosing was determined to be 70 mg, the maximum dose assessed in this study.
The median duration of everolimus exposure was 16 weeks (range: 2.7−36.0 weeks) in the daily everolimus schedule and 12 weeks (range: 4.4−57.2 weeks) in the weekly everolimus schedule. AEs and laboratory abnormalities led to dose reductions in nine patients (23.7%) and dose interruptions in 16 patients (41.0%). Across all dose levels in both schedules, the most common treatment-related AEs were myelosuppression, hypophosphataemia, hyperglycaemia, proteinuria, ALT elevation, skin rash, mucositis, anorexia and fatigue (Table 2). In both the daily and weekly schedules, most AEs were of grade 1 or 2 severity and manageable at a dose at or below the MTD. At both the daily and weekly MTDs, grade 3/4 thrombocytopenia and hypophosphataemia were observed in 33% of patients each; grade 3/4 diarrhoea, hyperglycaemia, hyponatraemia and proteinuria occurred in 17–33% of patients in the daily MTD schedule, but in no patients in the weekly schedule. Three patients, two with HBV and one with HCV, experienced cirrhotic progression during the study. The progression of one of the patients with HBV was thought to be due to HCC progression, not everolimus treatment. The other two cases occurred during everolimus treatment.
|2.5 mg n = 6||5.0 mg n = 6||7.5 mg n = 6||10 mg n = 3||20 mg n = 3||30 mg n = 6||50 mg n = 3||70 mg n = 6|
Grade 3/4 ALT elevation occurred in two patients in the weekly schedule and three patients in the daily schedule. Among these five patients, one was anti-HCV seropositive and four were HBsAg seropositive. Although not prespecified in the protocol, HBV and HCV viral loads were measured when ALT reached ≥200 IU/mL during everolimus treatment. In the patient seropositive for anti-HCV, the ALT increase was not associated with a concurrent elevation of the HCV RNA level. In the four HBsAg-seropositive patients, ALT elevations were accompanied by a >1-log increase in the serum HBV DNA level and were considered to be episodes of HBV flare. The median peak level of serum ALT in these four patients was 470 IU/mL (range: 255–656 IU/mL). Only one patient was symptomatic (grade 3 hyperbilirubinaemia). HBV reactivation during everolimus treatment occurred within 4–6 weeks in two patients and within 12–16 weeks in two patients (Figure 2). In all four HBsAg-seropositive patients, everolimus was withheld and lamivudine 100 mg/day was initiated upon the detection of HBV reactivation. Within 4–8 weeks of lamivudine initiation, the HBV DNA level declined and the ALT level returned to baseline. Two patients (cases #26 and #29) experienced radiographically confirmed disease progression within one month of withholding everolimus and were removed from the study. The other two patients (cases #9 and #38) received lamivudine and everolimus concurrently until disease progression without any further episodes of HBV reactivation.
After a single everolimus dose given on day 1 of cycle 1, a linear correlation between the everolimus dose and AUC0−24 was observed over the full dose range (Spearman's r = 0.937; P < 0.0001). Because of rapid disease progression, two patients discontinued the study after 4 weeks of treatment, leaving 37 patients who were available for pharmacokinetic analysis at week 5 (day 29). The steady state pharmacokinetic parameters on day 29 are listed in Table 3. Cmax, Cmin and AUC0−24 increased in an approximately dose-proportional manner over the full dose range in the daily cohort, whereas the increases in serum concentration and AUC0−72 were less than dose proportional in the weekly cohort (Figure S1). The concentrations of 168-h samples in the weekly cohort were below the lower level of quantitation limits. The 24-h and 168-h AUC curves at week 5 were used to calculate the apparent oral clearance (CL/F) of everolimus in the daily and weekly schedules respectively. Everolimus CL/F was not affected by dose in the daily schedule, but showed a decreasing trend over the 20- to 50-mg weekly doses (Table 3). The CL/F showed no association with CYP3AP1*3 and ABCB1 genotypes among tested individuals. Among the 12 patients who received everolimus 2.5 mg/day or 5.0 mg/day, the CL/F in patients of Child-Pugh class A (n = 7) and B (n = 5) were 7372.88 ± 5050.84 mL/h and 4047.24 ± 727.22 mL/h respectively (P = 0.18; Student t-test, two-sided).
