Women 18 years of age or older, who were either:
Postmenopausal, as defined by at least one of the following criteria:
Age ≥ 60 years;
Age <60 years and cessation of regular menses for ≥ 12 consecutive months with no alternative pathological or physiological cause; and serum oestradiol and follicle stimulating hormone (FSH) level within the laboratory’s reference range for postmenopausal females;
Documented bilateral oophorectomy;
Medically confirmed ovarian failure
Pre/ perimenopausal, that is not meeting the criteria for being postmenopausal.
if amenable to be treated with the LHRH agonist goserelin. Patients were to have commenced treatment with goserelin or an alternative LHRH agonist at least 4 weeks prior to randomisation. But, if patients had received an alternative LHRH agonist prior to study entry, they were to switch to goserelin for the duration of the study.
Histologically or cytologically proven diagnosis of breast cancer with evidence of metastatic or locally advanced disease, not amenable to resection or radiation therapy with curative intent.
Documentation of ER-positive and/or PR-positive tumour (≥ 1% positive stained cells) based on most recent tumour biopsy (unless bone-only disease, see below) utilising an assay consistent with local standards.
Documented HER2-negative tumour based on local testing on most recent tumour biopsy.
Patients were to satisfy the following criteria for prior therapy:
Progressed during treatment or within 12 months of completion of adjuvant therapy with an aromatase inhibitor if postmenopausal, or tamoxifen if pre- or perimenopausal.
Progressed while on or within 1 month after the end of prior aromatase inhibitor therapy for advanced/metastatic breast cancer if postmenopausal, or prior endocrine treatment for advanced/metastatic breast cancer if pre- or perimenopausal. One previous line of chemotherapy for advanced/metastatic disease was allowed in addition to endocrine therapy.
Except where prohibited by local regulations, all patients were to agree to provide and had available a formalin fixed paraffin embedded (FFPE) tissue biopsy sample taken at the time of presentation with recurrent or metastatic disease. A de novo biopsy was required if no archived tissue taken at the time of presentation with recurrent/metastatic disease was available. The sole exceptions were those patients with bone-only disease for whom provision of previous archival tissue only was acceptable. Patients who had surgery within the last 3 years (but without neoadjuvant chemotherapy prior to surgery) and relapsed while receiving adjuvant therapy may provide a tumor specimen from that surgery.
Measurable disease as defined by RECIST version 1.1, or bone-only disease. Patients with bone-only metastatic cancer were to have a lytic or mixed lytic-blastic lesion that could be accurately assessed by computed tomography (CT) or magnetic resonance imaging (MRI). Patients with bone-only disease and blastic-only metastasis were not eligible. Tumour lesions previously irradiated or subjected to other locoregional therapy were only deemed measurable if progression at the treated site after completion of therapy was clearly documented.
Eastern Cooperative Oncology Group (ECOG) performance status 0-1.
Adequate organ and marrow function defined as follows:
Absolute neutrophil count (ANC) ≥ 1,500/mm3 (1.5 x 109/L);
Platelets ≥ 100,000/mm3 (100 x 109/L);
Haemoglobin ≥ 9 g/dL (90 g/L);
Serum creatinine ≤1.5 x upper limit of normal (ULN) or estimated CrCL ≥ 60 mL/min;
Total serum bilirubin ≤1.5 x ULN (<3ULN if Gilbert’s disease);
AST and/or ALT ≤3 x ULN (≤5.0 x ULN if liver metastases present);
ALP ≤2.5 x ULN (≤5 x ULN if bone or liver metastases present).
Resolution of all acute toxic effects of prior therapy or surgical procedures to National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) Grade ≤1 (except alopecia).
Evidence of a personally signed and dated informed consent document.
Patients who were willing and able to comply with scheduled visits, treatment plan, laboratory tests, and other study procedures.
One previous line of chemotherapy for advanced/metastatic disease was allowed in addition to endocrine therapy. This means a proportion of these patients would be enrolling having had at least 2 prior treatments for their metastatic disease. It is difficult to determine under what circumstances postmenopausal women with ER-positive disease would receive chemotherapy followed by endocrine therapy in the metastatic setting, other than for visceral crises. If so, such patients may represent a poorer prognostic group; both due to the presence of such metastases and more lines of prior therapy.