|Dosing schedule||n||Cmax (ng/mL)||Cmin (ng/mL)||AUC0−ta (ng·h/mL)||Oral clearance (mL/h)||t1/2 (h)||Tmax (h)|
|2.5 mg||6||41.8 ± 14.2||14.6 ± 3.4||514.9 ± 127.4||5164.5 ± 1536.5||38.9 ± 22.6||1.7 ± 1.0|
|5.0 mg||6||106.1 ± 50.0||29.7 ± 15.9||1234.1 ± 598.3||6809.9 ± 5788.1||21.5 ± 6.3||1.5 ± 0.5|
|7.5 mg||7||147.4 ± 51.3||41.5 ± 21.1||1643.2 ± 560.5||5162.5 ± 2255.9||40.4 ± 26.4||1.5 ± 0.4|
|20.0 mg||4||550.6 ± 142.9||50.4 ± 17.1||7069.2 ± 1037.4||3335.9 ± 990.0||88.5 ± 14.1||1.0 ± 0.0|
|30.0 mg||5||684.2 ± 223.7||73.6 ± 26.2||10625.6 ± 1942.0||2897.5 ± 508.1||98.7 ± 28.6||1.2 ± 0.4|
|50.0 mg||3||2247.9 ± 478.0||184.5 ± 65.5||31329.4 ± 6022.2||1633.1 ± 290.4||112.8 ± 61.9||1.0 ± 0.0|
|70.0 mg||5||1334.9 ± 700.0||102.0 ± 75.5||18101.3 ± 9037.1||4421.0 ± 2016.4||106.8 ± 79.1||1.0 ± 0.0|
The best overall tumour response as determined by independent radiologist review in the total population (N = 39) was partial response in one patient (2.6%; everolimus dose, 20 mg/week), stable disease in 22 patients (56.4%) and progressive disease in 16 patients (41.0%; includes 1 patient whose response was non-evaluable). The disease control rate (complete/partial response + stable disease) was 71.4% in the daily schedule [n = 15; 95% confidence interval (CI), 52.1–90.7%] and 44·4% in the weekly schedule (n = 8; 95% CI, 21.4–67.4%). Overall, a >30% reduction in serum AFP was observed in 10 of 27 patients (37.0%) with a baseline serum AFP level ≥200 ng/mL, including 5 of 16 patients (31.2%) in the daily schedule and 5 of 11 patients (45.4%) in the weekly schedule. Eight of the 10 patients (80.0%) who experienced a >30% reduction in AFP had a best tumour response of partial response or stable disease. Waterfall plots of change in tumour size at best response and change in serum AFP level after 2 cycles of treatment are shown in Figure S2. Overall median PFS was 16.0 weeks (95% CI: 11.0−21.0 weeks) in the daily schedule and 8.3 weeks (95% CI: 0−19.5 weeks) in the weekly schedule (Figure 3a). Median OS was 33.4 weeks (95% CI: 9.2−57.6 weeks) and 24.6 weeks (95% CI: 0−53.5 weeks) in the daily and weekly schedules respectively (Figure 3b).
In the current study, we defined the MTD of everolimus in patients with advanced HCC to be 7.5 mg daily or 70 mg weekly and provide clinical evidence that everolimus may induce HBV reactivation and hepatitis flare in nontransplanted patients. Of note, the optimal dose of everolimus in the present population of patients with advanced HCC is 7.5 mg/day, which is lower than the 10-mg/day dose identified in patients with other solid malignancies[23-25] and in the phase 1/2 study of patients with advanced HCC performed by Zhu et al. The lower optimal dose of everolimus identified in the present study compared with the study by Zhu et al. is likely a reflection of the fact that patients in this study had more advanced disease than patients in the study by Zhu et al. More patients in the present study had an ECOG performance status of 1 (76.9% vs. 57.1%) or 2 (20.5% vs. 3.6%), HBsAg seropositivity (57.1% vs. 17.9%), BCLC stage C disease (100% vs. 93%) and a Cancer of the Liver Italian Program (CLIP) score of 4–5 (17.9% vs. 0%).
The most common treatment-related AEs observed in the present study were haematological toxicities, mucositis, skin rash, diarrhoea, fatigue, anorexia and laboratory abnormalities. This AE profile is largely consistent with the AE profile observed for everolimus in the study by Zhu et al., as well as the AE profiles observed in other studies of everolimus for patients with advanced solid tumours,[23-25] including renal cell carcinoma, pancreatic neuroendocrine tumours and hormone receptor-positive breast cancer. Most of the AEs were of grade 1 or 2 severity and manageable. However, grade 3/4 AEs were more common in our daily schedule than in the other phase 1 studies of everolimus. Of note, the frequencies of clinically relevant grade 3/4 thrombocytopenia (26.7%), diarrhoea (20.0%), anaemia (13.3%) and ALT elevation (13.3%) in the 15 patients in the present study who received everolimus 5–10 mg/day were higher than those observed in the studies by O'Donnell et al. (n = 37), Tabernero et al. (n = 24) and Okamoto et al. (n = 6).[23-25] The high incidence of such significant AEs in our population of patients with advanced HCC is likely a result of underlying chronic liver diseases and related complications, the altered pharmacokinetics of everolimus in patients with hepatic insufficiency or a combination of the two. These factors may also limit the ability to combine everolimus with other targeted therapies. For example, the phase 2 portion of a study of everolimus plus sorafenib for untreated advanced HCC (ClinicalTrials.gov identifier NCT00828594) was not conducted due to the toxicity observed in the phase 1 dose-finding portion of the study despite the fact that preclinical studies suggested antitumour synergism between these two therapies. Therefore, the partner for everolimus in clinical studies of combination therapy for advanced HCC should be carefully selected to maximise antitumour synergism and minimise the overlap of the safety profiles.
In the present study, grade 3/4 ALT elevations that occurred in 4 HBsAg-seropositive patients were accompanied by >1-log increases of serum HBV DNA levels and considered to be episodes of HBV hepatitis flare. HBV hepatitis flare is not an uncommon AE in HBV carriers receiving systemic chemotherapy or immunosuppressant therapy. Because everolimus is known to have immunosuppressive properties, we were surprised to note after an extensive literature review that there are very few other reports of everolimus-associated HBV hepatitis flare[28, 33] and that this appears to be the first detailed report of clinically relevant everolimus-associated HBV hepatitis flare. In recipients of chemotherapy or immunotherapy, most HBV hepatitis flares occur during drug holidays or even weeks or months after therapy is completed; thus, it is likely that these events are caused by rebounded cytotoxic T cells attacking hepatocytes with enhanced HBsAg expression, thereby causing liver parenchyma damage. Conversely, all the grade 3/4 ALT elevations in our HBsAg-seropositive patients occurred during everolimus treatment, including three patients receiving continuous daily dosing of everolimus and one patient on the weekly schedule. In two in vitro studies, Guo et al. and Teng et al. demonstrated that inhibition of the phosphatidylinositol 3-kinase/Akt/mTOR-signaling pathway could enhance the transcription of HBV RNA in HBV-expressing HepG2.2.15 cells and the expression of HBV surface protein in Huh-7 cells. Given the immunosuppressive nature of everolimus, we speculate that the observed everolimus-induced HBV hepatitis flare may represent early-stage fibrosing cholestatic or cytolytic hepatitis, in which the hepatocelluar damage is caused by cytopathic effects of highly expressed HBV proteins or vial particles within hepatocytes of immunocompromised patients[36-39] and detected by the rigorous liver function monitoring mandated as part of the clinical trial. In contrast with the majority of HBV-related fibrosing cholestatic hepatitis observed in post-transplant patients, which is frequently diagnosed at a late stage and associated with fatal outcomes, all events observed in this study resolved within 4–8 weeks of withholding everolimus and the concomitant administration of anti-HBV nucleotide/nucleoside therapy. Post hoc analysis showed the incidence of hepatitis flare in the present study to be significantly higher in HBsAg-seropositive patients with detectable baseline serum HBV DNA levels (46.2% vs. 7.1% of patients without detectable HBV DNA at baseline; P < 0.01, Fisher exact test). Based on our findings, prophylactic anti-HBV therapy and close monitoring of HBV DNA and ALT levels during treatment should be mandatory for HBsAg-seropositive patients who receive everolimus. In the phase 1/2 study of everolimus for advanced HCC conducted by Zhu et al., all patients with HBV received anti-viral therapy, and no HBV reactivation was observed.
In the current study, everolimus was rapidly absorbed across all dose levels, with a Tmax of 1–2 h. Similar to what was observed in other studies of everolimus for patients with solid malignancies,[23-25] steady state Cmax, Cmin and AUC0−t were largely dose-dependent in both the daily and weekly dosing schedules. Our data show that the steady state CL/F of everolimus in patients with HCC was lower than that observed in non-HCC patients, whereas the t1/2, Cmax and AUC0−24 were higher.[23-25] These findings concur with a study reported by Kovarik et al., which compared the pharmacokinetics of a single 2-mg dose of everolimus in healthy subjects and patients with hepatic impairment (Child-Pugh B) and suggest that the oral clearance of everolimus is significantly reduced in patients with hepatic dysfunction. Although in the present study there was no significant difference in everolimus CL/F in the patients with Child-Pugh A (n = 7) and B (n = 5) who received everolimus 2.5–5.0 mg/day, the analysis may be underpowered by the limited number of patients. The true impact of Child-Pugh stage on the pharmacokinetics of everolimus 7.5 mg/day in patients with advanced HCC will be further evaluated in an ongoing phase 2 trial (ClinicalTrials.gov identifier NCT00390195).