Postmenopausal women were only eligible if their disease had progressed on an aromatase inhibitor that is, they were not eligible if they had only received tamoxifen. Notably a premenopausal patient who had undergone bilateral oophorectomy but remained on tamoxifen (clinical practice would not necessarily be to change to an aromatase inhibitor in these circumstances) was considered to have been ineligible- such criteria are likely to lead to such protocol violations.
Patients who met any of the following exclusion criteria were not included in the study:
Prior treatment with any CDK inhibitor, or fulvestrant, or with everolimus, or any agent whose mechanism of action is to inhibit the PI3K-mTOR pathway.
Patients with advanced/metastatic, symptomatic, visceral spread, that were at risk of life-threatening complications in the short term (including patients with massive uncontrolled effusions [pleural, pericardial, peritoneal], pulmonary lymphangitis, and over 50% liver involvement).
Known active uncontrolled or symptomatic CNS metastases, carcinomatous meningitis, or leptomeningeal disease as indicated by clinical symptoms, cerebral oedema, and/or progressive growth. Patients with a history of CNS metastases or cord compression are eligible if they had been definitively treated (for example, radiotherapy, stereotactic surgery) and were clinically stable off anticonvulsants and steroids for at least 4 weeks before randomisation.
Current use of food or drugs known to be potent CYP3A4 inhibitors, drugs known to be potent CYP3A4 inducers and drugs that are known to prolong the QT interval.
Major surgery, chemotherapy, radiotherapy, or other anti-cancer therapy within 2 weeks before randomisation. Patients who received prior radiotherapy to ≥ 25% of bone marrow were not eligible independent of when it had been received.
Any other malignancy within 3 years prior to randomisation, except for adequately treated basal cell or squamous cell skin cancer, or carcinoma in situ of the cervix.
QTc interval >480 ms, family or personal history of long or short QT syndrome, Brugada syndrome or known history of QTc prolongation or Torsade de Pointes.
Any of the following within 6 months of randomisation: myocardial infarction, severe/unstable angina, ongoing cardiac dysrhythmias of NCI CTCAE Grade ≥ 2, atrial fibrillation of any grade, coronary/peripheral artery bypass graft, symptomatic congestive heart failure, cerebrovascular accident including transient ischemic attack, or symptomatic pulmonary embolism.
Impairment of gastrointestinal (GI) function or GI disease that might have significantly altered the absorption of palbociclib, such as history of GI surgery with might have resulted in intestinal blind loops and patients with clinically significant gastroparesis, short bowel syndrome, unresolved nausea, vomiting, active inflammatory bowel disease or diarrhea of CTCAE Grade >1.
Prior hematopoietic stem cell or bone marrow transplantation.
Known abnormalities in coagulation such as bleeding diathesis, or treatment with anticoagulants precluding intramuscular injections of fulvestrant or goserelin (if applicable).
Known or possible hypersensitivity to fulvestrant, goserelin, any of their excipients or to any palbociclib/placebo excipients.
Known human immunodeficiency virus infection.
Other severe acute or chronic medical or psychiatric condition, including recent or active suicidal ideation or behaviour, or laboratory abnormality that might have increased the risk associated with study participation or investigational product administration or might have interfered with the interpretation of study results and, in the judgment of the investigator, would have made the patient inappropriate for entry into this study.
Patients who were investigational site staff members directly involved in the conduct of the study and their family members, site staff members otherwise supervised by the Investigator, or patients who are Pfizer employees directly involved in the conduct of the study.
Participation in other studies involving investigational drugs (Phases 1-4) within 4 weeks before randomisation in the current study.
Comment: The exclusion criterion of suicidal ideation or behaviour is again noted (it was also in the study protocol of Study 1008 but not for 1003). The sponsor is requested to explain the rationale behind this exclusion criterion and provide details of any details where palbociclib might have been implicated in causing patients to commit suicide or become suicidal that is, while taking or after recently stopping palbociclib (Clinical Questions).
Patients in Arm A (at least 278) received palbociclib 125 mg/day orally for 3 weeks
followed by 1 week off plus fulvestrant 500 mg intramuscularly on Days 1 and 15 of Cycle 1, every 28 days (+/- 7 days) thereafter starting from Day 1 of Cycle 1.
Patients in Arm B (at least 139) received placebo orally daily for 3 weeks followed by 1 week off plus fulvestrant 500 mg intramuscularly on Days 1 and 15 of Cycle 1, every 28 days (+/- 7 days) thereafter starting from Day 1 of Cycle 1.