Efficacy data from this study show that everolimus is moderately active in stabilising the progression of advanced HCC. Although this study was not designed to compare the efficacy of the daily and weekly dosing schedules, the disease control rate was higher in the daily vs. weekly schedule (71.4% vs. 44.4%), and median PFS [16.0 weeks (3.7 months) vs. 8.3 weeks (1.9 months)] and OS [33.4 weeks (7.7 months) vs. 24.6 weeks (5.7 months)] were longer. The median PFS and OS observed in the daily schedule of the present study were nearly identical to those reported in a phase 1/2 study of everolimus for advanced HCC (3.8 months and 8.4 months, respectively) and longer than the median time to progression (3.0 months) and OS (5.3–6.5 months) observed in small studies of rapamycin for advanced HCC.[41, 42] Of note, the median OS observed for the daily schedule in the present study (7.7 months) was longer than the median OS observed for sorafenib- and placebo-treated patients in the phase 3 Asia-Pacific study of sorafenib for advanced HCC (6.5 months and 4.2 months, respectively). In comparison with the present study, the median OS in the phase 3 SHARP study was longer for sorafenib-treated patients (10.7 months) and similar for placebo-treated patients (7.9 months).8 However, comparison with the Asia-Pacific trial is likely more relevant as this patient population is more comparable to that of the present study. Overall and considering that all patients enrolled in the present study had locally advanced or metastatic disease and most were symptomatic (97.4% with ECOG performance status 1–2) and heavily pre-treated, including 23.1% and 30.8% who received prior systemic chemotherapy and molecularly targeted therapy respectively, the observed clinical activity is encouraging.
In conclusion, the results of the present study identified the daily and weekly MTDs of everolimus for patients with advanced HCC to be 7.5 mg and 70 mg respectively. At these doses, everolimus demonstrated acceptable tolerability and preliminary evidence of clinical activity. Because patients in the daily schedule had a higher disease control rate and longer median PFS and OS than patients in the weekly schedule, everolimus 7.5 mg daily is recommended for future studies conducted in advanced HCC. Based on the HBV flare observed in HBsAg-seropositive patients, prophylactic anti-viral therapy and close monitoring of HBV DNA and ALT levels should be mandatory for HBsAg-seropositive patients who receive everolimus. The efficacy and safety of everolimus 7.5 mg daily is being compared with that of placebo, both given with best supportive care, in patients with advanced HCC that progressed after sorafenib or who are sorafenib intolerant in the ongoing international, randomised, phase 3 EVOLVE-1 trial (ClinicalTrials.gov identifier NCT01035229).
Guarantor of the article: L.-T. Chen.
Author contributions: All authors participated in the design of the study and contributed to the writing of the manuscript. H.-S. Shiah, C.-Y. Chen, Y.-J. Lin, W.-C. Su, J.-Y. Chang, J. Whang-Peng, P.-W. Lin and L.-T. Chen participated in the collection and analysis of data from the patient records. C.-Y. Dai participated in the analysis of the hepatitis virus data. C.-F. Hsiao participated in analysis of the biostatistics. J.-D. Huang participated in the analysis of pharmacokinetics. All authors have approved the final version of the article, including the authorship list.
We would like to express our sincere thanks to the patients and families who participated in this study and the research nurses who contributed to this study: Ms. Tsai-Rong Chung and Hsiao-Wei Wu (National Heath Research Institutes). This investigator-initiated study was sponsored by Novartis Pharma AG, Basel, Switzerland, with administration and personnel support from the National Institute of Cancer Research, National Health Research Institutes, Taiwan (DOH99-TD-C-111-004).
Declaration of personal interests: J. Whang-Peng: has received honoraria for scientific presentations from Novartis Pharmaceuticals. L.-T. Chen has served as an consultant and advisory board member, received honoraria for scientific presentations and research funding from Novartis Pharmaceuticals.
Declaration of funding interests: This investigator-initiated study received funding and free everolimus samples from Novartis Pharmaceuticals and personnel support from the Department of Health, Executive Yuan, Taiwan (DOH99-TD-C-111-004). Initial data analyses were undertaken by the Biostatistical Center, Taiwan Cooperative Oncology Group, National Institue of Cancer Research. Editorial support was provided by Melanie Leiby of ApotheCom (Yardley, PA, USA) and funded by Novartis Pharmaceuticals.