In both arms, pre- and peri-menopausal women also received the LHRH agonist goserelin (Zoladex or generic) which must have been commenced at least 4 weeks prior to randomisation.
Patients were to continue to receive assigned treatment until objective progressive disease (PD), symptomatic deterioration, unacceptable toxicity, death, or withdrawal of consent, whichever occurs first. Crossover between treatment arms will not be allowed.
Patients continued to receive assigned treatment until objective Progressive Disease (PD), symptomatic deterioration, unacceptable toxicity, death, or withdrawal of consent, whichever occurs first.
Dose reductions were permitted for the palbociclib/placebo but not fulvestrant, and fulvestrant was not to be delayed by more than 7 days. Palbociclib doses could be reduced to 100 mg daily and 75 mg daily on 3/1 schedule, respectively, or to 75 mg on a 2-week on/2-week off (2/2) schedule. Where palbociclib/placebo dose delays occurred, administration of fulvestrant and goserelin was scheduled to continue according to the pre-planned schedule.
Efficacy variables and outcomes
Baseline disease assessment for all patients (within 28 days of randomisation):
CT or MRI scan of the chest, abdomen and pelvis.
CT or MRI scan of any other sites of disease as clinically indicated.
Clinical assessment of superficial disease which included photographs of all superficial metastatic lesions. All lesion measurements were recorded in the CRF.
Bone scans; any suspicious abnormalities (that is, hotspots) identified on the bone scans at baseline were confirmed by X-ray, CT scan with bone windows or MRI. Bone lesions identified at baseline as the only site of disease followed the same assessment schedule as for measurable lesions that is, bonescan plus additional imaging modality used to confirm/characterise at baseline.
Disease assessments were performed:
Every 8 weeks (±7 days) for the first year, then every 12 weeks (±7 days) until documented PD as per RECIST v.1.1, study treatment discontinuation (for patients continuing treatment beyond RECIST-defined disease progression), initiation of new anticancer therapy, or discontinuation of patient from overall study participation (for example, death, patient's request, lost to follow-up).
The same tumour assessment technique had to be used throughout the study for a given lesion/patient.
Bone scans were not routinely performed in those with no bone lesions at baseline, unless clinically or biochemically indicated but were required to confirm a CR
Interpretation of PD for bone-only disease was if:
The malignant nature of one or more new lesions identified with bone scan is
confirmed with X-ray, or CT, or MRI scan;
Flare observed in bone scan is followed by confirmation of progression with other imaging modalities;
Clinical worsening of the disease is assessed by bone scan and disease progression (that is, new lesion(s)) is confirmed with other imaging modalities;
Unequivocal progression of existing bone lesions is observed.
Interpretation will be SD if:
The malignant nature of all the new lesions identified with bone scan is not confirmed.
If receiving a clinical benefit, patients can continue treatment with the study treatment beyond RECIST-proven PD; or in presence of toxicities, patients can continue to receive fulvestrant alone. Patients discontinuing the active treatment phase (that is, discontinuing both palbociclib/placebo and fulvestrant) entered a follow-up phase during which survival and new anti-cancer therapy information will be collected, initially every 3 months and then every 6 months.
In the following cases the patient were to be censored at the date of prior tumour assessment with no PD: 1) on-study fracture; 2) on-study management of pain (palliative radiation therapy, palliative surgery), 3) clinical worsening not objectively confirmed; 4) on-study change of therapy. In all the censored cases (no objectively documented PD) tumor assessment will be performed until PD. Also, it will be at the discretion of the investigator to discontinue the study treatment.
Patients who discontinued study treatment for reasons other than objectively documented disease progression as per RECIST definitions:
were not to be recorded on the CRF as PD but as ‘off treatment due to Global Deterioration of Health Status’;
continued to have tumour assessments performed during the follow-up visits every 8 weeks (±7 days) for the first year, and then every 12 weeks (±7 days) until
documented disease progression, initiation of new anticancer therapy or discontinuation of patient from overall study participation (for example, death, patient's request, lost to follow-up).
Comment: The detail regarding bony progression and interpretation study is similar to that in the protocol for Study 1008 (see Study 1008 Efficacy Variables and Outcomes). Assessment of the BICR analyses will help determine whether this reduces some of the discordance observed between investigator assessments and BICR assessments in Study 1003.
The censoring rules for PFS determination were included to inform regarding definitions of PFS.
Following the End of Treatment visit, survival status will be collected in all patients (telephone contact is acceptable) every 3 months (Month 3, 6, and 9, ±14 days) then every 6 months starting at Month 15 (±14 days) from the last dose of study treatment. Information on start, stop and type of subsequent anticancer therapy was also to be collected.
Comment: The collection of OS data in this study was more intensive than in Study 1008 (every 6 months following end of treatment) and accordingly, has potential to provide more accurate assessments of OS.
All primary and secondary endpoints based on radiological (and photographical where applicable) assessments of tumour burden (that is, PFS, OR, DR, CBR) were derived using the local radiologist’s/investigator’s tumour assessments as primary data source.
Blinded independent central review (BICR)
BICR of radiographic and clinical data for a randomly selected subgroup of patients (approximately 40%) was undertaken. The independent reviewers assessed tumour progression based on the review of scans, physical examination data and other data from the final data cut for this randomly selected subgroup of the study population. These were to verify investigators’ findings and be used for supportive analyses.
Trough concentrations of palbociclib were collected from all patients for exposure/response analysis for safety and efficacy findings. The SAP describes these as exploratory.
Patient Reported Outcomes (PRO) will be collected in this trial using:
the EuroQol-5D (EQ-5D);
European Organisation for Research and Treatment of Cancer Quality of Life questionnaire (EORTC-QLQ-C30);
breast cancer (EORTC-QLQ-BR23).
Under Secondary endpoints, the protocol states:
‘Tumor tissue biomarkers, including genes (for example, copy numbers of CCND1 and CDKN2A, PIK3CA mutations), proteins (for example, Ki67, pRb, CCNE1), and RNA expression (for example, cdk4, cdk6)’ (Study Protocol dated 30 September 2014; 2 page numbers listed on same page: 9 and 11).
Therefore, no specific biomarkers other than ER and HER2 were prespecified in the protocol.
Comment: These biomarker studies are important but essentially exploratory in nature from a statistical perspective.
Progression-Free Survival (PFS) as assessed by the Investigator.
Overall Survival (OS);
1-year, 2-year, and 3-year survival probabilities;
Objective Response (OR): Complete Response (CR) or Partial Response (PR);
Duration of Response (DR);
Clinical Benefit Response (CBR): CR or PR or Stable Disease (SD) >24 weeks;
Type, incidence, severity [NCI CTCAE] v4.0), seriousness and relationship to study medications of Adverse Events (AEs) and any laboratory abnormalities;
Trough plasma concentration of palbociclib, fulvestrant and goserelin (if applicable) in the subgroup of approximately 40 patients included in the initial safety review;
Tumour tissue biomarkers;
Randomisation and blinding methods
Patients deemed eligible by the sponsor on the basis of information provided in forms submitted from the trial site, were randomised into the study by interactive randomisation technology (IRT).
This form included the following information needed for patient stratification:
documented sensitivity to prior hormonal therapy (yes vs. no);
menopausal status at study entry (pre-/peri- vs. postmenopausal);
presence of visceral metastases (yes vs. no).
Blinding codes could be broken in emergency situations for patient safety, or where on disease progression, it was deemed necessary to select the next therapy, after discussion with the sponsor. Wherever the blinding code was broken, site staff were to document the reasons and date but not communicate the results to sponsor personnel.
Intent-to-Treat Population (Full Analysis Set)
The intent-to-treat (ITT) population or full analysis set will include all patients who are randomised, with study drug assignment designated according to initial randomisation, regardless of whether patients receive study drug or receive a different drug from that to which they were randomised. The ITT population will be the primary population for evaluating all efficacy endpoints and patient characteristics.
As-Treated (AT) Population (Safety Analysis Set)
The as-treated (AT) population or safety analysis set will include all patients who receive at least 1 dose of study medication, with treatment assignments designated according to actual study treatment received. The AT population will be the primary population for evaluating treatment administration/compliance and safety. Efficacy and clinical benefit endpoints may be assessed in this population as well.
Biomarker Analysis Set
A subset of AT patients, who have both baseline and at least one follow-up values for > 1 biomarker.
Patient Reported Outcome (PRO) Evaluable Population (PRO Analysis Set)
The PRO –evaluable population is defined as a subset of ITT patients, who have completed a baseline and at least one post –baseline PRO assessment prior to end of study treatment.
Comment: The proportion of patients from the ITT completing the assessment will be evaluated when determining the relevance of the findings and the validity of this PRO analysis set. It is noted that ‘prorating’ was to be used for missing data for the QLQ-C30 and QLQ-BR23 if at least half of the scales had been answered. This introduces potential bias and makes assumptions about patients’ highly personal, individual and subjective responses. If a substantial number of these are incomplete, then the appropriateness of the tool should be re-examined rather than extrapolations made.
At least 417 patients were initially planned to be randomisation in a 2:1 ratio and stratified by documented sensitivity to prior hormonal therapy (yes vs. no), menopausal status at study entry (pre-/peri- vs. postmenopausal), and presence of visceral metastases (yes vs. no).
The sample size for this study was determined based on the median PFS for the control arm in this study being assumed to be 6.0 months. Therefore, an improvement of 56% to a median PFS of 9.38 months (corresponding to a HR=0.64) would be considered clinically meaningful. A total of 238 PFS events will be required in the two treatment arms for the study to have a 90% power to detect an increase in PFS assuming a true HR of 0.64 (representing a 56% increase in median PFS from 6 to 9.38 months ), if tested at a 1-sided significance level of p=0.025.
Assuming a non-uniform accrual accomplished over a period of about 14 months, data follow-up for approximately 20 months from the start of study randomisation for final PFS analysis, and a non-uniform dropout with dropout rate of 25% at 18 months for PFS, a total sample size of 417 patients (278 in the fulvestrant plus palbociclib arm and 139 in the placebo plus fulvestrant arm) is required.
Comment: The anticipated dropout rate was high at 25% for those with metastatic disease and high degree of motivation to continue treatment if no disease progression. The sponsor is requested to provide a rationale for this. Was this to reflect anticipated side effects related to the use of fulvestrant, the administration of which is associated with significant discomfort?
The median OS for women with recurrent advanced or metastatic breast cancer treated with AI and fulvestrant monotherapy is assumed to be 24 months. With an overall one-sided p of 0.025 and one interim analysis of OS (at the time of final PFS analysis), the study will have approximately 80% power to detect a Hazard Ratio (HR) of 0.65 (representing a 54% increase in median OS from 24 months to 37 months) when 198 deaths have occurred.
The study was designed to have one interim analysis to allow for early stopping of the study due to efficacy, or to potentially re-estimate the sample size of the trial based upon the primary endpoint of PFS. The safety of the combination was also to be assessed at the interim analysis. The interim analysis of PFS will be performed after approximately 143 patients have documented PD or death (approximately 60% of the total events expected). The information fraction for the interim analysis may be adjusted if needed. The sample size of the study may also be adjusted as appropriate. Rules for determining PFS and censoring were included.
Only one interim analysis of OS was planned. This was to be hierarchically tested for significance at the time of PFS analyses, provided the primary PFS endpoint is statistically significant at the interim and/or final PFS analyses. At that point, it was estimated that 97 deaths would have occurred; if OS is not significant at the interim analysis, a final analysis will be performed after 198 deaths. With an overall one-sided α of 0.025 and one interim analysis of OS (at the time of PFS final analysis), the study will have approximately 80% to detect a HR of 0.65 (representing a 54% increase in median OS from 24 months to 37 months) when 198 deaths have occurred.
An external DMC was appointed to undertake the following:
review PK data after approximately 40 patients had been enrolled to determine safety of the study combination;
make recommendation as to whether or not the trial should continue based on ongoing reviews of safety data;
evaluate interim efficacy data and make a recommendation regarding study continuation based on observed results of the study;
When the protocol was amended to include ophthalmic assessments, a review of the number required for the sample size for a pre-specified analysis was only to be shared with the DMC before the interim analysis. The SAP states that there will be descriptive statistics and potentially pooled results if there were insufficient numbers to undertake comparative analyses.
A BICR was undertaken in 40% of the total population – assuming investigator assessment and BICR results are similar and the estimated log of investigator-based HR is -0.45 (HR=0.64), the audit size of 40% will ensure that the upper bound of a one-sided 95% CI for BICR-based treatment effect (log-hazard ratio) has 90% probability of being below zero if the correlation (ρ) between investigator assessment and BICR is 0.76 and the standard error is 0.39.